Guide to Six Sigma and Process Improvement – Global Homework Experts

From the Library of Pearson HED
A Guide to Six
Sigma and Process
Improvement
for Practitioners
and Students
Second Edition
From the Library of Pearson HED
This page intentionally left blank
From the Library of Pearson HED
A Guide to Six
Sigma and Process
Improvement
for Practitioners
and Students
Foundations, DMAIC, Tools,
Cases, and Certification
Second Edition
Howard S. Gitlow
Richard J. Melnyck
David M. Levine
From the Library of Pearson HED
Publisher: Paul Boger
Editor-in-Chief: Amy Neidlinger
Executive Editor: Jeanne Glasser Levine
Operations Specialist: Jodi Kemper
Cover Designer: Alan Clements
Managing Editor: Kristy Hart
Senior Project Editor: Betsy Gratner
Copy Editor: Geneil Breeze
Proofreader: Laura Hernandez
Indexer: Erika Millen
Senior Compositor: Gloria Schurick
Manufacturing Buyer: Dan Uhrig
© 2015 by Pearson Education, Inc.
Old Tappan , New Jersey 07675
For information about buying this title in bulk quantities, or for special sales opportunities (which may include
electronic versions; custom cover designs; and content particular to your business, training goals, marketing
focus, or branding interests), please contact our corporate sales department at [email protected] or
(800) 382-3419.
For government sales inquiries, please contact [email protected] .
For questions about sales outside the U.S., please contact [email protected] .
Company and product names mentioned herein are the trademarks or registered trademarks of their respective
owners.
All rights reserved. No part of this book may be reproduced, in any form or by any means, without permission in
writing from the publisher.
Printed in the United States of America
First Printing May 2015
ISBN-10: 0-13-392536-6
ISBN-13: 978-0-13-392536-4
Pearson Education LTD.
Pearson Education Australia PTY, Limited
Pearson Education Singapore, Pte. Ltd.
Pearson Education Asia, Ltd.
Pearson Education Canada, Ltd.
Pearson Educación de Mexico, S.A. de C.V.
Pearson Education—Japan
Pearson Education Malaysia, Pte. Ltd.
Library of Congress Control Number: 2015932281
From the Library of Pearson HED
This book is dedicated to:
Shelly Gitlow
Ali Gitlow
Abraham Gitlow
Beatrice Gitlow
Jack Melnyck
Eileen Melnyck
Lee Levine
Reuben Levine
From the Library of Pearson HED
Contents
Section I Building a Foundation of Process Improvement
Fundamentals
Chapter 1 You Don’t Have to Suffer from the Sunday Night Blues! . . . . . . . . .1
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
Sarah’s Story . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
Nine Principles of Process Improvement to Get the Most Out of This Book . . . . . . . .3
Structure of the Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14
Let’s Go! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16
Chapter 2 Process and Quality Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . .17
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
Process Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
What Is a Process? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
Where Do Processes Exist? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
Why Does Understanding Processes Matter?. . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
What Is a Feedback Loop and How Does It Fit into the Idea of a Process? . . . .19
Some Process Examples to Bring It All Together! . . . . . . . . . . . . . . . . . . . . . . . . .19
Variation Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
What Is Variation in a Process? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
Why Does Variation Matter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
What Are the Two Types of Variation? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
How to Demonstrate the Two Types of Variation . . . . . . . . . . . . . . . . . . . . . . . . .27
Red Bead Experiment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30
Quality Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
Goal Post View of Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
Continuous Improvement Definition of Quality—Taguchi Loss Function . . . .32
More Quality Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34
From the Library of Pearson HED
Contents vii
Chapter 3 Defining and Documenting a Process . . . . . . . . . . . . . . . . . . . . . . . .35
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
A Story to Illustrate the Importance of Defining and Documenting a Process . . . . .35
Fundamentals of Defining a Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36
Who Owns the Process? Who Is Responsible for the Improvement
of the Process?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36
What Are the Boundaries of the Process? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
What Are the Process’s Objectives? What Measurements Are Being
Taken on the Process with Respect to Its Objectives? . . . . . . . . . . . . . . . . . . . . . .38
Fundamentals of Documenting a Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39
How Do We Document the Flow of a Process? . . . . . . . . . . . . . . . . . . . . . . . . . . .39
Why and When Do We Use a Flowchart to Document a Process? . . . . . . . . . . .39
What Are the Different Types of Flowcharts and When Do We Use Each? . . . .40
What Method Do We Use to Create Flowcharts?. . . . . . . . . . . . . . . . . . . . . . . . . .43
Fundamentals of Analyzing a Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
How Do We Analyze Flowcharts? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
Things to Remember When Creating and Analyzing Flowcharts . . . . . . . . . . . .45
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46
Section II Creating Your Toolbox for Process Improvement
Chapter 4 Understanding Data: Tools and Methods. . . . . . . . . . . . . . . . . . . . .47
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47
What Is Data?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47
Types of Numeric Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47
Graphing Attribute Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50
Bar Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50
Pareto Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
Line Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52
Graphing Measurement Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
Histogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
Dot Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55
Run Chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56
Measures of Central Tendency for Measurement Data . . . . . . . . . . . . . . . . . . . . . . . . .59
Mean. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59
Median . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
From the Library of Pearson HED
viii Contents
Measures of Central Tendency for Attribute Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61
Proportion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61
Measures of Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62
Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62
Sample Variance and Standard Deviation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63
Understanding the Range, Variance, and Standard Deviation . . . . . . . . . . . . . .64
Measures of Shape. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66
Skewness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66
More on Interpreting the Standard Deviation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68
How-To Guide for Understanding Data: Minitab 17 User Guide . . . . . . . . . . . . . . . .70
Using Minitab Worksheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70
Opening and Saving Worksheets and Other Components . . . . . . . . . . . . . . . . . .71
Obtaining a Bar Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74
Obtaining a Pareto Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .76
Obtaining a Line Graph (Time Series Plot). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .78
Obtaining a Histogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .79
Obtaining a Dot Plot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .82
Obtaining a Run Chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .84
Obtaining Descriptive Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .85
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .88
Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .88
Chapter 5 Understanding Variation: Tools and Methods . . . . . . . . . . . . . . . .89
What Are the Objectives of This Chapter?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .89
What Is Variation? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .89
Common Cause Variation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .89
Special Cause Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .90
Using Control Charts to Understand Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .90
Attribute Control Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .90
Variables Control Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91
Understanding Control Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91
Rules for Determining Out of Control Points. . . . . . . . . . . . . . . . . . . . . . . . . . . . .93
Control Charts for Attribute Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98
P Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98
C Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .104
U Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .106
From the Library of Pearson HED
Contents ix
Control Charts for Measurement Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .108
Individuals and Moving Range (I-MR) Charts. . . . . . . . . . . . . . . . . . . . . . . . . . .109
X Bar and R Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .112
X Bar and S Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115
Which Control Chart Should I Use? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119
Control Chart Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119
Measurement Systems Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126
Measurement System Analysis Checklist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126
Gage R&R Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .127
How-To Guide for Understanding Variation: Minitab User Guide
(Minitab Version 17, 2013) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .131
Using Minitab to Obtain Zone Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .131
Using Minitab for the P Chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .131
Using Minitab for the C Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .134
Using Minitab for the U Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .134
Using Minitab for the Individual Value and Moving Range Charts . . . . . . . . .136
Using Minitab for the X Bar and R Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .137
Using Minitab for the X Bar and S Charts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .139
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .142
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143
Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143
Chapter 6 Non-Quantitative Techniques: Tools and Methods . . . . . . . . . . .145
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .145
High Level Overview and Examples of Non-Quantitative Tools and Methods . . . .145
Flowcharting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .146
Voice of the Customer (VoC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .146
Supplier-Input-Process-Output-Customer (SIPOC) Analysis . . . . . . . . . . . . . .149
Operational Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .151
Failure Modes and Effects Analysis (FMEA) . . . . . . . . . . . . . . . . . . . . . . . . . . . .153
Check Sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .153
Brainstorming. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .155
Affinity Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .156
Cause and Effect (Fishbone) Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .157
Pareto Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .159
Gantt Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .159
Change Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .160
Communication Plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .163
From the Library of Pearson HED
x Contents
How-To Guide for Using Non-Quantitative Tools and Methods . . . . . . . . . . . . . . . .165
How to Do Flowcharting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .165
How to Do a Voice of the Customer (VoC) Analysis. . . . . . . . . . . . . . . . . . . . . .166
How to Do a SIPOC Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .172
How to Create Operational Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173
How to Do a Failure Modes and Effects Analysis (FMEA) . . . . . . . . . . . . . . . . .174
How to Do Check Sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .177
Brainstorming. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .179
How to Do Affinity Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181
How to Do Cause and Effect Diagrams (C&E Diagrams) . . . . . . . . . . . . . . . . . .182
How to Do Pareto Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .182
How to Do Gantt Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .185
How to Use Change Concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .185
How to Do Communication Plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .198
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .200
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .201
Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .201
Chapter 7 Overview of Process Improvement Methodologies. . . . . . . . . . . .203
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .203
SDSA Cycle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .203
SDSA Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .204
PDSA Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .206
PDSA Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .207
Kaizen/Rapid Improvement Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .209
Kaizen/Rapid Improvement Events Example. . . . . . . . . . . . . . . . . . . . . . . . . . . .210
DMAIC Model: Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .212
Define Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .213
Measure Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .213
Analyze Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .214
Improve Phase. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .216
Control Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .216
DMAIC Model Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .216
DMADV Model: Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .218
Define Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .218
Measure Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .218
Analyze Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .218
Design Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .219
From the Library of Pearson HED
Contents xi
Verify/Validate Phase. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .219
DMADV Model Example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .219
Lean Thinking: Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .221
The 5S Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .221
Total Productive Maintenance (TPM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .223
Quick Changeover (Single Minute Exchange of Dies—SMED) . . . . . . . . . . . . .224
Poka-Yoke . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .225
Value Streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .226
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .227
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .228
Chapter 8 Project Identification and Prioritization:
Building a Project Pipeline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .231
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .231
Project Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .231
Internal Proactive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .232
Internal Reactive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .234
External Proactive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .235
External Reactive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .236
Using a Dashboard for Finding Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .238
Structure of a Managerial Dashboard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .238
Example of a Managerial Dashboard. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .239
Managing with a Dashboard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .240
Project Screening and Scoping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .240
Questions to Ask to Ensure Project Is Viable . . . . . . . . . . . . . . . . . . . . . . . . . . . .241
Estimating Project Benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .242
Project Methodology Selection—Which Methodology Should I Use? . . . . . . .243
Estimating Time to Complete Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .245
Creating a High Level Project Charter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .246
Problem Statement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .247
Prioritizing and Selecting Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .247
Prioritizing Projects Using a Project Prioritization Matrix . . . . . . . . . . . . . . . .248
Final Project Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .250
Executing and Tracking Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .250
Allocating Resources to Execute the Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . .250
Monthly Steering Committee (Presidential) Reviews . . . . . . . . . . . . . . . . . . . . .251
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .251
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .252
From the Library of Pearson HED
xii Contents
Section III Putting It All Together—Six Sigma Projects
Chapter 9 Overview of Six Sigma Management . . . . . . . . . . . . . . . . . . . . . . . .253
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253
Non-Technical Definition of Six Sigma Management . . . . . . . . . . . . . . . . . . . . . . . . .253
Technical Definition of Six Sigma. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253
Where Did Six Sigma Come From? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253
Benefits of Six Sigma Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .254
Key Ingredient for Success with Six Sigma Management . . . . . . . . . . . . . . . . . . . . . .255
Six Sigma Roles and Responsibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .255
Senior Executive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .255
Executive Steering Committee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .256
Project Champion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .256
Process Owner. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .257
Master Black Belt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .257
Black Belt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .258
Green Belt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .259
Green Belt Versus Black Belt Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .260
Six Sigma Management Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .260
Next Steps: Understanding the DMAIC Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .264
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .264
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .265
Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .265
Appendix 9.1 Technical Definition of Six Sigma Management . . . . . . . . . . . . . . . . .266
Chapter 10 DMAIC Model: “D” Is for Define. . . . . . . . . . . . . . . . . . . . . . . . . . .273
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .273
Purpose of the Define Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .273
The Steps of the Define Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .274
Activate the Six Sigma Team. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .274
Project Charter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .276
SIPOC Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .283
Voice of the Customer Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .286
Definition of CTQ(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .288
Create an Initial Draft of the Project Objective. . . . . . . . . . . . . . . . . . . . . . . . . . .289
Tollgate Review: Go-No Go Decision Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . .290
Keys to Success and Pitfalls to Avoid in the Define Phase. . . . . . . . . . . . . . . . . . . . . .291
From the Library of Pearson HED
Contents xiii
Case Study of the Define Phase: Reducing Patient No Shows in an
Outpatient Psychiatric Clinic—Define Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .292
Activate the Six Sigma Team. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .292
Project Charter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .293
SIPOC Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .299
Voice of the Customer Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .299
Definition of CTQ(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .308
Initial Draft Project Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .308
Tollgate Review: Go-No Go Decision Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . .308
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .309
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .310
Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .310
Chapter 11 DMAIC Model: “M” Is for Measure . . . . . . . . . . . . . . . . . . . . . . . . .311
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .311
Purpose of the Measure Phase. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .311
The Steps of the Measure Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .312
Operational Definitions of the CTQ(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .312
Data Collection Plan for CTQ(s). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .312
Validate Measurement System for CTQ(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .313
Collect and Analyze Baseline Data for the CTQ(s). . . . . . . . . . . . . . . . . . . . . . . .317
Estimate Process Capability for CTQ(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .321
Keys to Success and Pitfalls to Avoid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .324
Case Study: Reducing Patient No Shows in an Outpatient Psychiatric
Clinic—Measure Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .324
Operational Definition of the CTQ(s) and Data Collection
Plan for the CTQ(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .325
Validate Measurement System for CTQ(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .326
Collect and Analyze Baseline Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .328
Tollgate Review: Go-No Go Decision Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . .330
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .330
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .331
Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .331
Chapter 12 DMAIC Model: “A” Is for Analyze. . . . . . . . . . . . . . . . . . . . . . . . . .333
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .333
Purpose of the Analyze Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .333
The Steps of the Analyze Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .334
Detailed Flowchart of Current State Process . . . . . . . . . . . . . . . . . . . . . . . . . . . .334
Identification of Potential Xs for CTQ(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .335
From the Library of Pearson HED
xiv Contents
Failure Modes and Effects Analysis (FMEA) to Reduce the Number of Xs . . .338
Operational Definitions for the Xs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .338
Data Collection Plan for Xs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .339
Validate Measurement System for X(s). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .340
Test of Theories to Determine Critical Xs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .340
Develop Hypotheses/Takeaways about the Relationships between the
Critical Xs and CTQ(s). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .342
Go-No Go Decision Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .342
Keys to Success and Pitfalls to Avoid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .343
Case Study: Reducing Patient No Shows in an Outpatient Psychiatric Clinic—
Analyze Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .344
Detailed Flowchart of Current State Process . . . . . . . . . . . . . . . . . . . . . . . . . . . .344
Identification of Xs for CTQ(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .344
Failure Modes and Effects Analysis (FMEA) to Reduce the Number of Xs . . .346
Operational Definitions of the Xs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .346
Data Collection Plan for Xs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .348
Validate Measurement System for Xs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .348
Test of Theories to Determine Critical Xs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .348
Develop Hypotheses/Takeaways about the Relationships between the
Critical Xs and CTQ(s). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .354
Tollgate Review—Go-No Go Decision Point . . . . . . . . . . . . . . . . . . . . . . . . . . . .354
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .355
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .356
Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .356
Chapter 13 DMAIC Model: “I” Is for Improve . . . . . . . . . . . . . . . . . . . . . . . . . .357
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .357
Purpose of the Improve Phase. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .357
The Steps of the Improve Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .358
Generate Alternative Methods for Performing Each Step in the Process . . . . .358
Select the Best Alternative Method (Change Concepts) for All of the CTQs . .360
Create a Flowchart for the Future State Process. . . . . . . . . . . . . . . . . . . . . . . . . .361
Identify and Mitigate the Risk Elements for New Process . . . . . . . . . . . . . . . . .362
Run a Pilot Test of the New Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .362
Collect and Analyze the Pilot Test Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .362
Go-No Go Decision Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .364
Keys to Success and Pitfalls to Avoid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .365
From the Library of Pearson HED
Contents xv
Case Study: Reducing Patient No Shows in an Outpatient Psychiatric Clinic—
Improve Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .366
Generate Alternative Methods for Performing Each Step in the Process . . . . .366
Select the Best Alternative Method (Change Concept) for All the CTQs . . . . .367
Create a Flowchart of the New Improved Process . . . . . . . . . . . . . . . . . . . . . . . .368
Identify and Mitigate the Risk Elements for the New Process . . . . . . . . . . . . . .369
Run a Pilot Test of the New Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .369
Collect and Analyze the Pilot Test Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .372
Tollgate Review—Go-No Go Decision Point . . . . . . . . . . . . . . . . . . . . . . . . . . . .373
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .373
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .374
Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .374
Chapter 14 DMAIC Model: “C” Is for Control . . . . . . . . . . . . . . . . . . . . . . . . . .375
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .375
Purpose of the Control Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .375
The Steps of the Control Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .376
Reduce the Effects of Collateral Damage to Related Processes . . . . . . . . . . . . .376
Standardize Improvements (International Standards Organization [ISO]). . .379
Develop a Control Plan for the Process Owner . . . . . . . . . . . . . . . . . . . . . . . . . .380
Identify and Document the Benefits and Costs of the Project . . . . . . . . . . . . . .383
Input the Project into the Six Sigma Database . . . . . . . . . . . . . . . . . . . . . . . . . . .383
Diffuse the Improvements throughout the Organization . . . . . . . . . . . . . . . . . .383
Conduct a Tollgate Review of the Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .384
Keys to Success and Pitfalls to Avoid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .385
Case Study: Reducing Patient No Shows in an Outpatient Psychiatric Clinic—
Control Phase. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .386
Reduce the Effects of Collateral Damage to Related Processes . . . . . . . . . . . . .386
Standardize Improvements (International Standards Organization [ISO]). . .386
Develop a Control Plan for the Process Owner . . . . . . . . . . . . . . . . . . . . . . . . . .387
Financial Impact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .387
Input the Project into the Six Sigma Database . . . . . . . . . . . . . . . . . . . . . . . . . . .390
Diffuse the Improvements throughout the Organization . . . . . . . . . . . . . . . . . .390
Champion, Process Owner, and Black Belt Review the Project . . . . . . . . . . . . .390
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .391
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .392
Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .392
From the Library of Pearson HED
xvi Contents
Chapter 15 Maintaining Improvements in Processes, Products-Services,
Policies, and Management Style. . . . . . . . . . . . . . . . . . . . . . . . . . . .393
What Is the Objective of This Chapter? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .393
Improving Processes, Products-Services, and Processes: Revisited. . . . . . . . . . . . . .393
Case Study 1: Failure in the Act Phase of the PDSA Cycle in Manufacturing. . . . . .393
Case Study 2: Failure in the Act Phase of the PDSA Cycle in
Accounts Receivable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .394
A Method for Promoting Improvement and Maintainability. . . . . . . . . . . . . . . . . . .396
Dashboards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .396
Presidential Review of Maintainability Indicators. . . . . . . . . . . . . . . . . . . . . . . .397
The Funnel Experiment and Successful Management Style . . . . . . . . . . . . . . . . . . . .398
Rule 1 Revisited. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .398
Rule 4 Revisited. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .398
Succession Planning for the Maintainability of Management Style. . . . . . . . . . . . . .399
Succession Planning by Incumbent Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .399
Succession Planning by Creating Talent Pools Model. . . . . . . . . . . . . . . . . . . . .400
Succession Planning Using the Top-Down/Bottom-Up Model . . . . . . . . . . . . .400
Process Oriented Top-Down/Bottom-Up Succession Planning Model . . . . . .401
Egotism of Top Management as a Threat to the Maintainability of
Management Style. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .403
Six Indicators of Egotism That Threaten the Maintainability of
Management Style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .403
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .404
The Board of Directors Fails to Understand the Need for Maintainability in the
Organization’s Culture and Management Style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .404
Definition of Culture/Management Style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .404
Components of Board Culture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .405
Shared Mission and Shared Values/Beliefs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .405
Allocation of Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .405
Reducing Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .406
Engagement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .406
Trust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .406
Takeaways from This Chapter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .406
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .407
From the Library of Pearson HED
Contents xvii
Section IV The Culture Required for Six Sigma
Management
Chapter 16 W. Edwards Deming’s Theory of Management: A Model for
Cultural Transformation of an Organization . . . . . . . . . . . . . . . .409
Background on W. Edwards Deming. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .409
Deming’s System of Profound Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .409
Purpose of Deming’s Theory of Management . . . . . . . . . . . . . . . . . . . . . . . . . . .410
Paradigms of Deming’s Theory of Management . . . . . . . . . . . . . . . . . . . . . . . . .410
Components of Deming’s Theory of Management . . . . . . . . . . . . . . . . . . . . . . .411
Deming’s 14 Points for Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .413
Deming’s 14 Points and the Reduction of Variation . . . . . . . . . . . . . . . . . . . . . . . . . .430
Transformation or Paradigm Shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .433
The Prevailing Paradigm of Leadership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .433
The New Paradigm of Leadership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .434
Transformation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .434
Quality in Service, Government, and Education. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .434
Quotes from Deming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .434
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .435
References and Additional Readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .436
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .439
Section V Six Sigma Certification
Chapter 17 Six Sigma Champion Certification . . . . . . . . . . . . . . . . . . . . . . Online
Chapter 18 Six Sigma Green Belt Certification . . . . . . . . . . . . . . . . . . . . . . Online
ACCESS TO DATA FILES AND CHAPTERS 17 AND 18
Go to www.ftpress.com/sixsigma and click the Downloads tab to access Minitab practice
data files and Chapters 17 and 18.
From the Library of Pearson HED
xviii Acknowledgments
Acknowledgments
First, we thank the late W. Edwards Deming for his philosophy and guidance. Second,
we thank everyone at the University of Miami and the University of Miami Miller School
of Medicine for collaborating with us in all of our process improvement efforts. We have
learned something from every single one of you! Third, we thank all the people who provided
life lessons to us to make this book a reality. Finally, we thank Jeanne Glasser Levine for giving us the opportunity to write this second edition. Thank you one and all.
From the Library of Pearson HED
About the Authors xix
About the Authors
Dr. Howard S. Gitlow is Executive Director of the Institute for the Study of Quality, Director of the Master of Science degree in Management Science, and a Professor of Management
Science, School of Business Administration, University of Miami, Coral Gables, Florida. He
was a visiting professor at the Stern School of Business at New York University from 2007
through 2013, and a visiting professor at the Science University of Tokyo in 1990 where he
studied with Dr. Noriaki Kano. He received his PhD in Statistics (1974), MBA (1972), and
BS in Statistics (1969) from New York University. His areas of specialization are Six Sigma
Management, Dr. Deming’s theory of management, Japanese Total Quality Control, and
statistical quality control.
Dr. Gitlow is a Six Sigma Master Black Belt, a fellow of the American Society for Quality, and
a member of the American Statistical Association. He has consulted on quality, productivity, and related matters with many organizations, including several Fortune 500 companies.
Dr. Gitlow has authored or coauthored 16 books, including
America’s Research Universities:
The Challenges Ahead,
University Press of America (2011); A Guide to Lean Six Sigma, CRC
Press (2009);
Design for Six Sigma for Green Belts and Champions, Prentice-Hall, (2006); Six
Sigma for Green Belts and Champions,
Prentice-Hall, (2004); Quality Management: Tools
and Methods for Improvement,
3rd edition, Richard. D. Irwin (2004); Quality Management
Systems,
CRC Press (2000), Total Quality Management in Action, Prentice-Hall, (1994); The
Deming Guide to Quality and Competitive Position,
Prentice-Hall (1987); Planning for Quality, Productivity, and Competitive Position, Dow Jones-Irwin (1990); and Stat City: Understanding Statistics Through Realistic Applications, 2nd edition, Richard D. Irwin (1987). He
has published more than 60 academic articles in the areas of quality, statistics, management,
and marketing.
While at the University of Miami, Dr. Gitlow has received awards for outstanding teaching,
outstanding writing, and outstanding published research articles.
Richard J. Melnyck is Assistant Vice President for Medical Affairs and Executive Director
of Process Improvement at the University of Miami Miller School of Medicine and Health
System. He is a Six Sigma Master Black Belt, the University of Miami faculty advisor for the
American Society for Quality, the University of Miami Miller School of Medicine faculty
advisor for the Institute for Healthcare Improvement, and a member of the Beta Gamma
Sigma International Honor Society. Melnyck has taught process improvement in both the
School of Business and the Miller School of Medicine at the University of Miami. He has consulted on quality, productivity, and related matters with many organizations. He received his
MS in Management Science (2008), MBA (2002), and MS in Computer Information Systems
(2002) from the University of Miami.
From the Library of Pearson HED
xx About the Authors
David M. Levine is Professor Emeritus of Statistics and Computer Information Systems at
Baruch College (City University of New York). He received B.B.A. and M.B.A. degrees in
Statistics from City College of New York and a PhD from New York University in industrial
engineering and operations research. He is nationally recognized as a leading innovator
in statistics education and is the coauthor of 14 books, including such bestselling statistics
textbooks as
Statistics for Managers Using Microsoft Excel, Basic Business Statistics: Concepts
and Applications
, Business Statistics: A First Course, and Applied Statistics for Engineers and
Scientists Using Microsoft Excel and Minitab
. He also is the coauthor of Even You Can Learn
Statistics & Analytics: A Guide for Everyone Who Has Ever Been Afraid of Statistics
, currently
in its third edition, and
Design for Six Sigma for Green Belts and Champions, and the author of
Statistics for Six Sigma Green Belts, all published by Pearson, and Quality Management, third
edition, McGraw-Hill/Irwin. He is also the author of
Video Review of Statistics and Video
Review of Probability
, both published by Video Aided Instruction, and the statistics module
of the MBA primer published by Cengage Learning. He has published articles in various
journals, including
Psychometrika, The American Statistician, Communications in Statistics,
Decision Sciences Journal of Innovative Education, Multivariate Behavioral Research, Journal
of Systems Management
, Quality Progress, and The American Anthropologist, and he has
given numerous talks at the Decision Sciences Institute (DSI), American Statistical Association (ASA), and Making Statistics More Effective in Schools and Business (MSMESB) conferences. Levine has also received several awards for outstanding teaching and curriculum
development from Baruch College.
From the Library of Pearson HED
1
1
You Don’t Have to Suffer
from the Sunday Night Blues!
What Is the Objective of This Chapter?
We all know someone who dreads Sunday night because he or she isn’t looking forward
to going to work the next day. In fact, many of us know that person very well because that
person is us!
Many employees are highly respected and well paid, and you may believe that they are happy
with their jobs, but do not be fooled by their smiles. Many of them dislike their jobs. Many
people are “burned out” at work. So, if you are an employee just trying to do your job and
you think your job is boring, draining, and depressing, just think—you may have to do it for
the rest of your work life! How’s that for something to look forward to?
Well, we are here to tell you that you don’t have to suffer from the Sunday night blues!
Before we tell you what you can do to make that happen we need to first tell you a little bit
about intrinsic motivation. Intrinsic motivation comes from the sheer joy or pleasure of
performing an act, in this case such as improving a process or making your job better. It
releases human energy that can be focused into improvement and innovation of a system. As
amazing as it may seem, work does not have to be a drain on your energy. If you can release
the intrinsic motivation that lies within all of us it can actually fill you with energy so you
can enjoy what you do and look forward to doing it, day after day and year after year. Many
artists, athletes, musicians, and professors enjoy their work over the course of their lives. You
can enjoy your work also, or at least you can enjoy it much more than you currently do. It just
requires a redefinition of work and a management team that promotes the redefined view of
work to release the intrinsic motivation within each of us.
In today’s world, many of us are asked to self-manage to a great extent, meaning we are given
the autonomy and opportunity to direct our work to accomplish important organizational
objectives. However, many of us do not take advantage of that opportunity. Why? The reason is that we do not have the tools to release that intrinsic motivation to make our jobs, our
organizations, and most importantly our lives better. Now we do!
This book not only explains how it is possible for you to make both your work life and your
personal life better using process improvement and Six Sigma, but it gives you the tools and
methods to make it happen.
From the Library of Pearson HED
2 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Sarah’s Story
Most people go into work every day and are confronted with a long list of crises that require
immediate attention. Consider the story of Sarah who is an administrative assistant in a
department in a large, urban, private university. Please note that Sarah has not read this
book—yet. So she comes to work every day only to be greeted by a long to-do list of mini
crises that are boring and repetitive. Sound familiar?
The mini crises include answering the same old questions from faculty and students, week
after week after week:
What room is my class in?
Does the computer in room 312 work?
What are my professor’s office hours?
Are the copies I need for class (and requested only 5 minutes ago) ready? Blah,
blah, blah.
These crises prevent Sarah from doing her “real” work, which keeps piling up. It is frustrating
and depressing. If you ask Sarah what her job is, she will say: “I do whatever has to be done
to get through the day without a major disaster.”
No one is telling Sarah she cannot improve her processes so that she doesn’t have to answer
the same questions over and over again. In fact, her bosses would rather her not focus on
answering the same old questions and instead prefer her to work on projects that actually
add value. The problem is not that she doesn’t want to improve her processes; the problem
is that she doesn’t know how.
Then one day somehow the stars align and Sarah finds a copy of our book on her desk, so she
reads it. She starts to apply some of the principles of the book to her job and to her life, and
guess what? Things begin to change for the better.
For example, instead of having people call her to see what room their class is in she employs
something that she learns in the book called
change concepts, which are approaches to change
that have been found to be useful in developing solutions that lead to improvements in processes. In this case, she uses a change concept related to automation and sends out a daily
autogenerated email to all students and staff to let them know what room their classes are
located in. Utilizing the change concept eliminates the annoying calls she used to receive to
see what room classes are in.
Can you identify with Sarah? Do you want to learn tools and methods that will help you
transform your job, your organization, and your life? The upcoming chapters take you on
that journey, the journey of process improvement.
Before we go through the structure of the book, it is important for you to understand some
key fundamental principles. These are principles that you need to understand as a prerequisite to reading this book and are principles you need to keep referring back to if you want to
From the Library of Pearson HED
Chapter 1 You Don’t Have to Suffer from the Sunday Night Blues! 3
transform your job (to the extent management allows you to do it), your organization (if it
is under your control), and your life through process improvement.
A young violinist in New York City asks a stranger on the street how to get to Carnegie
Hall; the stranger’s reply is, “Practice, practice, practice.” The same thing applies to process
improvement. The only way you get better at it is through practice, practice, practice, and it
starts with the nine principles outlined in this chapter.
Nine Principles of Process Improvement to Get the Most
Out of This Book
Process improvement and Six Sigma embrace many principles, the most important of which
in our opinion are discussed in this section. When understood, these principles may cause a
transformation in how you view life in general and work in particular (Gitlow, 2001; Gitlow,
2009).
The principles are as follows:
Principle 1—Life is a process (a process orientation).
Principle 2—All processes exhibit variation.
Principle 3—Two causes of variation exist in all processes.
Principle 4—Life in stable and unstable processes is different.
Principle 5—Continuous improvement is always economical, absent capital
investment.
Principle 6—Many processes exhibit waste.
Principle 7—Effective communication requires operational definitions.
Principle 8—Expansion of knowledge requires theory.
Principle 9—Planning requires stability. Plans are built on assumptions.
These principles are outlined in the following sections and appear numerous times throughout the book. Illustrated from the point of view of everyday life, it is your challenge to apply
them to yourself, your job, and your organization.
Principle 1: Life is a process. A process is a collection of interacting components that transform inputs into outputs toward a common aim called a mission statement. Processes exist
in all facets of life in general, and organizations in particular, and an understanding of them
is crucial.
The transformation accomplished by a process is illustrated in Figure 1.1 . It involves the
addition or creation of time, place, or form value. An output of a process has
time value if
it is available when needed by a user. For example, you have food when you are hungry, or
equipment and tools available when you need them. An output has
place value if it is available where needed by a user. For example, gas is in your tank (not in an oil field), or wood
From the Library of Pearson HED
4 A Guide to Six Sigma and Process Improvement for Practitioners and Students
chips are in a paper mill. An output has form value if it is available in the form needed by a
user. For example, bread is sliced so it can fit in a toaster, or paper has three holes so it can
be placed in a binder.

order now
Inputs
Manpower/services
Equipment
Materials/goods
Methods
Environment

 

Process
Transformation
of inputs into
output by adding
time, form, or
place value

Outputs
Manpower/services
Equipment
Materials/goods
Methods
Environment
Figure 1.1 Basic process
An example of a personal process is Ralph’s “relationship with women he dates” process.
Ralph is 55 years old. He is healthy, financially stable, humorous, good looking (at least he
thinks so!), and pleasant. At age 45 he was not happy because he had never had a long-term
relationship with a woman. He wanted to be married and have children. Ralph realized that
he had been looking for a wife for 20 years, with a predictable pattern of four to six month
relationships—that is, two relationships per year on average; see Figure 1.2 . That meant he
had about 40 relationships over the 20 years.

Use an e‐dating service,
go to social gatherings,
or get fixed up on a blind
date.

 

Get depressed over
breakup. Obsess about
reason for breakup.

 

Start dating a woman.
Continue dating one
special woman. Break up
with the special woman.

 

Figure 1.2 Ralph’s relationship with women process
Ralph continued living the process shown in Figure 1.2 for more than 20 years. It depressed
and frustrated him, but he did not know what to do about it. Read on to the next principles
to find out more about Ralph’s situation.
From the Library of Pearson HED
Chapter 1 You Don’t Have to Suffer from the Sunday Night Blues! 5
Principle 2: All processes exhibit variation. Variation exists between people, outputs, services, products, and processes. It is natural and should be expected, but it must be reduced.
The type of variation being discussed here is the unit-to-unit variation in the outputs of
a process (products or services) that cause problems down the production or service line
and for customers. It is
not diversity, for example, racial, ethnic, or religious, to name a few
sources of diversity. Diversity makes an organization stronger due to the multiple points of
view it brings to the decision making process.
Let’s go back to our discussion of unit-to-unit variation in the outputs of a process. The critical question to be addressed is: “What can be learned from the unit-to-unit variation in the
outputs of a process (products or services) to reduce it?” Less variability in outputs creates a
situation in which it is easier to plan, forecast, and budget resources. This makes everyone’s
life easier.
Let’s get back to Ralph’s love life or lack thereof. Ralph remembered the reasons for about 30
of his 40 breakups with women. He made a list with the reason for each one. Then he drew a
line graph of the number of breakups by year; see Figure 1.3 .
Year
Number of Breakups
1981 1983 1985 1987 1989 1991 1993 1995 1997 1999
3.0
2.5
2.0
1.5
1.0
0.5
0.0 0
Time Series Plot of Number of Breakups
Figure 1.3 Number of breakups by year
As you can see, the actual number of breakups varies from year to year. Ralph’s ideal number
of breakups per year is zero; this assumes he is happy and in a long-term relationship with
a woman whom he has children with. The difference between the actual number of breakups and the ideal number of breakups is unwanted variation. Process improvement and
Six Sigma management help you understand the causes of unwanted waste and variation,
From the Library of Pearson HED
6 A Guide to Six Sigma and Process Improvement for Practitioners and Students
thereby giving you the insight you need to bring the actual output of a process and the ideal
output of a process closer to each other.
Another example: Your weight varies from day to day. Your
ideal daily weight would be
some medically determined optimum level; see the black dots on Figure 1.4 . Your
actual
daily weights may be something entirely different. You may have an unacceptably high average weight with great fluctuation around the average; see the fluctuating squares on Figure
1.4 . Unwanted variation is the difference between your ideal weight and your actual weights.
Process improvement and Six Sigma management help you understand the causes of this
variation, thereby giving you the insight you need to bring your actual weight closer to your
ideal weight.
Index
Data
2 4 6 8 10 12 14 16 18 20 22 24
200
190
180
170
160
Variable
Ideal
Actual
Actual vs. Ideal Weights by Day
Figure 1.4 Actual versus ideal weights by day
Principle 3: Two causes of variation exist in all processes; they are special causes and common causes of variation. Special causes of variation are due to assignable causes external
to the process. Common causes of variation are due to the process itself—that is, variation
caused by the structure of the process. Examples of common causes of variation could be
stress, values and beliefs, or the level of communication between the members of a family.
Usually, most of the variation in a process is due to common causes. A process that exhibits
special and common causes of variation is unstable; its output is not predictable in the future.
A process that exhibits only common causes of variation is stable (although possibly unacceptable); its output is predictable in the near future.
Let’s visit Ralph again. Ralph learned about common and special causes of variation and
began to use some basic statistical thinking and tools to determine whether his pattern of
From the Library of Pearson HED
Chapter 1 You Don’t Have to Suffer from the Sunday Night Blues! 7
breakups with women was a predictable system of common causes of variation. Ralph constructed a control chart (see Figure 1.5 ) of the number of breakups with women by year. After
thinking about himself from a statistical point of view using a control chart, he realized his
relationships with women were not unique events (special causes); rather, they were a common cause process (his relationship with women process).
Year
Sample Count
1981 1983 1985 1987 1989 1991 1993 1995 1997 1999
5 4 3 2 1 0
_
C=1.5
UCL=5.174
LCL=0
C Chart of Number of Breakups
Figure 1.5 Number of breakups with women by year
Control charts are statistical tools used to distinguish special from common causes of variation. All control charts have a common structure. As Figure 1.5 shows, they have a center
line, representing the process average, and upper and lower control limits that provide information on the process variation. Control charts are usually constructed by drawing samples
from a process and taking measurements of a process characteristic, usually over time. Each
set of measurements is called a
subgroup, for example, a day or month. In general, the center
line of a control chart is taken to be the estimated mean of the process; the upper control limit
(UCL) is a statistical signal that indicates any point(s) above it are likely due to special causes
of variation, and the lower control limit (LCL) is a statistical signal that indicates any point(s)
below it are likely due to special causes of variation. Additional signals of special causes of
variation are not discussed in this chapter, but are discussed later in the book.
Back to Ralph’s love life; Figure 1.5 shows that the number of breakups by year are all between
the UCL = 5.174 and the LCL = 0.0. So, Ralph’s breakup process with women only exhibits common causes of variation; it is a stable and predictable process, at least into the near
future. This tells Ralph that he should analyze all 30 data points for all 20 years as being part
From the Library of Pearson HED
8 A Guide to Six Sigma and Process Improvement for Practitioners and Students
of his “relationship with women” process; he should not view any year or any relationship
as special.
Ralph was surprised to see that the reasons he listed for the 30 breakups collapsed down to
five basic categories, with one category containing 24 (80%) of the relationships. The categories (including repetitions) are grouped into the frequency distribution shown in Table 1.1 .
Table 1.1 Frequency Distribution of Reasons for Breakups with Women for 20 Years

Reason Frequency Percentage
Failure to commit 24 80.00
Physical 03 10.00
Sexual 01 3.33
Common interests 01 3.33
Other relationships 01 3.33
Total 30 100.00

Ralph realized that there were not 30 unique reasons (special causes) that moved him to
break up with women. He saw that there were only five basic reasons (common causes of
variation in his process) that contributed to his breaking up with women, and that “failure
to commit” is by far the most repetitive common cause category.
Principle 4: Life in stable and unstable processes is different. This is a big principle. If a
process is stable, understanding this principle allows you to realize that most of the crises that
bombard you on a daily basis are nothing more than the random noise (common causes of
variation) in your life. Reacting to a crisis like it is a special cause of variation (when it is in
fact a common cause of variation) will double or explode the variability of the process that
generated it. All common causes of variation (formerly viewed as crises) should be categorized to identify
80-20 rule categories, which can be eliminated from the process. Eliminating an 80-20 rule category eliminates all, or most, future repetition of the common causes
(repetitive crises) of variation generated by the problematic component of the process.
Let’s return to the example of Ralph. Ralph realized that the 30 women were not individually to blame (special causes) for the unsuccessful relationships, but rather, he was to blame
because he had not tended to his emotional well-being (common causes in his stable emotional process); refer to Figure 1.5 . Ralph realized he was the process owner of his emotional
process. Armed with this insight, he entered therapy and worked on resolving the biggest
common cause category (80-20 rule category) for his breaking up with women, failure to
commit.
The root cause issue for this category was that Ralph was not getting his needs met by the
women. This translated into the realization that his expectations were too high because he
had a needy personality. In therapy he resolved the issues in his life that caused him to be
needy and thereby made a fundamental change to himself (common causes in his emotional
From the Library of Pearson HED
Chapter 1 You Don’t Have to Suffer from the Sunday Night Blues! 9
process). He is now a happily married man with two lovely children. Ralph studied and
resolved the common causes of variation between his ideal and real self, and moved himself
to his ideal; see the right side of Figure 1.6 . He did this by recognizing that he was the process
owner of his emotional process and that his emotional process was stable, and required a
common cause type fix, not a special cause type fix. Ralph is the manager of his life; only he
can change how he interacts with the women he forms relationships with.
Year
Sample Count
1982 1985 1988 1991 1994 1997 2000 2003 2006
5 4 3 2 1 0
C Chart of Number of Breakups with Women by Before and After

Figure 1.6 Number of breakups with women before and after therapy
Principle 5: Continuous improvement is always economical, absent capital investment.
Continuous improvement is possible through the rigorous and relentless reduction of common causes of variation and waste around a desired level of performance in a stable process.
It is always economical to reduce variation around a desired level of performance,
without
capital investment
, even when a process is stable and operating within specification limits.
For example, elementary school policy states that students are to be dropped off at 7:30 a.m.
If a child arrives before 7:25 a.m., the teacher is not present and it is dangerous because it is
an unsupervised environment. If a child arrives between 7:25 a.m. and 7:35 a.m., the child
is on time. If a child arrives after 7:35 a.m., the entire class is disrupted. Consequently, parents think that if their child arrives anytime between 7:25 a.m. and 7:35 a.m. it is acceptable
(within specification limits). However, principle 5 promotes the belief that for every minute
a child is earlier or later than 7:30 a.m., even between 7:25 am and 7:35 am, a loss is incurred
by the class. The further from 7:30 a.m. a child arrives to school, the greater the loss. Please
note that the loss may not be symmetric around 7:30 a.m. Under this view, it is each parent’s
job to continuously reduce the variation in the child’s arrival time to school. This minimizes
From the Library of Pearson HED
10 A Guide to Six Sigma and Process Improvement for Practitioners and Students
the total loss to all stakeholders of the child’s classroom experience (the child, classmates,
teacher, and so on). Table 1.2 shows the loss incurred by the class of children in respect to
accidents from early arrivals of children and the disruptions by late arrivals of children for
a one year period.
Table 1.2 Loss from Minutes Early or Late

Arrival Times (a.m.) # Minutes Early or Late Loss to the Classroom
7:26 4 2 accidents
7:27 3 2 accidents
7:28 2 1 accident
7:29 1 1 accident
7:30 0 0 accidents
7:31 1 1 minor disruption
7:32 2 1 minor disruption
7:33 3 1 medium disruption
7:34 4 1 major disruption
Total 6 accidents
2 minor disruptions
1 medium disruption
1 major disruption

If parents can reduce the variation in their arrival time processes from the distribution in
Table 1.2 to the distribution in Table 1.3 , they can reduce the loss from early or late arrival
to school. Reduction in the arrival time process requires a fundamental change to parents’
arrival time behavior, for example, laying out their child’s clothes the night before to eliminate time. As you can see, Table 1.2 shows 6 accidents, 2 minor disruptions, 1 medium
disruption, and 1 major disruption, while Table 1.3 shows 4 accidents, 2 minor disruptions,
and 1 medium disruption. This clearly demonstrates the benefit of continuous reduction of
variation, even if all units conform to specifications.
Table 1.3 Improved Loss from Minutes Early or Late

Arrival Times (a.m.) # Minutes Early or Late Loss to the Classroom
7:26 4 0 accidents
7:27 3 2 accidents
7:28 2 1 accident
7:29 1 1 accident

From the Library of Pearson HED
Chapter 1 You Don’t Have to Suffer from the Sunday Night Blues! 11

Arrival Times (a.m.) # Minutes Early or Late Loss to the Classroom
7:30 0 0 accidents
7:31 1 1 minor disruption
7:32 2 1 minor disruption
7:33 3 1 medium disruption
7:34 4 0 disruptions
Total 4 accidents
2 minor disruptions
1 medium disruption

Principle 6: Many processes exhibit waste. Processes contain both value added activities
and non-value added activities. Non-value added activities in a process include any wasteful
step that
Customers are not willing to pay for
Does not change the product or service
Contains errors, defects, or omissions
Requires preparation or setup
Involves control or inspection
Involves overproduction, special processing, and inventory
Involves waiting and delays
Value added activities include steps that customers are willing to pay for because they positively change the product or service in the view of the customer. Process improvement and
Six Sigma management promote reducing waste through the elimination of non-value added
activities (streamlining operations), eliminating work in process and inventory, and increasing productive flexibility and speed of employees and equipment.
Recall Ralph and his love life dilemma. If you consider Ralph’s failure to commit as part of
his relationship with women process, you can clearly see that it is a non-value added activity.
This non-value added activity involves some wasteful elements. First, the women Ralph dates
do not want to spend their valuable time dating a man who cannot commit to a long-term
relationship. Second, the women ultimately feel tricked or lied to because Ralph failed to discuss his commitment issues early in the relationship. Third, the women resent the emotional
baggage (unwanted inventory) that Ralph brings to the prospective relationship. Clearly,
Ralph needed to eliminate these forms of waste from his love life.
Principle 7: Effective communication requires operational definitions. An operational
definition promotes effective communication between people by putting communicable
meaning into a word or term. Problems can arise from the lack of an operational definition
From the Library of Pearson HED
12 A Guide to Six Sigma and Process Improvement for Practitioners and Students
such as endless bickering and ill will. A definition is operational if all relevant users of the
definition agree on the definition. It is useful to illustrate the confusion that can be caused
by the absence of operational definitions. The label on a shirt reads “75% cotton.” What
does this mean? Three quarters cotton on average over this shirt, or three quarters cotton
over a month’s production? What is three quarters cotton? Three quarters by weight? Three
quarters at what humidity? Three quarters by what method of chemical analysis? How many
analyses? Does 75% cotton mean that there must be some cotton in any random cross-section
the size of a silver dollar? If so, how many cuts should be tested? How do you select them?
What criterion must the average satisfy? And how much variation between cuts is permissible? Obviously, the meaning of 75% cotton must be stated in operational terms; otherwise
confusion results.
An operational definition consists of
A criterion to be applied to an object or a group
A test of the object or group in respect to the criterion
A decision as to whether the object or group did or did not meet the criterion
The three components of an operational definition are best understood through an example.
Susan lends Mary her coat for a vacation. Susan requests that it be returned clean. Mary
returns it dirty. Is there a problem? Yes! What is it? Susan and Mary failed to operationally
define clean. They have different definitions of clean. Failing to operationally define terms
can lead to problems. A possible operational definition of clean is that Mary will get the coat
dry-cleaned before returning it to Susan. This is an acceptable definition if both parties agree.
This operational definition is shown here:
Criteria: The coat is dry-cleaned and returned to Susan.
Test: Susan determines if the coat was dry-cleaned.
Decision: If the coat was dry-cleaned, Susan accepts the coat. If the coat was not drycleaned, Susan does not accept the coat.
From past experience, Susan knows that coats get stained on vacation and that dry cleaning
may not be able to remove a stain. Consequently, the preceding operational definition is
not acceptable to Susan. Mary thinks dry cleaning is sufficient to clean a coat and feels the
preceding operational definition is acceptable. Since Susan and Mary cannot agree on the
meaning of clean, Susan should not lend Mary the coat.
An operational definition of clean that is acceptable to Susan follows:
Criteria: The coat is returned. The dry-cleaned coat is clean to Susan’s satisfaction or
Mary must replace the coat, no questions asked.
Test: Susan examines the dry-cleaned coat.
Decision: Susan states the coat is clean and accepts the coat. Or, Susan states the coat
is not clean and Mary must replace the coat, no questions asked.
From the Library of Pearson HED
Chapter 1 You Don’t Have to Suffer from the Sunday Night Blues! 13
Mary doesn’t find this definition of clean acceptable. The moral is: Don’t do business with
people without operationally defining critical quality characteristics.
Operational definitions are not trivial. Statistical methods become useless tools in the absence
of operational definitions because data does not mean the same thing to all its users.
Principle 8: Expansion of knowledge requires theory. Knowledge is expanded through revision and extension of theory based on systematic comparisons of predictions with observations. If predictions and observations agree, the theory gains credibility. If predictions and
observations disagree, the variations (special and/or common) between the two are studied,
and the theory is modified or abandoned. Expansion of knowledge (learning) continues
forever.
Let’s visit Ralph again. He had a theory that each breakup had its own and unique special
cause. He thought deeply about each breakup and made changes to his behavior based on his
conclusions. Over time, Ralph saw no improvement in his relationships with women; that
is, the difference between the actual number of breakups by year was not getting any closer
to zero; that is a long-term relationship. Coincidently, he studied process improvement and
Six Sigma management and learned that there are two types of variation in a process, special
and common causes. He used a control chart to study the number of breakups with women
by year; refer to the left side of Figure 1.6 . Ralph developed a new theory for his relationship
with women process based on his process improvement and Six Sigma studies. The new
theory recognized that all Ralph’s breakups were due to common causes of variation. He
categorized them, went into therapy to deal with the biggest common cause problem, and
subsequently, the actual number of breakups with women by year equaled the ideal number
of breakups with women by year; refer to the right side of Figure 1.6 . Ralph tested his new
theory by comparing actual and ideal numbers, and found his new theory to be helpful in
improving his relationship with women process.
Principle 9: Planning requires stability. Plans are built on assumptions. Assumptions are
predictions concerning the future conditions, behavior, and performance of people, procedures, equipment, or materials of the processes required by the plan. The predictions have
a higher likelihood of being realized if the processes are stable with low degrees of variation.
If you can stabilize and reduce the variation in the processes involved with the plan, you
can affect the assumptions required for the plan. Hence, you can increase the likelihood of
a successful plan.
Example: Jan was turning 40 years old. Her husband wanted to make her birthday special.
He recalled that when Jan was a little girl she dreamed of being a princess. So, he looked for
a castle that resembled the castle in her childhood dreams. After much searching, he found a
castle in the middle of France that met all the required specifications. It had a moat, parapets,
and six bedrooms; perfect. Next, he invited Jan’s closest friends, three couples and two single
friends, filling all six bedrooms. After much discussion with the people involved, he settled
on a particular three day period in July and signed a contract with the count and countess
who owned the castle. Finally, he had a plan and he was happy.
As the date for the party drew near, he realized that his plan was based on two assumptions.
The first assumption was that the castle would be available. This was not a problem because
From the Library of Pearson HED
14 A Guide to Six Sigma and Process Improvement for Practitioners and Students
he had a contract. The second assumption was that all the guests would be able to go to the
party. Essentially, each guest’s life is a process. The question is: Is each “guest’s life process”
stable with a low enough degree of variation to be able to predict attendance at the party. This
turned out to be a substantial problem. Due to various situations, several of the guests were
not able to attend the party. One couple began to have severe marital problems. One member
of another couple lost his job. Jan’s husband should have realized that the likelihood of his
second assumption being realized was problematic and subject to chance; that is, he would
be lucky if all the guests were okay at the time of the party. He found out too late that the
second assumption was not met at the time of the party. If he had he realized this, he could
have saved money and heartache by renting rooms that could be cancelled in a small castletype hotel. As a postscript, the party was a great success!
Structure of the Book
We structured the book strategically into five main sections, each building upon each other
and each expanding your knowledge so that eventually you can complete a process improvement project on your own.
We use the analogy of building a house in how we structured this book.
Section I —Building a Foundation of Process Improvement Fundamentals
Chapter 1 —You Don’t Have to Suffer from the Sunday Night Blues!
Chapter 2 —Process and Quality Fundamentals
Chapter 3 —Defining and Documenting a Process
One of the first steps to building a house is to lay down a foundation. The first section creates your foundation in process improvement by taking you through the process and quality
fundamentals you need as you build up your knowledge base. It goes into further detail on
many of our nine principles for process improvement, principles critical to your understanding of this material.
Section II —Creating Your Toolbox for Process Improvement
Chapter 4 —Understanding Data: Tools and Methods
Chapter 5 —Understanding Variation: Tools and Methods
Chapter 6 —Non-Quantitative Techniques: Tools and Methods
Chapter 7 —Overview of Process Improvement Methodologies
Chapter 8 —Project Identification and Prioritization: Building a Project Pipeline
You cannot build a house without tools and without understanding how and when to use
them, right? The second section creates your toolbox for process improvement by not only
teaching you the tools and methods you need to improve your processes but teaching you
when and how to use them.
From the Library of Pearson HED
Chapter 1 You Don’t Have to Suffer from the Sunday Night Blues! 15
Section III —Putting It All Together—Six Sigma Projects
Chapter 9 —Overview of Six Sigma Management
Chapter 10 —DMAIC Model: “D” Is for Define
Chapter 11 —DMAIC Model: “M” Is for Measure
Chapter 12 —DMAIC Model: “A” Is for Analyze
Chapter 13 —DMAIC Model: “I” Is for Improve
Chapter 14 —DMAIC Model: “C” Is for Control
Chapter 15 —Maintaining Improvements in Processes, Products-Services, Policies,
and Management Style
When you build a house you need a framework or guide to follow to make sure you build the
house correctly; it’s called a blueprint! Once that beautiful house is built you need to maintain
it so it stays beautiful, right?
The third section is analogous to the blueprint of a house, and it is where we put everything
you have learned together to complete a project. We use a specific set of steps—kind of like
a blueprint—to keep us focused and make sure we do the project correctly. Those steps
are called the Six Sigma management style. Like the maintenance of a new house, once we
improve the process, the last thing we want is for the process to backslide to its former problematic state. We show you how to maintain and sustain those improvements.
Section IV —The Culture Required for Six Sigma Management
Chapter 16 —W. Edwards Deming’s Theory of Management: A Model for Cultural
Transformation of an Organization
The fourth section of this book discusses an appropriate culture for a successful Six Sigma
management style. We can use the house building analogy because a house has to be built
on a piece of property that can support all its engineering, social, psychological, and so on
needs and wants. Without a proper piece of property, the house could fall into a sinkhole.
Section V —Six Sigma Certification
Chapter 17 —Six Sigma Champion Certification (online-only chapter)
Chapter 18 —Six Sigma Green Belt Certification (online-only chapter)
The fifth section discusses how you can become Six Sigma certified at the Champion and
Green Belt levels of certification. Certification is like getting your house a final inspection and
receiving a Certificate of Occupancy so you can move in. (This section can be found online
at www.ftpress.com/sixsigma.)
From the Library of Pearson HED
16 A Guide to Six Sigma and Process Improvement for Practitioners and Students
We hope you enjoy this book. Feel free to contact the authors concerning any mistakes you
have found, or any ideas for improvement. Thank you for reading our book. We hope you
find it an invaluable asset on your journey toward a Six Sigma management culture.
Howard S. Gitlow, PhD
Professor
Six Sigma Master Black Belt
Department of Management Science
University of Miami
[email protected]
Richard J. Melnyck, MBA, MS in MAS, and MAS in CIS
Six Sigma Master Black Belt
Assistant Vice President for Medical Affairs
Executive Director of Process Improvement
Office of the Senior Vice President for Medical Affairs and Dean
University of Miami Miller School of Medicine
[email protected]
David M. Levine, PhD
Professor Emeritus
Department of Statistics and Computer Information Systems
Baruch College
City University of New York
[email protected]
Let’s Go!
We are excited to begin this journey with you—the journey of process improvement that we
hope transforms your job and more importantly your life! While this is a technical book, we
want to make it fun and interesting so that you will remember more of what we are teaching you. We tried to add humor and stories to make the journey a fun one. So what are we
waiting for? Let’s go!
References
Gitlow, H. (2009), A Guide to Lean Six Sigma Management Skills (New York: CRC Press).
Gitlow, H., “Viewing Statistics from a Quality Control Perspective,”
International Journal of
Quality and Reliability Management,
vol.18, issue 2, 2001.
From the Library of Pearson HED
17
2
Process and Quality Fundamentals
What Is the Objective of This Chapter?
The objective of this chapter is to teach you the meaning of (1) a process or system, (2) variation in a process or system, and (3) quality of a process, product, service, policy, procedure,
and so on. These concepts prepare you for what is to come in the rest of the book. Before
you can improve a process, it is crucial that you understand the building blocks of process,
variation in a process, and the definitions of quality and their significance.
Now That’s an Interesting Process!

A man walks into a bar and engages in an interesting process. He orders three beers,
takes a drink out of the first one and sets it down, takes a drink out of the second one
and sets it down, takes a drink out of the third one and sets it down. He comes in
multiple times per week and performs the same process for a month before the bar
tender gets curious and asks him, “I can’t help but notice your odd process of order
ing three beers, taking a drink out of each, and then leaving. What’s the deal?” The
man responds, “Growing up, my two brothers and I were very close and we used to
drink together all the time. When we all got married and moved far away from each
other we made a pact that we would each do this to remember the good ol’ times.”
The man repeated the same process for another month and then all of a sudden he
changed his process and ordered only two beers. Again the bartender got curious
and asked, “I don’t mean to pry, but I cannot help but notice you have changed your
process and are only ordering two beers. Did one of your brothers die?” The man
responded with a smile, “No, no, they are both alive and well. The reason I changed
my process was because my drinking was affecting my marriage, and my wife said I
couldn’t drink anymore. But she didn’t say anything about my brothers!”
Moral of the story: When all else fails, just blame it on the process!

From the Library of Pearson HED
18 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Process Fundamentals
Process fundamentals include the following topics: (1) What is a process? (2) Where do
processes exist? (3) Why does understanding processes matter? (4) What is a feedback loop,
and how does it fit into the idea of a process? These questions need to be answered to begin
to perform process improvement activities.
What Is a Process?
A process is a collection of interacting components that transform inputs (elements that the
process needs to execute) into
outputs (the results of the process) toward a common aim,
called a mission statement (Gitlow et al., 2015). A mission statement for a hospital could be
“to be a world class provider of medical services to the community.” A mission statement
for an individual, in this case for me, is “to generate positive energy into the universe.” For
example, this is helpful when I am driving. I am an impatient driver and get frustrated when
someone is in my way. My mission statement directs me to just relax and take it easy; getting
to my destination 2 minutes later isn’t worth the frustration. Figure 2.1 shows how a process
transforms inputs into outputs.

Inputs
Manpower/services
Equipment
Materials/goods
Methods/policies
Environment
And so on…

 

Process
Transformation
of inputs into
outputs

Outputs
Manpower/services
Equipment
Materials/goods
Methods/policies
Environment
And so on…
Figure 2.1 Transforming inputs into outputs
Where Do Processes Exist?
Now that we know what a process is, the next question is where do processes exist (Gitlow et
al., 2015). Processes exist in all facets of organizations, as well as in everyday life. An understanding of processes is crucial if you want to improve the quality of your organization and/
or the quality of your life. Remember, work and life processes are not mutually exclusive,
so feel free to improve them both at the same time. Many of us mistakenly think only of
production processes. However, if you think about it processes are everywhere; administration, sales, service, human resources, training, maintenance, paper flows, interdepartmental communication, and vendor relations are all processes you see at work. Importantly,
From the Library of Pearson HED
Chapter 2 Process and Quality Fundamentals 19
relationships between people are also processes. How about your daily life? From when you
wake up in the morning to when you leave for school or work is a process, cleaning your
house is a process (not exactly a fun one), and making plans with your friends is a process
(depending on your friends it can be a complicated one!).
It is critical for top management of an organization to have a process-oriented view of their
organization and life if they are to be successful with the style of management presented in
this book.
Why Does Understanding Processes Matter?
Most processes can be studied, documented, defined, standardized, improved, and innovated. Any situation in your life or work, where an input is turned into an output, involves
a process. Consequently, acquiring the tools to improve processes makes your life and work
a whole lot better!
What Is a Feedback Loop and How Does It Fit into the Idea
of a Process?
A feedback loop is the part of a system that takes the actual outputs (data in one form or
another) of a step in a process and feeds them back/forward/sideways to another step in the
process to determine whether the desired outputs were generated by the process step (Gitlow
et al., 2015). The purpose of a feedback loop is to use the data from it to close the gap between
what was desired of the process output and what outputs were actually delivered by the process. In layman’s terms, did you get what you wanted out of the process? If not, the feedback
loop provides data to help you improve the process. As a result of the process improvement,
the next time there is a better chance you will get the desired process output.
Some Process Examples to Bring It All Together!
Let’s review some examples that will help you understand the process concepts covered so
far, namely, inputs, outputs, feedback loops, and aims/missions.
Process Example #1—The Background Check Process
An example of an important human resources process is the background check that is usually
the last major process completed before hiring a new employee. The hiring manager gives
human resources the name of the person she wants to hire and then human resources begins
the background check process. The
aim of this process is to make sure the new employee
does not have any showstopper skeletons in his closet that puts the organization at risk.
Figure 2.2 shows the inputs, process, and outputs of performing a background check on an
employee. The inputs include the candidate’s hiring application, the candidate’s consent
form for the background check, and the recommendation to hire by the hiring manager. The
process involved in this transformation of inputs into outputs includes reviewing criminal
From the Library of Pearson HED
20 A Guide to Six Sigma and Process Improvement for Practitioners and Students
records, reviewing educational records, reviewing driving records, and so on. The output in
this example is the completed background check report. An important aspect of this process
is the
feedback loop that enables the human resources manager to report back to the hiring
manager on the employee’s appropriateness for a given job.

Inputs
Hiring application
Consent form
Recommendation to
hire

 

Process
Review criminal
records
Review educational
records
Review driving
records

Outputs
Background check
report

Feedback Loop
This person has lied about his education
and has been in jail 10 times! Not only can
we not hire him, we need to look at our
screening processes.

Figure 2.2 Background check process
Process Example #2—Leonard’s Head Searching Process
The execution of different types of diagnostic testing can also be thought of as processes.
Consider the case of Leonard who goes to the doctor. The doctor senses something is really
“off” with Leonard, so his
aim is to figure out what it is. The inputs to the process are the doctor checking Leonard, looking at his past medical records, and ordering an MRI of Leonard’s
head. The
process involves Leonard being escorted to the Radiology department, the technologist prepping and positioning Leonard, the technologist telling him to hold his breath
and then taking his picture. Unfortunately, the technologist does not tell Leonard to stop
holding his breath, so his face starts to turn blue and he almost passes out. Luckily, the technologist is on her game so she immediately tells him to breathe and he quickly recovers. The
radiologist then reads the report and the
output involves the report being sent to the doctor.
As for the
feedback loop, it seems like the doctor’s hunch was correct as to what is going on
in Leonard’s head—nothing! Since there is nothing going on in Leonard’s head, the doctor
sends Leonard to see a psychiatrist.
Figure 2.3 depicts the process of trying to figure out what is going on in Leonard’s head.
From the Library of Pearson HED
Chapter 2 Process and Quality Fundamentals 21

Inputs
Doctor checks
Leonard.
Doctor orders x‐ray.

 

Process
Leonard is escorted to Radiology
department.
Technologist preps Leonard.
Technologist positions Leonard.
Technologist moves camera into
place.
Leonard holds breath.
X-ray taken.
Technologist forgets to tell
Leonard to stop holding his breath.
Leonard turns blue and almost
passes out.
Technologist tells him to stop
holding his breath.
Radiologist reviews x-ray and
sends report to doctor.

Outputs
• Report from
radiologist
• Referral to a
psychiatrist

Feedback Loop
Doctor gets the report to
see what is going on in
Leonard’s head. All is
okay. Doctor sends
Leonard to see a
psychiatrist.

Figure 2.3 Leonard’s head searching process
Process Example #3—Patient’s Day of Surgery Process
Another example of a process is the presurgery process that a patient goes through the day
of surgery. The
aim is to ensure that the patient is ready to proceed with the surgery as
planned. Figure 2.4 shows the inputs, process, outputs, and feedback loop of the patient’s day
of surgery process. The
inputs include the patient’s surgery information packet, the blood
bank band (if required), medications currently taken and pre-op testing clearance form. The
process involved in this transformation of inputs into outputs includes being driven to the
hospital by a responsible adult, waiting, filling out necessary paperwork, waiting, filling out
more paperwork, waiting, receiving ID bracelet with your name misspelled, waiting, receiving new ID bracelet with your name now spelled correctly, waiting, getting in awkward looking gown, waiting, giving your belongings to a nurse, praying you will see your belongings
From the Library of Pearson HED
22 A Guide to Six Sigma and Process Improvement for Practitioners and Students
again, waiting, speaking to anesthesiologist, waiting, going to pre-op area and confirming
with staff the procedure you are having with them, waiting. The
output in this example would
be the patient ready for surgery. An important aspect of this process is the
feedback loop that
enables the pre-op area to give feedback to staff upstream in the process if the patient is not
ready for surgery.

Inputs
Surgery info packet
Blood bank band
Medications taken
Pre‐op clearance form

 

Process
Driven to hospital by responsible
adult
Waiting
Filling out paperwork
Waiting
Filling out more paperwork
Waiting
Receiving ID bracelet with name
misspelled
Waiting
Receiving new ID bracelet with
name spelled correctly
Putting on awkward gown
Waiting
Giving belongings to nurse
Praying you will see them again
Waiting
Speaking to anesthesiologist
Waiting
Going to pre‐op area
Waiting
Confirming procedure with pre‐op
staff

 

Outputs
Patient ready for
surgery

 

Feedback Loop
If patient is not ready,
why? How can the
process be improved
so surgery doesn’t
need to be cancelled?

Figure 2.4 Patient’s day of surgery process
From the Library of Pearson HED
Chapter 2 Process and Quality Fundamentals 23
An Anecdote about No Feedback Loops

John and Mary have a troubled marriage, and they have had it for 40 years. They
never talk to each other; in other words, they have no feedback loops in their mar
riage to use to possibly make their lousy situation better. Over time, their relation
ship continues to deteriorate; now they have a loveless marriage highlighted by angry
silence.

An Anecdote about Only Common Cause Feedback Loops

I grew up in a two-story house. My bedroom was on the second floor. When I was
about five years old, one night I snuck downstairs to see what my parents were doing.
Lucky for me they weren’t making me a baby sister, so I missed that trauma. Well,
they were having a conversation about my father’s upcoming trip to China. He said
to my mother: “I have to get a Gama globulin shot and the needle is a foot long.” At
that point two things stuck in my mind. First, I heard a really big word and wanted
to brag to my friends that I knew a big word like Gama globulin, even though I had
no idea what it meant. Second, I learned that needles existed that were a foot long,
and I hoped I would never see one.
Fast forward 50 years, and I am booked to go on a trip to China. My physician tells
me that I need a Gama globulin shot. As amazing as it may seem, I actually remem
bered that the needle was a foot long. Dear reader, please understand that this foot
long fact was hard-wired into my brain. I didn’t think about it rationally; I just
unconsciously accepted it.
Needless to say, I was sweating bullets when I went to get the shot. I assumed the
position for the shot: arms on the examining table with my pants and underwear
around my ankles, butt pushed out and ready for the railroad spike to be inserted
into my butt cheek. In my opinion, the nurse was taking too long to give me the shot,
so I angrily said to her: “Give me the damn shot.” She said, “I already did.” I called
her a liar because if someone was going to stick a railroad spike in my butt I would
be the first to know about it. She got offended and came over to me and flicked my
butt where she had given me the shot, and sure enough she wasn’t lying. Well, I went
from hating her to thinking she was Florence Nightingale; the best nurse and shot
giver ever born.
I left the office and sent an email to my physician, who also is the Chief of the Clinic,
saying that the nurse was the best shot giver in the world and they should all con
gratulate themselves on being world class!

From the Library of Pearson HED
24 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Variation Fundamentals
Variation fundamentals are critical to understanding how to improve a process. Understanding that there are two types of variation in a process, not one as is usually assumed, allows
process improvers to stabilize and improve a process, as opposed to over-reacting to process
variation and making the process worse that it was before. These concepts are explained in
the following sections.
What Is Variation in a Process?
As we saw in the previous examples, each process has one or many outputs, and each of these
outputs may be measured. The distribution of these measurements varies, and the differences
between these measurements are what we call
process variation.
The next part of the story is my fantasy of what happened after the email was sent. In
reality I have no idea what happened, but here is my fictional account of the transpiring events. Everyone in the office knew that I was an expert in process improvement,
so my email may have carried some weight. My doctor went out into the office and
read the email out loud to the entire office staff, congratulating them on jobs well
done.
The next bit of the story takes a dark turn; I imagined that another patient came in
and got a Gama globulin shot and moved while it was being inserted into her butt,
causing great pain. She yelled at the nurse and complained to my doctor, the Chief
of the Clinic, about the horrible and painful medical service delivered by the clinic.
At this point he assembled all the office staff and berated them for sloppy work. They
went from elated to depressed. The moral to this story is that if you always respond
to common causes of variation as if they were special causes of variation you are
micromanaging people, and it will drive them crazy and increase the variation in
their performance of their jobs.
An Anecdote about Your Daily Moods

Each day you wake up and your mood is slightly different, or maybe a lot different,
than it was the day before, and the day before that, and so on. Your daily moods
form a distribution of daily moods. Most of the time your mood is around your
average mood. The further your daily mood is from your average daily mood, the
less frequently this particular mood happens. Your daily moods may well form a
bell shaped distribution. If the distribution is skinny, your daily mood may well be
predictable into the near future with a high degree of belief. If the distribution is fat,
your daily mood may not be so predictable into the near future with any degree of

From the Library of Pearson HED
Chapter 2 Process and Quality Fundamentals 25
Why Does Variation Matter?
Variation matters because it is a fact of life. All processes exhibit variation, even if it is too
small to measure. When making a plan, you must consider whether the assumptions of your
plan will be in place when you execute the plan. Unfortunately, each assumption of your plan
is the output of some process, which exhibits variation.
For example, I plan to go to a 7:00 p.m. movie with my friend Neal, but he shows up at 7:25
p.m. Variation has ruined my movie plans. Variation matters; it needs to be reduced so prediction into the near future is possible with some degree of comfort.
What Are the Two Types of Variation?
The two types of variation (Deming, 1982, 1986, and 1994; Gitlow and Gitlow, 1987) are
common causes and special causes of variation.
Common variation is due to the design, management, policies, and procedures of the system
itself; this type of variation is the responsibility of management. An employee cannot change
the system he works in; only management can do that.
Special variation is external to the system; it disrupts the system from its routine generation
of common variation. Special variation is the responsibility of front line employees; however,
front line employees may need management’s help sometimes to deal with a special cause
of variation.
The outputs from all processes and their component parts vary over time. Your body has
a process that does a good job of generating and removing heat. The measurement used to
measure your body’s ability to do this is your body temperature. Despite large variation in
temperatures outside of you, your body does a great job of keeping your body temperature
within a safe and stable range. If your body is too hot your blood vessels expand to carry the
excess heat to the surface of your skin in the form of sweat, and when the sweat evaporates it
cools your body. If your body is too cold your blood vessels contract so that the blood flow
to your skin slows to conserve heat; your body may shiver in response to create more heat.
Under normal conditions both sweating and shivering help keep your body temperature
within safe and stable levels.
Normal body temperature is said to be 98.6 degrees Fahrenheit, so it can go as high as 99.6
degrees Fahrenheit and as low as 97.6 degrees Fahrenheit throughout the day depending on
belief. That’s life! Process improvement is largely about making your distribution of
daily moods skinny around a mean of high happiness. Remember, that 10% of the
time you will be in your bottom 10% of your moods; this is a mathematical fact. The
only way you can do anything about this is to shift your mood distribution further
in the direction of your happiness metric. Process improvement can help.
From the Library of Pearson HED
26 A Guide to Six Sigma and Process Improvement for Practitioners and Students
the time of day and how active you are. This slight variation in your body temperature occurs
all the time and is an example of common variation, or variation due to the system.
Now, let’s say you get sick and your body temperature jumps up to 101 degrees Fahrenheit.
This is a special cause of variation because it is caused by a change external to your body’s
heat management process—that is, getting sick.
If you had not gotten sick your body’s heat management process would have continued on its
former path of common variation. Note that if you get sick a lot, it might be a common cause
of variation. Later in the book we give you specific tools to understand how to distinguish
between special and common causes of variation.
The capability of a process is determined by inherent common causes of variation such as
poor hiring or training processes, inadequate work environments, poor information technology systems, or a lack of policies and procedures to name just a few. Front line employees
cannot change those types of processes, so they should not be held accountable for their
outcomes. Managers need to realize that unless changes are made to those processes, the
capability of the outcomes will remain the same. And only they can change the processes!
Special causes of variation on the other hand are due to events outside the system, such as a
natural disaster, problematic new supplies or equipment, or problems with a new IT system
to name just a few. Please note that special variation is
always detected by statistical signal,
not human logic.
Due to the fact that variation causes the customer to lose confidence in the reliability of the
dependability and uniformity of outputs, managers need to understand how to identify and
reduce variation. Using statistical methods, employees can identify different types of variation so that special causes of variation can be resolved (they can be good or bad) and common
variation can be reduced via process improvement projects resulting in more uniform and
reliable outputs.
Anecdote on Special Versus Common Causes of Variation

To demonstrate the difference between special and common causes of variation
let’s give another example. A friend of the authors was sitting at the kitchen table
on a Sunday morning practicing his penmanship, more specifically his little letter
“a.” He kept writing the little a over and over hoping he could perfect it when all of
a sudden his wife walked by and accidently hit the arm he was writing with, causing
a much larger tail on one of the a’s (see Figure 2.5 ). This example helps explain the
difference between special and common causes of variation. Most of the variation
in the size and shape of the a’s is due to common causes of variation as they are all
pretty close in size and shape, but when something external happens to the process
(his wife hitting his arm) you see a big difference in the size and shape of that letter a.
You are probably wondering why someone would spend time actually practicing one
letter; the only explanation we can come up with is that he is a Type A personality!

From the Library of Pearson HED
Chapter 2 Process and Quality Fundamentals 27
How to Demonstrate the Two Types of Variation
Dr. Deming used to conduct workshops to demonstrate special and common causes of variation; the workshops are called the Funnel Experiment and the Red Bead experiment.
The Funnel Experiment
Purpose and Introduction: The Funnel Experiment (Boardman and Iyer, 1986; Gitlow et al.,
2015) shows that adjusting (or tampering with) a stable process that is exhibiting only common causes of variation will increase the variation in the output of the process, depending
on how the process is adjusted.
In the experiment a marble is dropped through a funnel onto a piece of paper that has a point
that serves as the designated target. The objective is to get the marble as close to the target as
possible, the experiment uses various “rules” to attempt to minimize the distance between the
spread of the marble drops (results) and the position of the target. For each rule, the marble
is dropped 50 times, and its landing spot is marked on the piece of paper.
Special variation
(My wife hit my writing arm.)
Common variation in size
and shape of a’s
Largest a
Smallest a
Figure 2.5 Distinguishing between special and common causes of variation
From the Library of Pearson HED
28 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Rule #1:
36.0
24.0
12.0
0.0
-12.0
-24.0
-36.0
-44.0 -33.0 -22.0 -11.0 0.0 11.0 22.0 33.0 44.0
Y-POSITION
X-POSITION
Description: The marble is dropped, but the position
of the funnel does not change after each drop. This rule
serves as our baseline.
Results: A small circle of points emerges.
Takeaway: The size of the circle is due entirely to com
mon causes of variation in the process. This is analogous to
management understanding common and special variation
and knowing how to manage each without tampering with
the process.
Rule #2:
36.0
24.0
12.0
0.0
-12.0
-24.0
-36.0
-44.0 -33.0 -22.0 -11.0 0.0 11.0 22.0 33.0 44.0
Y-POSITION
X-POSITION
Description: We take the distance and direction away
from the target where the previous marble landed and
move the funnel to the equal and opposite direction for the
next drop.
Results: A circle of points emerges with double the varia
tion of the results in Rule #1.
Takeaway: The size of the circle is twice as large as Rule
#1, which is analogous to management tampering with the
process.
Real life example: The level of overtime at your hospital
was 12% over budget last month, so this month you make
sure to reduce overtime to be less than 12% of budget.
Rule #3
36.0
24.0
12.0
0.0
-12.0
-24.0
-36.0
-44.0 -33.0 -22.0 -11.0 0.0 11.0 22.0 33.0 44.0
Y-POSITION
X-POSITION
Description: We take the distance and direction of the
spot where the previous marble landed away from the tar
get and move the funnel first back to the target, and second
in an equal and opposite direction from the target for the
next drop.
Results: This rule produces an unstable, explosive pattern
of resting points on the surface.
Takeaway: This is analogous to management tampering
with the process causing variation to explode.
Real life example: Making up the previous month’s miss
to budget on inpatient admissions during the current
period.

From the Library of Pearson HED
Chapter 2 Process and Quality Fundamentals 29

Rule #4:
36.0
24.0
12.0
0.0
-12.0
-24.0
-36.0
-44.0 -33.0 -22.0 -11.0 0.0 11.0 22.0 33.0 44.0
Y-POSITION
X-POSITION
Description: We take the spot where the previous marble
landed and move the funnel to that spot for the next drop.
Results: This rule produces an unstable, explosive pattern
of resting points on the surface and it eventually moves
farther and farther away from the target in one direction
Takeaway: Again, this is analogous to management tam
pering with the process causing variation to explode.
Real life example: On the job training where worker
trains worker and the job skills deteriorate without
bounds.

Conclusion and takeaways: The Funnel Experiment illustrates how a system behaves when
it is tampered with.
Rules 2, 3, and 4 are analogous to management tampering with a process without
profound knowledge of how to improve the process through statistical thinking. This
increases the process’s variation and reduces the ability to manage that process.
Instead of using rules 2, 3, and 4, the better approach would be to create a control
chart of the output of rule 1. We would see that the process is stable and in control,
and instead of tampering the best course of action would be to improve the process.
Some ways to improve the process could be lowering the height of the funnel or using
a funnel with a smaller diameter.
An Anecdote about Rule 4 of the Funnel Experiment

On one occasion I was consulting for a factory and met an engineer who was work
ing 12- to 14-hour days, seven days a week, for many years. He was still married,
which surprised me. One day he asked me a statistics question that didn’t seem to have
anything to do with his job. I answered his question, but started to observe him so I
could understand his job better. I discovered that every time a manager senior to him
changed positions, which was relatively frequently given the number of years he was
on the job, the new manager would give Doug some new tasks to do. Doug was a “get
things done” kind of guy. As the years progressed his list of things to do increased to
the point of absurdity. Doug’s job is a perfect example of rule 4 of the Funnel Experi
ment; it kept expanding without bounds. He never asked his new manager if all the
data he was collecting and the reports he was writing were still desired by the his new
manager. Most likely, the new manager would have eliminated many of the tasks he
was doing for previous bosses, but he never asked and the manager never questioned
Doug about his workload. Doug was a wheel that didn’t squeak. This is a perfect exam
ple of rule 4 in operation.

From the Library of Pearson HED
30 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Red Bead Experiment
Purpose and Introduction: W. Edwards Deming used the Red Bead Experiment to show
the negative effects of treating common variation as special variation. It is discussed in this
section to further enhance your understanding of common causes and special causes of variation (Deming, 1994; Gitlow et al., 2015).
The experiment involves using a paddle to select beads from a box that contains 4,000 beads.
The box contains 3,200 white beads and 800 red beads. This fact is unknown to the participants in the experiment. Components of the experiment include a paddle with 50 bead
size depression, another box for mixing the beads, as well as a foreman in the Quality Bead
Company who hires, trains, and supervises four “willing workers” to produce white beads,
two inspectors of the willing workers’ output, one chief inspector to inspect the findings of
the inspectors, and one recorder of the chief inspector’s findings.
The job of the workers is to produce white beads, since red beads are unacceptable to customers. Strict procedures are to be followed. Work standards call for the production of 50
beads by each worker (a strict quota system): no more and no less than 50. Management has
established a standard that no more than three red beads (6%) per worker are permitted on
any given day. The paddle is dipped into the box of beads so that when it is removed, each
of the 50 holes contains a bead. Once this is done, the paddle is carried to each of the two
inspectors, who independently record the count of red beads. The chief inspector compares
their counts and announces the results to the recorder who writes down the number and
percentage of red beads next to the name of the worker.
The Results: On each day, some of the workers were above the average and some below the
average. On day 1, Sharyn did best with 7 red beads, but on day 2, Peter (who had the worst
record on day 1) was best with 6 red beads, and on day 3, Alyson was best with 6 red beads.
Table 2.1 displays the results for the four workers over three days.
Table 2.1 Red Bead Experiment Results for Four Workers Over Three Days

Day
Name 1 2 3 All 3 Days
Alyson 9 (18%) 11 (22%) 6 (12%) 26 (17.33%)
David 12 (24%) 12 (24%) 8 (16%) 32 (21.33%)
Peter 13 (26%) 6 (12%) 12 (24%) 31 (20.67%)
Sharyn 7 (14%) 9 (18%) 8 (16%) 24 (16.00%)
All 4 workers 41 38 34 113
Average (X–) 10.25 9.5 8.5 9.42
Proportion 20.5% 19% 17% 18.83%

From the Library of Pearson HED
Chapter 2 Process and Quality Fundamentals 31
Some Takeaways from the Red Bead Experiment:
1.
Common variation is an inherent part of any process. The variation between the
number of red beads by worker by day is only due to common variation.
2. Managers are responsible for the common variation in a system; they set the policies
and procedures. If managers are unhappy with the number of red beads, they should
take actions; for example, get a new supplier of white beads with less red bead per
load.
3. Workers are not responsible for the problems of the system—that is, common causes
of variation. The system primarily determines the performance of workers. The quota
of no more than three red beads per day per worker is insane. This is the case because
it is beyond the capability of this company to achieve that quota; 20% of the beads are
red, so in a load of 50 beads you would expect an average of 10 red beads per worker,
per day.
4. Some workers will always be above the average, and some workers will always be
below the average.
5. Some workers will always be in the bottom 10% because all distributions have a bottom 10% of units.
Quality Fundamentals
Quality is a term we hear frequently: That is a quality automobile, she is a quality person, and
this is a quality stock. Most people equate high quality with a big price tag, and low quality
with a small price tag. The purpose of this section is to debunk this outdated notion of quality
and to explain what it really means.
Goal Post View of Quality
When watching a game of American football, as long as the ball is kicked between the goalposts, no matter how close to either goalpost, it is considered “good” (Gitlow et al., 2015).
In quality the same used to be true as quality meant “conformance to valid customer requirements.” That is, as long as an output fell within acceptable limits (the goal posts), called
specification limits, around a desired value, called the nominal value (denoted by “m”), or target
value
, it was deemed conforming, good, or acceptable. The nominal value and specification
limits are set based on the perceived needs and wants of customers.
Figure 2.6 shows the goal post view of losses arising from deviations from the nominal value.
That is, losses are zero until the
lower specification limit (LSL) or upper specification limit
(USL)
is reached. Then suddenly they become positive and constant, regardless of the magnitude of the deviation from the nominal value.
From the Library of Pearson HED
32 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Quality characteristic
LSL m USL

 

Good, no loss No
good,
loss
No
good,
loss

LOSS
Figure 2.6 Goal posts view of losses arising from deviations from the nominal value
As an example of the goal post view, the desired diameter of a prescription container in a
hospital pharmacy is 25 mm (the nominal value). A tolerance of 2 mm above or below 25
mm is acceptable to pharmacists. Thus, if the diameter of a prescription container measures
between 23 mm and 27 mm (inclusive), it is deemed conforming to specifications. If the
diameter of a prescription container measures less than 23 mm or greater than 27 mm, it is
deemed not conforming to specifications as the lids will hardly fit and is scrapped at a cost
of $1.00 per container, which is a stiff pill to swallow. In this example, 23 mm is the lower
specification limit and 27 mm is the upper specification limit.
Takeaway: Output merely has to be between specification limits; that is, the diameter of the
prescription container must be anywhere from 23 to 27 inclusive.
Continuous Improvement Definition of Quality—
Taguchi Loss Function
As the definition of quality has evolved, its meaning has shifted. A modern definition of quality states that “quality is a predictable degree of uniformity and dependability, at low cost and
suited to the market” (Deming, 1982). Figure 2.7 shows a more realistic loss curve developed
by Dr. Genichi Taguchi in 1950 (Taguchi and Wu, 1980) called the Taguchi Loss Function
(TLF). Using the TLF view of quality, losses begin to accrue as soon as a quality characteristic
of a product or service deviates from the nominal value. As with the goal post view of quality,
once the specification limits are reached the loss suddenly becomes positive and constant,
regardless of the deviation from the nominal value beyond the specification limits.
Curve A represents the distribution of output of the process before process improvement.
Curve B represents the distribution of output of the process after process improvement. The
shaded area under the loss curve framed by the distribution of output represents the cost of
poor quality. As you can see, the area under the Taguchi Loss Function for curve A is shaded
both gray and black; however, the area under the Taguchi Loss Function for curve B is only
shaded black. So, decreasing the variation in the distribution of output (from curve A to
curve B) decreases cost, absent capital investment.
From the Library of Pearson HED
Chapter 2 Process and Quality Fundamentals 33
Takeaway: Reducing unit-to-unit variation around the nominal value always makes sense,
absent capital investment. At that point it becomes a cost-benefit decision for management.
More Quality Examples
An individual visiting an outpatient clinic expects to wait an hour to see the physician, but if
she has to wait five hours she will perceive the quality to be poor and probably won’t come
back.
If a materials management worker receives orders from a supplier without any missing supplies, her needs will be met and she will perceive the quality of that supplier as good, that is,
unless the supplies are bad.
If a patient in the hospital finds a clean, comfortable room on a quiet floor, he will feel that
his expectations were met. But if the room is not clean or there is constant noise that affects
his ability to sleep, the patient will perceive that the quality is poor and seek revenge by giving
the hospital crappy patient satisfaction scores.
Takeaways from This Chapter
Processes transform inputs into outputs.
Processes are everywhere in our lives and our jobs; improving our lives and/or jobs is
accomplished by improving our processes.

Nominal
Loss incurred
from unit-to
unit variation
after
improvement
Loss incurred
from unit-to
unit variation
before
improvement
Loss
curve
B A
Loss Probability

 

Nominal Value
(m)
Quality
Characteristic
LSL USL

Figure 2.7 Continuous improvement view of losses from deviations from the nominal value
From the Library of Pearson HED
34 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Feedback loops utilize data to help us improve our processes.
Variation consists of the differences in the distribution of the measurements of outputs from a process.
Variation limits your ability to predict the future outcomes of a process.
There are two types of variation: special causes and common causes.
Common causes are inherent to the system and are the responsibility of management.
Special causes are external to the system and the responsibility of front line workers.
Dr. Deming used the Funnel Experiment to show that tampering with a process
makes things worse and the Red Bead Experiment to show that common causes of
variation are inherent in any system and that you cannot expect employees to produce
beyond the capability of the system.
The goal post view of quality states that as long as the output is within specification
limits it is considered good or conforming no matter how close it is to either specification limit.
The Taguchi Loss Function shows that it is always advised to reduce variation around
nominal, absent capital investment.
References
Boardman, T. J. and H. Iyer (1986), The Funnel Experiment (Fort Collins, CO: Colorado
State University).
Deming, W. E. (1982),
Quality, Productivity, and Competitive Position (Cambridge, MA:
Massachusetts Institute of Technology).
Deming, W. E. (1986),
Out of the Crisis (Cambridge, MA: Massachusetts Institute for Technology Center for Advanced Engineering Study).
Deming, W. E. (1994),
The New Economics for Industry, Government, Education, 2nd ed.
(Cambridge, MA: Massachusetts Institute for Technology Center for Advanced Engineering Study).
Gitlow, H., and S. Gitlow (1987),
The Deming Guide to Quality and Competitive Position
(Englewood Cliffs, NJ: Prentice-Hall).
Gitlow, H., A. Oppenheim, R. Oppenheim, and D. Levine (2015),
Quality Management: Tools
and Methods for Improvement
, 4th ed. (Naperville, IL: Hercher Publishing Company).
This book is free online at hercherpublishing.com.
Taguchi, G. and Y. Wu (1980),
Introduction to Off-Line Quality Control (Nagoya, Japan:
Central Japan Quality Control Association).
From the Library of Pearson HED
35
3
Defining and Documenting a Process
What Is the Objective of This Chapter?
To improve a process you must define it, then you must document it, and finally you must
analyze it.
The objective of this chapter is to teach you how to do the following:
Define a process by looking at who owns and is responsible for the process, by understanding the boundaries of the process, and by understanding the objectives and
metrics of the process used to measure its success.
Document a process by understanding flowcharts.
Analyze a process by understanding how to analyze a flowchart to begin improving a
process.
A Story to Illustrate the Importance of Defining and
Documenting a Process
Defining and documenting a process is a critical step toward improvement and/or innovation of a process. The following example demonstrates this point. In a study of a hospital
laundry office, an analyst began to diagram the flow of paperwork using a flowchart. While
walking the green copy of an invoice through each step of its life cycle, he came upon an
administrative assistant transcribing information from the green invoice copy into large
black loose-leaf books. The analyst asked what she was doing so that he could record it on
his flow diagram. She responded, “I’m recording the numbers from the green papers into
the black books.” He asked, “What are the black books?” She said, “I don’t know.” He asked,
“What are the green papers?” She said, “I have no idea.” He asked, “Who looks at the black
books?” She said, “Well, Mr. Johnson used to look at the books, but he died 5 years ago.”
The analyst asked, “Who has looked at them for the last 5 years?” She said, “Nobody.” He
asked, “How long have you been doing this job?” She said, “Seven and one half years.” He
asked, “What percentage of your time do you spend working on the black books?” She said,
“100 percent.”
From the Library of Pearson HED
36 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Next, the analyst did two things. First, he examined the black books. From examining the
black books he realized that they were inventory registers. Every day all laundry pulled from
central supply was recorded on the green papers (invoices) by item onto the appropriate page
in the black books. At the end of each month, page totals were calculated, yielding monthly
used laundry inventory by item. Second, he asked the administrative assistant how long ago
the person who hired and trained her had left the hospital, to which she responded, “Five
years ago.” At this point, the consultant realized he had solved a problem the hospital did not
know it had. Nobody was looking at the books because nobody knew what the administrative
assistant was doing. The current manager assumed the administrative assistant was doing
something important. Surprisingly, the current manager had computerized the inventory
registers as soon as he came on the job five years ago! So, the administrative assistant had
been doing a redundant job for five years. Nobody bothered to document the process. The
administrative assistant was a wheel that didn’t squeak; so why study her? As an epilogue,
the administrative assistant was reassigned to other needed duties because she was a good
employee.
This problem is an example of a failure to define and document a process to make sure that
it is logical, complete, and efficient.
Fundamentals of Defining a Process
Defining a process requires answers to the following questions: (1) Who own the process?
(2) What are the boundaries of the process? (3) What are the process’s objectives? (4) What
measurements are being taken on the process with respect to its objectives? A process definition is created by answering those questions.
Who Owns the Process? Who Is Responsible for the Improvement
of the Process?
Every process must have an owner; that is, an individual responsible for the process (Gitlow
and Levine, 2004). Process owners can be identified because they can change the flowchart
of a process using only their signature. Process owners may have responsibilities extending
beyond their departments, called cross-functional responsibilities; their positions must be
high enough in the organization to influence the resources necessary to take action on a
cross-functional process. In such cases, a process owner is the focal point for the process, but
each function of the process is controlled by the line management within that function. The
process owner requires representation from each function; these representatives are assigned
by the line managers. They provide functional expertise to the process owner and are the
promoters of change within their functions. A process owner is the coach and counsel of her
process in an organization.
The identification and participation of a process owner are critical in defining a process. It
is usually a waste of time to be involved in defining and documenting a process, as part of
process improvement activities, without the complete commitment of the process owner.
From the Library of Pearson HED
Chapter 3 Defining and Documenting a Process 37
What Are the Boundaries of the Process?
Next, boundaries must be established for processes; in other words, before a flowchart of the
process can be created, the process owner must help you identify where the process starts
and stops (Gitlow and Levine, 2004). These boundaries make it easier to establish process
ownership and highlight the process’s key interfaces with other (customer/vendor) processes.
Process interfaces frequently are the source of process problems, which result from a failure
to understand downstream requirements; they can cause turf wars. Process interfaces must
be carefully managed to prevent communication problems. Construction of operational
definitions for critical process characteristics agreed upon by process owners on both sides of
a process boundary goes a long way toward eliminating process interface problems. Operational definitions are discussed later in this book.
An Anecdote about Process Boundaries

One time I was consulting at a paper mill, and I noticed that the entire mill was sur
rounded by a nine-foot-tall chain link fence. I realized that since a paper mill is a
dangerous place, even potential intruders have to be protected.
As was my custom, I started my tour from the beginning of the process, in this case
the wood procurement area. This is the where trees enter the mill on flatbed trucks
and are cut into 40-foot lengths by large saws. Everyone I met was nice and helpful.
The next part of the process was the wood yard. This is the area where the 40-foot
logs are turned into wood chips for making paper. I noticed that there was a chain
link fence between these two areas. That was to prevent truckers from wandering
into the wood yard. However, the chain link fence also had a door that was pad
locked, and on top of the chain link fence was concertina barbed wire, coils of wire
with razor blades attached. It would slice to pieces anyone trying to get over it.
I wondered why the wood procurement area would be more concerned about the
wood yard employees than they would be about outsiders. The wood procurement
folks opened the padlocked door between the two areas and let me enter. I thought
I was about to be attacked by wild lions. In the distance I saw a man walking toward
me, and as he got close, I stuck out my hand to shake the wood yard manager’s hand.
The wood yard manager did not reciprocate and called me an ethnic slur that was
relevant only to a small area in Europe that my ancestors are from. How this man
would have known this term was a mystery to me. Apparently he was a savant of
bigotry.
Now I knew why there was concertina wire between the two mill areas; it was a
statement of mutual hatred. When I asked the wood yard manager why there was
so much hatred between the two mill areas, he responded: “They are morons! They

From the Library of Pearson HED
38 A Guide to Six Sigma and Process Improvement for Practitioners and Students
What Are the Process’s Objectives? What Measurements Are Being
Taken on the Process with Respect to Its Objectives?
A key responsibility of a process owner is to clearly state the objectives of the process and
indicators that are consistent with organizational objectives (Gitlow and Levine, 2004). An
example of an organizational objective is “Provide our customers with higher-quality products/services at an attractive price that will meet their needs.” Each process owner can find
meaning and a starting point in the adaptation of this organizational objective to his process’s
objectives. For example, a process owner in the purchasing department of a health system
could translate the preceding organizational objective into the following subset of objectives
and metrics:
Objective: Decrease the number of days from purchase request to item/service
delivery.
Metric: Number of days from purchase request to item/service delivery by
delivery overall, and by type of item purchased, by purchase.
Objective: Increase ease of filling out purchasing forms.

Metric: Number of questions received about how to fill out forms by form
by month.
Objective: Increase employee satisfaction with purchased material
Metric: Number of employee complaints about purchased material by month.
Objective: Continuously train and develop purchasing personnel with respect to job
requirements.
Number of errors per purchase order by purchase order.
Number of minutes to complete a purchase order by purchase order.
Metric:
Metric:

Whatever the objectives of a process are, all involved persons must understand them and
devote their efforts toward those objectives. A major benefit of clearly stating the objectives
of a process is that everybody works toward the same aim/mission.
cut the logs into 40-foot lengths and we need 20-foot lengths for our equipment to
operate at its best.” When I asked why they didn’t know to cut the logs into 20-foot
lengths the manager said, “They should know.”
At that point, I realized that the wood procurement area was so put off by the
manager of the wood yard that they had no communication whatsoever, even to
the extent of which size saws to purchase. This is a classic example of a clear dysfunctional process boundary. Most disagreements occur at process boundaries.
From the Library of Pearson HED
Chapter 3 Defining and Documenting a Process 39
Fundamentals of Documenting a Process
Now that we have identified the process owner, know where the process starts and stops, and
understand the objectives/metrics to measure success, it is time to document the process.
Documenting a process requires input from all stakeholders of the process, as they may have
different points of view on the flow of the process.
How Do We Document the Flow of a Process?
To document a process we use a flowchart, which is a pictorial summary of the steps, flows,
and decisions that comprise a process (Fitzgerald and Fitzgerald, 1973; Silver and Silver,
1976). Figure 3.1 shows a simple generic flowchart.
Process Step

Start
Stop
Decision
No
Yes
Figure 3.1 A s imple generic flowchart
Why and When Do We Use a Flowchart to Document a Process?
Documenting a process using a flowchart, as opposed to using written or verbal descriptions
has several advantages:
A flowchart makes it easier for people who are unfamiliar with a process to understand it.
A flowchart allows for employees to visualize what actually happens in a process, as
opposed to what is supposed to happen.
From the Library of Pearson HED
40 A Guide to Six Sigma and Process Improvement for Practitioners and Students
A flowchart functions as a communications tool. It provides an easy way to convey
ideas between engineers, managers, hourly personnel, vendors, and others in the
interdependent system of stakeholders for the organization. It is a concrete, visual
way of representing a complex system.
A flowchart functions as a planning tool. Designers of processes are greatly aided by
flowcharts. They enable a visualization of the elements of new or modified processes
and their interactions.
A flowchart removes unnecessary details and breaks down the system so designers
and others get a clear, unencumbered look at what they’re creating.
A flowchart defines roles. It demonstrates the functions of personnel, workstations,
and subprocesses in a system. It also shows the personnel, operations, and locations
involved in the process.
Flowcharts can be used in the training of new and current employees.
A flowchart helps you understand what data needs to be collected when trying to
measure and improve a process.
A flowchart can also be used to be compliant with regulatory agencies, i.e., JCAHO
(Joint Commission on Accreditation of Healthcare Organizations)
A flowchart can be used to compare the current state of the process (how the process
is), the desired state of the process (how the process should be with standardization),
and the future state of the process (how the process could be with improvement).
And last, but not least, a flowchart helps to identify the weak points in a process; that
is, the points in the process that are causing problems for the stakeholders of the
process.
Flowcharts can be applied in any type of process to aid in defining and documenting it, and
ultimately to improve and innovate the process.
What Are the Different Types of Flowcharts and When Do We
Use Each?
This chapter covers two types of flowcharts used in process improvement activities: process
flowcharts and deployment flowcharts. Each type of flowchart has different features discussed in the following sections.
Process Flowchart
What is it?
A flowchart that lays out process steps and decision points in a downward direction
from the starting point to the stopping point of the process.
From the Library of Pearson HED
Chapter 3 Defining and Documenting a Process 41
When to use it?
When you want to depict a process at a high level or when you want to drill down into
a detailed portion of a process.
What does it look like?
An example of a process flowchart for a typical inpatient cardiology consult process
is shown in Figure 3.2 .

No Sr. patient rep
ships forms to
billing office.

Yes
START
Inpatient Cardiology Consult Process
STOP
Physician
prepares
handwritten
note.
Sr. patient rep
prepares a facesheet with name
and DOB of
patient.
Physician
completes
encounter form
after consult.
All forms are
batched
(facesheets and
encounters).
Administrator
collects all forms
every Thursday
morning in
person.
Did
Administrator
collect form?
Figure 3.2 Process flowchart example
From the Library of Pearson HED
42 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Deployment Flowchart (Also Known As Cross-Functional or Swim Lane
Flowcharts)
What is it?
A flowchart organized into “lanes” that show processes that involve various departments, people, stages, or other categories.
When to use it?
When you want to show who is responsible for different parts of a process as well as
track the number and location of handoffs within the process.
What does it look like?
An example of a deployment flowchart for a surgical biopsy sign-out process is shown
in Figure 3.3 .

START GI Clinic
GI Pathology Biopsy-Process Map–After Change
Pathology Department END
Slides of new
cases distributed
by secretary.
Cases
reviewed by
fellow daily.
Cases
reviewed by
resident daily.
Cases with
odd numbers??
Request slides
for extra staining.
Put case
on hold.
Case diagnosed
by pathologist 1
next day from
8:00-12:00.
6
Transcription
of final diagnosis
by transcriptionist.
Case diagnosed
by pathologist 2
next day from
8:00-12:00.
Pending cases
diagnosed by
pathologist and
resident/fellow
daily in second shift.
Transcription of
final diagnosis with
ancillary results by
transcriptionist.
Sign off final
diagnosis
electronically by
pathologist.
Final diagnosis
with ancillary results
by pathologists.
Slides
received by
secretary.
Does case
need ancillary
testing?
Is the staining
critical?
Slides of
pending cases
distributed by
secretary.
Y
N
Y
Y
N
GI Path Lab 4
5

Case obtained
by medical service.
Received cases
signed off in lab
notebook by lab
asst.
Accessioning
conducted by
lab asst.
Tissue along with
pathology request form is
placed at sample collection
room and case logged in
log book by GI nurse.
1
Case transported
to lab by hospital courier
once a day.
Dropped cases logged
in path lab notebook the
by courier.
2
3
Macroscopic
evaluation by
path asst.
Macroscopic
evaluation
transcribed by
transcriptionist.
Tissue
processing using
short cycle
by histo tech.
Slides
sent to case
manager by
histo tech.
Reconciliation of
slides with macroscopic
evaluation by histo tech.
Histologic
preparation and
QC of slides
by histo tech.
Slides
transported from
hospital path lab
by AP courier.
Transcript on
of final diagnosis
without ancillary testing
by transcriptionist.
Sign off the
addendum
electronically by
pathologist.
N
Figure 3.3 Deployment flowchart example
From the Library of Pearson HED
Chapter 3 Defining and Documenting a Process 43
The starbursts in Figure 3.3 show areas of the process that are suspected of causing problems
in the process’s outputs. The starbursts can come from process experts, reviews of the available literature, benchmarking with similar processes in other organizations, or many other
possible sources. All the possible sources of the starbursts are discussed later in this book.
What Method Do We Use to Create Flowcharts?
The American National Standards Institute, Inc. (ANSI), developed a standard set of flowchart symbols used for defining and documenting a process, some of which are shown in
Table 3.1 . The shape of the symbol and the information written within the symbol provide
information about that particular step or decision in a process.
Table 3.1 American National Standards Institute Approved Standard Flowchart Symbols

Symbol Function
Start/stop symbol The general symbol used to indicate the beginning and end of a
process is an oval.
Basic processing symbol The general symbol used to depict a processing operation is a
rectangle.
Decision symbol A diamond is the symbol that denotes a decision point in the
process. This includes decisions such as pass-fail or yes-no, which
creates branches on the flowchart.
Flowline symbol A line with an arrowhead is the symbol that shows the direction of
the stages in a process. The flowline connects the elements of the
system.

Flowcharts provide a common language for the stakeholders of a process to discuss
the process, for example, where it begins and ends, who is responsible for each step in
the process, how does the process flow, how should the process flow, to name a few
benefits of flowcharts.
Flowcharts help managers see that they are responsible for the outputs of the process,
as opposed to the employees who work in the process, because the managers design
and manage the process.
Flowcharts provide stakeholders a tool to walk the process from front to back (process
owner’s point of view) and back to front (customer’s point of view).
From the Library of Pearson HED
44 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Flowcharts provide an opportunity for the stakeholders of a process to identify the
problematic steps in the process (starbursts).
Fundamentals of Analyzing a Process
One of the best methods to define and analyze a process is a flowchart. We discuss flowcharts
in detail in this section.
How Do We Analyze Flowcharts?
Process improvers can use a flowchart to change a process by paying attention to the following five points:
1. Process improvers find the steps of the process that are weak (for example, parts of
the process that generate a high defect rate).
2. Process improvers improve the steps of the process that are within the process owner’s
control; that is, the steps of the process that can be changed without higher approval
than the process owner.
3. Process improvers isolate the elements in the process that affect customers.
4. Process improvers find solutions that don’t require additional resources.
5. Process improvers don’t have to deal with political issues.
If these five conditions exist simultaneously, an excellent opportunity to constructively modify a process has been found. Again, process improvements have a greater chance of success
if they are either nonpolitical or have the appropriate political support, and either do not
require capital investment or have the necessary financial resources.
Other questions that the process improver can ask are the following:
Why are steps done? How else could they be done?
Is each step necessary?
Is it value added and necessary? Is it repetitive?
Would a customer pay for this step specifically? Would the customer notice if
it’s gone?
Is it necessary for regulatory compliance?
Does the step cause waste or delay?
Does the step create rework?
Could the step be done in a more efficient and less costly way?
From the Library of Pearson HED
Chapter 3 Defining and Documenting a Process 45
Is the step happening at the right time? (sequence)
Could this step be done in parallel with another step to cut cycle time?
Are the right people doing the right thing?
Could this step be automated?
Does the process contain proper feedback loops?
Are roles and responsibilities for the process clear and well documented?
Are there obvious bottlenecks, delays, waste, or capacity issues that can be identified
and removed?
What is the impact of the process on stakeholders?
Things to Remember When Creating and Analyzing Flowcharts
Work with people who really know and live the process such as front-line employees.
Managers may think they know how it works, or how it is supposed to work, but those
on the front lines can tell you how it really works.
Make people understand you are only there to help. The last thing you want people to
think is that their jobs may be in jeopardy if the process gets improved so much that it
is no longer necessary. You are there to help them do their jobs better, not to eliminate
them. Explain they have employment security, not job security. Job security leads to
redundancy and unnecessary work; for example, if a unionized electrician knocks a
water pipe loose, she can’t fix it due to union rules. A plumber must be called in to fix
it and much damage could result in the factory in the interim. However, if employees
are guaranteed employment and wage security, they are more open to cross-training
and the preceding scenario would not happen.
Start high level to identify major components of the process; then drill down.
Keep asking questions, and question everything.
Involve enough people so you get a complete understanding of the process.
Validate and verify with key stakeholders to make sure the process is understood.
Keep the flowchart as simple and understandable as possible so anyone can follow it.
Walk the process from front to back (process owner’s point of view) and from back
to front (customer’s point of view).
Focus on the needs of the customer.
Only improve processes with data and facts.
From the Library of Pearson HED
46 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Takeaways from This Chapter
Four steps are used to define a process:
1. Identify the process.
2. Identify the process owner.
3. Identify where a process starts and where it ends.
4. Identify the objectives and metrics to measure the success of a process.
Flowcharts are used to document a process.
The two main flowcharts used in process improvement activities are process flowcharts and deployment flowcharts.
References
Deming, W. E. (1982), Quality, Productivity, and Competitive Position (Cambridge, MA:
Massachusetts Institute of Technology, Center for Advanced Engineering Study).
Deming, W. E. (1986),
Out of the Crisis (Cambridge, MA: Massachusetts Institute of Technology, Center for Advanced Engineering Study).
Fitzgerald, J. M. and A. F. Fitzgerald (1973),
Fundamentals of Systems Analysis (New York:
John Wiley and Sons).
Gitlow, H., A. Oppenheim, R. Oppenheim, and D. Levine (2015),
Quality Management: Tools
and Methods for Improvement
, 4th ed. (Naperville, IL: Hercher Publishing Company).
This book is free online at hercherpublishing.com.
Gitlow, H. and D. Levine (2004),
Six Sigma for Green Belts and Champions: Foundations, DMAIC, Tools and Methods, Cases and Certification (Upper Saddle River, NJ:
Prentice-Hall).
Silver, G. A. and J. B. Silver (1976),
Introduction to Systems Analysis (Englewood Cliffs, NJ:
Prentice-Hall).
From the Library of Pearson HED
47
4
Understanding Data: Tools and Methods
What Is the Objective of This Chapter?
The objective of this chapter is to introduce you to some statistical tools that help you understand the data that you will use in the Six Sigma DMAIC model. The chapter is split into two
sections. The first section is a high level overview of the tools and methods used in understanding data by looking at what each tool and method is, why each tool and method is used,
and examples of each tool and method. The second section shows you how to use Minitab to
utilize the different tools and methods you learn about in the first section.
What Is Data?
Data is information collected about a product, service, process, individual, item, or thing.
Because no two things are exactly alike, data inherently varies. Each characteristic of interest
is referred to as a
variable. Data can also be words, sounds, pictures, to name a few types of
data. We focus mainly on numerical data in this chapter and “word” data later in this book.
Types of Numeric Data
There are two basic types of numeric data; they are attribute data and measurement data.
Each type of data is discussed in the following sections.
Attribute Data
Attribute data (also referred to as classification or count data) occurs when a variable is
either classified into categories (defective or conforming) or used to count occurrences of a
phenomenon (number of patient falls on a particular hospital floor in a particular month).
Attribute Classification Data
Attribute classification data places an item into one of two or more categories; for example,
not defective (fit for use) or defective (not fit for use). Some examples of attribute classification data are the following:
From the Library of Pearson HED
48 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Percent of accounts receivable older than 90 days per month. Either the account is
over 90 days or it isn’t over 90 days; there are only two categories.
Percent of employees off sick by supervisor by day. Either the employee is off sick or
not; there are only two categories.
Percent of occurrences of surgery delays in an operating room by month. Either the
surgery is delayed or not; again, there are only two categories.
Table 4.1 shows the classification of defective items from a daily sample of 100 units for 11
days.
Table 4.1 Defective Items from a Daily Sample of 100 Units for 11 Days

Day Number Defective Sample Size Proportion Defective
1 11 100 0.11
2 21 100 0.21
3 13 100 0.13
4 20 100 0.20
5 14 100 0.14
6 21 100 0.21
7 19 100 0.19
8 18 100 0.18
9 30 100 0.30
10 21 100 0.21
11 23 100 0.23

Attribute Count Data
Attribute count data consists of the number of defects in an item or area of opportunity (for
example, a microscope, a room, a stretch of highway, a hospital, and so on). An item can
have multiple defects and still not be defective. However, it is possible that one or more of the
defects in an item make the item defective. For example, if a water bottle leaks it is defective.
However, if the water bottle has a dent and a scratch on the label, it has two defects, neither
of which makes the item defective (it is still fit for use).
Some examples of attribute count data are
The number of data entry errors on a patient chart by chart
The number of cars entering a hospital parking garage by day
The number of surgeries performed on the wrong patient per year
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 49
The number of patient falls per week in a hospital is attribute count data and can be seen in
Table 4.2 .
Table 4.2 Patient Falls Per Week in a Hospital

Week Falls
1 10
2 6
3 9
4 11
5 14
6 7
7 10
8 12
9 9
10 11

Measurement Data
Measurement data (also referred to as continuous or variables data) results from a measurement taken on an item of interest, or the computation of a numerical value from two or
more measurements of variables data. Any value can theoretically occur, limited only by the
precision of the measuring process. This type of data can have decimal points—for example,
height, weight, temperature, waiting time, service time, diameter, revenues, costs, and cycle
time.
Some examples of measurement data are
Height by person
Waiting time by patient
Revenue by month
Cost by line item by store by month
Other examples of measurement data include miles since refueling by ambulance, gallons
consumed per ambulance, and a computation of miles per gallon per ambulance as shown
in Table 4.3 .
From the Library of Pearson HED
50 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Table 4.3 Miles Since Refueling, Gallons Consumed, and MPG

Measurement of a
Characteristic
Measurement of a
Characteristic
Computation of a
Characteristic
Truck Miles Since Refueling Gallons Consumed MPG
1 308 17.3 17.8
2 256 15.3 16.7
3 274 16.5 16.6
4 310 16.9 18.3
5 302 17.1 17.7
6 296 17.3 17.1

Graphing Attribute Data
When dealing with attribute data, responses are tallied into two or more categories, and the
frequency or percentage in each category is obtained. Three widely used graphs—the bar
chart, the Pareto diagram, and the line chart—are presented in this section.
Bar Chart
What: A bar chart presents each category of an attribute variable as a bar whose length is the
frequency or percentage of observations falling into a particular category. The width of the
bar is meaningless for a bar chart, but all bars should be the same width.
Why: A bar chart is used to graphically display the frequency or percentages of items that fall
into two or more categories.
Example: To illustrate a bar chart we examine data from a hospital pharmacy regarding
reasons on delays to orders for the hospital’s chemotherapy treatment unit. Table 4.4 shows
the data collected on the reasons for delays.
Table 4.4 Pharmacy Delay Reasons (January 1 Through June 30, 2014)

Delay Reason Number Percentage
Missing D.O.S. 74 41%
Missing height and weight 66 37%
Dose change 15 8%
Order clarification 9 5%
No consent form 7 4%
Labs pending 6 3%
Labs high 2 1%

From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 51
The bar chart in Figure 4.1 shows that missing D.O.S. (date of service) and missing height
and weight are the two most problematic reasons for delays in the pharmacy from the data
in Table 4.4 .
Figure 4.1 Delay reasons
Pareto Diagrams
What: Pareto diagrams are used to identify and prioritize issues that contribute to a problem
we want to solve. Created by Italian economist Vilfredo Pareto, Pareto analysis focuses on
distinguishing the vital few causes of problems from the trivial many causes of problems.
The vital few are the few causes that account for the largest percentage of the problem, while
the trivial many are the myriad of causes that account for a small percentage of the problem.
Why: A Pareto diagram is used to graphically display attribute data; specifically it is used
to distinguish the significant few categories from the trivial many categories. Consequently,
you are able to prioritize efforts on the most important causes (categories) of the problem.
Pareto diagrams rank problematic categories from the largest to the smallest, except for the
last category that may be “other.” This is the source of the famous 80-20 rule. The general
idea is that 80% of your problems come from 20% of your causes.
Example: The Director of a hospital pharmacy is interested in learning about the reasons on
delays to orders for the hospital’s chemotherapy treatment unit. Using the data from Table
4.4 he creates the Pareto diagram in Figure 4.2 .
From the Library of Pearson HED
52 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 4.2 Delay reasons Pareto diagram
The Pareto diagram in Figure 4.2 shows that the major reasons for delays are missing (D.O.S.)
date of service and missing height and weight on the orders. These two categories account for
78.2% of delays. There are six categories, so all things being equal you would expect one-sixth
(16.6%) of the data to be in each category, but 78.2% are in the first two categories instead
of 33.2%. Graphically displaying data by using a Pareto diagram promotes prioritization of
effort that discourages micromanagement.
Line Graphs
What: A line graph is a graph of any type of variable plotted on the vertical axis and usually
time plotted on the horizontal axis.
Why: A line graph is generally used to graphically display data over time.
Example: A line graph is illustrated by using data concerning a medical transcription service
that enters medical data on patient files for hospitals. The service studied ways to improve
the turnaround time (defined as the time between receiving data and time the client receives
completed files). After studying the process, the service determined that transcription errors
increased turnaround time. Table 4.5 presents the number and proportion of transcription
with errors by day.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 53
Table 4.5 Transmission Errors

Month Date Number of Errors Proportion of Errors
August 1 6 0.048
August 2 3 0.024
August 5 4 0.032
August 6 4 0.032
August 7 9 0.072
August 8 0 0.000
August 9 0 0.000
August 12 8 0.064
August 13 4 0.032
August 14 3 0.024
August 15 4 0.032
August 16 1 0.008
August 19 10 0.080
August 20 9 0.072
August 21 3 0.024
August 22 1 0.008
August 23 4 0.032
August 26 6 0.048
August 27 3 0.024
August 28 5 0.040
August 29 1 0.008
August 30 3 0.024
September 3 14 0.112
September 4 6 0.048
September 5 7 0.056
September 6 3 0.024
September 9 10 0.080
September 10 7 0.056
September 11 5 0.040
September 12 0 0.000
September 13 3 0.024

From the Library of Pearson HED
54 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 4.3 presents the line chart for these data. The line graph clearly shows a great deal of
fluctuation in the proportion of transcription errors from day to day. The highest number of
errors occurred on day 23 (September 3), but a large number of errors also occurred on days
5, 8, 13, 14, and 27. The medical transcription service needs to study the process to determine
the reasons for this variation. Are the variations due to special causes? Or are the variations
due to common causes? Methods for studying these issues using control charts are covered
in Chapter 5 , “Understanding Variation: Tools and Methods,” so stay tuned!
Figure 4.3 Line graph for transcription errors
Graphing Measurement Data
Many people do not have the ability to look at data and make much sense of it. Consequently,
process improvers and other scientists create graphical or pictorial representations of data
to help them understand it. In this section, we discuss several of these graphical or pictorial
representations used for measurement data; they are histograms, dot plots, and run charts.
You will find these tools extremely valuable in understanding your data; they let the data talk
to you about what is going on in the process that generated them.
Histogram
What: A histogram is a special bar chart for measurement data. In the histogram, the data is
grouped into adjacent numerical categories of equal size, for example, 100 to less than 200,
200 to less than 300, 300 to less than 400, and so on. The difference between a bar chart and
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 55
a histogram is that the X axis on a bar chart is a listing of categories, whereas the X axis on
a histogram is a measurement scale. In addition there are no gaps between adjacent bars.
Why: A histogram is used to graphically display measurement data to understand the distribution of the data.
Example: To illustrate a histogram we examine 100 patient wait times in minutes at an
outpatient clinic in Figure 4.4 . For these data, notice tick marks at 30, 40, 50, 60, 70, and 80.
The distribution appears to be approximately bell-shaped with a heavy concentration of wait
times between 40 and 70.
Figure 4.4 Histogram of patient wait times
Dot Plot
What: Similar to a histogram, a dot plot is a graph of measurement data in which dots that
represent data values are stacked vertically on the horizontal axis for each value of the variable of interest.
Why: A dot plot is used to graphically display measurement data to understand the distribution of the data. Dots plots are generally used when you have smaller sets of data as once data
sets become larger the dot plot becomes cluttered.
Example: To illustrate a dot plot we examine the same 100 patient wait times in minutes at
an outpatient clinic in Figure 4.5 (refer to the histogram in Figure 4.4 ). Note that the dot plot
looks different from the histogram. This occurs because the histogram groups the data into
From the Library of Pearson HED
56 A Guide to Six Sigma and Process Improvement for Practitioners and Students
class intervals, whereas the dot plot presents each data value, with the height representing
the frequency at each horizontal value. Nevertheless, the dot plot shows a concentration of
values in the center of the distribution between 40 and 70 minutes.
Figure 4.5 Dot plot of patient wait times
Run Chart
What: A run chart is a type of line chart that may have multiple measurements for each
time period, called subgroups. Time is the x-axis and the variable of interest is plotted on
the y-axis; remember there may be multiple data points on the y-axis for each time period
on the x-axis.
Why: A run chart is used to graphically display measurement data where there may be multiple data points per subgroup. When data are collected over time, the variable of interest
should be plotted in time order before any other graphs are plotted, descriptive statistics are
calculated, or statistical analyses are performed.
Example: To illustrate a run chart, we examine a sample of three waiting times (in minutes)
for patients per day in a medical clinic; we determine the waiting time for the first patient to
enter at 9:00 a.m., then at noon, and then at 4:00 p.m. The data appears in Table 4.6 .
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 57
Table 4.6 Waiting Times for a Sample of Three Patients per Day in a Medical Clinic (in minutes)

Waiting Time Waiting Time Waiting Time
Day (9:00 a.m.) (noon) (4:00 p.m.)
1 109.909 94.580 108.207
2 104.939 106.269 08.483
3 96.738 103.557 99.587
4 104.494 94.569 84.540
5 112.930 104.091 103.763
6 88.272 105.702 85.598
7 106.013 112.941 121.007
8 98.202 102.188 101.633
9 102.216 112.204 86.541
10 79.289 100.382 117.723
11 93.124 95.098 106.073
12 100.466 88.852 95.471
13 112.393 102.399 105.336
14 100.049 100.118 100.838
15 90.865 91.397 115.002
16 91.230 117.406 99.431
17 88.502 100.035 100.881
18 116.279 113.906 97.700
19 79.182 84.080 93.334
20 107.262 102.475 96.598
21 115.299 108.362 106.973
22 88.139 109.658 95.129
23 112.439 99.519 88.262
24 103.661 105.989 106.239
25 97.802 100.906 99.214
26 99.055 101.062 105.244
27 106.934 108.751 98.336
28 92.009 91.027 119.083
29 102.465 121.023 106.972
30 95.953 98.419 102.357
31 109.776 95.586 101.465
32 114.890 107.868 97.132
33 110.809 94.834 96.335
34 99.719 101.790 99.900
35 83.911 89.563 100.017

From the Library of Pearson HED
58 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Waiting Time Waiting Time Waiting Time
36 119.598 91.605 113.346
37 108.419 93.357 86.829
38 93.971 88.679 108.423
39 108.036 106.149 95.813
40 96.530 89.928 100.436
41 88.415 95.474 95.795
42 83.524 103.116 107.381
43 88.941 99.971 95.070
44 94.799 110.642 90.903
45 97.267 109.286 92.706
46 91.730 112.460 103.507
47 105.964 116.872 101.706
48 114.394 72.322 101.364
49 108.729 88.161 92.020
50 88.860 121.195 85.736

As you can see from Figure 4.6 , there are the three data points per day, one for the 9:00 a.m.
patient, one for the noon patient, and one for the 4:00 p.m. patient. We learn how to analyze
this data further in Chapter 5 . However, for now we can see that the data is not trending up
or trending down, and it is not getting more variable (like a cone going from left [small] to
right [large] or less variable (like a cone going from left [large] to right [small]).
Figure 4.6 Run chart of patient wait times in minutes
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 59
Measures of Central Tendency for Measurement Data
Although the graphs studied thus far are useful for getting a visual picture of what the distribution of a variable looks like, it is important to compute measures of central tendency
or location. Three measures of central tendency are discussed—the mean, the median, and
the mode.
Mean
What: The most common numerical representation of central tendency is the arithmetic
average or mean. It is the sum of the numerical values of the items being measured divided by
the number of items. Because its computation is based on every observation, the arithmetic
mean is greatly affected by any extreme value or values, so be careful because you may get a
distorted representation of what the data are conveying!
Why: The mean is useful to convey the average value of measurement data especially if the
data is symmetric around the average; for example, a bell shaped histogram or dot plot of
data.
Example: To illustrate the computation of the mean, consider the following example related
to your personal life: the time it takes to get ready for work in the morning. Many people
wonder why it seems to take longer than they anticipate getting ready to leave for work, but
virtually no one has measured the time it takes to get ready in the morning. Suppose the time
to get ready in the morning is operationally defined as the time in minutes (rounded to the
nearest minute) from when you get out of bed to when you leave your house. Suppose you
collect these data for a period of two weeks, ten working days, with the results in Table 4.7 .
Table 4.7 Time to Get Ready for Ten Days
Day 1 2 3 4 5 6 7 8 9 10
Time (minutes) 39 29 43 52 39 44 40 31 44 35
To calculate the arithmetic mean, you simply sum the observations and divide by the number
of observations. In this case you would add up all the minutes and divide by the number of
days, which is 396/10 = 39.6.
The mean (or average) time to get ready is 39.6 minutes, even though not one individual day
in the sample actually had that value. Note that the calculation of the mean is based on all
the observations in the set of data. No other commonly used measure of central tendency
uses this characteristic.
From the Library of Pearson HED
60 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Median
What: The median is the middle value in a set of data that has been ordered from the lowest
to the highest value. Half the observations will be smaller than the median, and half will be
larger. The median is not affected by any extreme values in a set of data.
Why: To convey the underlying character of measurement data by representing the middle
value such that 50% of the observations are smaller and 50% of the observations are larger.
Whenever an extreme value is present, the median is useful in describing the central tendency of the data.
Example: To calculate the median from a set of data, you must first organize the data into
an ordered array that lists the values from smallest to largest. If there is an odd number of
observations in the data set, the median is the middle most number in the data set. If there is
an even number of observations in the data set, the median is the average of the two middle
most numbers in the data set.
Going back to our example of the time it takes to get ready for work, we first arrange the
values from smallest to largest as seen in Table 4.8 .
Table 4.8 Ordered Time to Get Ready for Ten Days
Day 1 2 3 4 5 6 7 8 9 10
Time (minutes) 29 31 35 39 39 40 43 44 44 52
Since there is an even number of observations in the data set we then find the two middle
most numbers, which happen to be 39 and 40; we then average them to get a median of 39.5.
Suppose our data set contained only the first nine observations as seen in Table 4.9 .
Table 4.9 Time to Get Ready for Ten Days from Smallest to Largest
Day 1 2 3 4 5 6 7 8 9
Time (minutes) 29 31 35 39 39 40 43 44 44
In this case we would have an odd number of observations in the data set, so to calculate the
median we would simply find the middle most observation, which would give us a median
of 39.
Mode
What: The mode is the value in a set of data that appears most frequently in a data set. Unlike
the arithmetic mean, the mode is not affected by the occurrence of any extreme values.
Why: The mode is used only to find the most commonly occurring value in a data set.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 61
Example: Referring back to the times for the days, there are two modes: 39 minutes and 44
minutes because each of those values occurs twice. This is called a
bimodal distribution.
Measures of Central Tendency for Attribute Data
Recall that there are two types of attribute data: classification data and count data. Classification and count data are subject to all the measures of central tendency used for measurement
data, but classification data is a bit special because it consists of only 0s and 1s; that is, yes
or no, and so on. This makes the computations a little simpler. Instead of the mean for classification data, we use the proportion.
Proportion
What: Often data are classified into two non-numerical conditions, such as broken versus
not broken, defective versus conforming, operating versus not operating. The proportion or
fraction of the data possessing one of two such conditions is a meaningful measure of central
tendency.
Why: The proportion is used to understand the degree to which the output of a population
or process is in either one or two possible states, for example, defective or conforming, and
so on.
Example: A Chief Medical Officer at a large hospital was concerned with readmissions of a
certain patient population. He took a random sample of 30 patients to look at the proportion
that was readmitted to the hospital within 30 days; see Table 4.10 .
Table 4.10 Hospital Readmissions Within 30 Days

Patient # Condition
1 Readmitted
2 Not Readmitted
3 Not Readmitted
4 Not Readmitted
5 Not Readmitted
6 Not Readmitted
7 Readmitted
8 Not Readmitted
9 Not Readmitted
10 Not Readmitted
11 Not Readmitted
12 Readmitted

From the Library of Pearson HED
62 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Patient # Condition
13 Not Readmitted
14 Not Readmitted
15 Readmitted
16 Not Readmitted
17 Not Readmitted
18 Readmitted
19 Not Readmitted
20 Readmitted
21 Not Readmitted
22 Not Readmitted
23 Not Readmitted
24 Readmitted
25 Readmitted
26 Not Readmitted
27 Not Readmitted
28 Not Readmitted
29 Not Readmitted
30 Not Readmitted

To calculate the proportion (p), he would divide the number readmitted by the total number
of patients in the sample. In this case 8/30 = .27 = 27%.
Measures of Variation
A second important property that describes a set of numerical data is variation. Variation
is the amount of dispersion, or spread, in the data. Three measures of variation include the
range, the variance, and the standard deviation.
Range
What: The range is simply the difference between the largest and the smallest data points
in the data set. It is calculated by subtracting the smallest number from the largest number.
Why: The range is used to analyze the spread of the data. The range assumes that there are
no extreme values in the data because that would distort the range and make it a meaningless
measure of variation.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 63
Example: Using the data on time to get ready in the morning (refer to Table 4.8 ), we would
calculate the range as follows:
Range = largest value – smallest value
Range = 52 – 29 = 23 minutes
This means that the largest difference between any two days in the time to get ready in the
morning is 23 minutes.
Sample Variance and Standard Deviation
What: To understand how the values in the data are spread around the mean we have to
look at two other commonly used measures of variation called the
variance and the standard
deviation
.
The variance and standard deviation measure how far the values in a data set are spread out
around the average. A variance or standard deviation of zero indicates all the values in the
data set are the same.
To calculate the sample variance you simply obtain the difference between each value and
the mean, you square each difference, you add the squared differences, and then you divide
this total by the number of observations minus 1. To calculate the sample standard deviation
you simply take the square root of the variance.
Why: The variance and standard deviation tell us how the values are distributed around the
mean. The variance and standard deviation are less affected by extreme value in the data set
that is the range.
Example: Let’s go back to our time to get ready in the morning example to help us understand variance and standard deviation. The times to get ready are listed in Table 4.11 .
Table 4.11 Time to Get Ready for Ten Days Used to Calculate Variance and Standard Deviation

Time (x) Difference Between
x and Mean
Squared Differences
around the Mean
39 -0.6 0.36
29 -10.6 112.36
43 3.4 11.56
52 12.4 153.76
39 -0.6 0.36
44 4.4 19.36
40 0.4 0.16
31 -8.6 73.96
44 4.4 19.36
35 -4.6 21.16
Mean = 39.6 Sum of differences = 0 Sum of squared differences = 412.4

From the Library of Pearson HED
64 A Guide to Six Sigma and Process Improvement for Practitioners and Students
To calculate the variance we take the following steps:
1. Subtract each value from the mean; see the middle column in Table 4.11 .
2. Square those numbers and add them up to get the sum of squared differences, which
is 412.4 as you see at the bottom of the right column in Table 4.11 .
3. Divide the sum of squared differences by the number of observations (10) minus one
which gives us 412.4/9 = 45.82.
To calculate the standard deviation simply take the square root of the variance, which is 6.77.
Understanding the Range, Variance, and Standard Deviation
So what does this all tell us? The following statements summarize what you need to know
about the range, variance, and standard deviation:
The more spread out, or dispersed, the data are, the larger will be the range, variance,
and standard deviation.
The more concentrated or homogeneous the data is, the smaller will be the range,
variance, and standard deviation.
If the values are all the same (so that there is no variation in the data), the range, variance, and standard deviation will all be zero.
The range, variance, or standard deviation will always be greater than or equal to zero.
The range can be a deceptive measure of dispersion if there are extreme values in the
data set.
Here are some examples to help you understand the standard deviation.
Example #1: Distribution with Mean = 100 and Standard Deviation =10
Mean = 100; Standard Deviation = 10
In Figure 4.7 you see that the data is more spread out than in Figure 4.8 because the standard
deviation = 10.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 65
Example #2: Distribution with Mean = 100 and Standard Deviation = 1
In Figure 4.8 you see that the data is less spread than in Figure 4.7 because the standard
deviation = 1.
Example #3: Distribution with Mean = 100 and Standard Deviation = 0
Mean = 100, Standard Deviation = 0
Figure 4.7 Histogram for Example #1
Figure 4.8 Histogram for Example #2
From the Library of Pearson HED
66 A Guide to Six Sigma and Process Improvement for Practitioners and Students
In Figure 4.9 you see that the standard deviation = 0, so all the data are the same.
Figure 4.9 Histogram for Example #3
Measures of Shape
The third important property of data that we look at it is shape. The shape is the manner in
which the data are distributed. Either a histogram or a dot plot can be used to study the shape
of a distribution of data.
Skewness
Skewness is a measure of the size of the right or left tail of a unimodal (one hump) distribution. We examine three types of skewness: symmetrical, positive or right skewness, and
negative or left skewness.
Symmetrical
A symmetrical distribution arises when the mean, median, and mode are all equal, see
Figure 4.10 .
Positive or Right Skewness
Positive or right skewness occurs when the data has some unusually high values, which
causes the mean to be greater than the median as is seen in Figure 4.11 .
When data is skewed to the right the mean is larger than the median, and the median is larger
than the mode.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 67
Negative or Left Skewness
Negative or left skewness occurs when the data has some unusually low values, which causes
the mean to be less than the median as is seen in Figure 4.12 .
When data is skewed to the left the mode is larger than the median, and the median is larger
than the mean.
Mean
Median
Mode
X
Figure 4.10 Symmetrical distribution

Figure 4.11 Positive or right skewed distribution
From the Library of Pearson HED
68 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Figure 4.12 Negative or left skewed distribution
More on Interpreting the Standard Deviation
If the distribution of output from a process is unimodal, symmetric, and bell shaped, we call
this the
normal distribution. The normal distribution has some properties that are worth
mentioning here to improve your understanding of the standard deviation.
First, if you create a region under the normal distribution that is plus or minus one standard
deviation from the mean, then 68.26% of the output from the process that generated the
normal distribution will lie in that area. Figure 4.13 shows a normal distribution with a mean
of 100 and a standard deviation of 10. In this distribution 68.26% of the data will lie between
90 (Mean – 1 standard deviation = [100 – 10] = 90) and 110 (Mean + 1 standard deviation
= [100 + 10] = 110).
Figure 4.13 Mean +/- 1 standard deviation from the mean
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 69
Second, if you create a region under the normal distribution that is plus or minus two standard deviations from the mean, then 95.44% of the output from the process that generated
the normal distribution will lie in that area. Figure 4.14 shows a normal distribution with a
mean of 100 and a standard deviation of 10. In this distribution 95.44% of the data will lie
between 80 (Mean – 2 standard deviations = [100 – 20] = 80) and 120 (Mean + 2 standard
deviations = [100 + 20] = 120).
Figure 4.14 Mean +/- 2 standard deviations from the mean
Third, if you create a region under the normal distribution that is plus or minus three standard deviations from the mean, then 99.73% of the output from the process that generated
the normal distribution will lie in that area. Figure 4.15 shows a normal distribution with a
mean of 100 and a standard deviation of 10. In this distribution 99.73% of the data will lie
between 70 (Mean – 3 standard deviations = [100 – 30] = 70) and 130 (Mean + 3 standard
deviations = [100 + 30] = 130).
Statisticians calculated these probabilities a long time ago. It doesn’t matter what the mean
is or what the standard deviation is, the preceding probabilities apply.
From the Library of Pearson HED
70 A Guide to Six Sigma and Process Improvement for Practitioners and Students
How-To Guide for Understanding Data: Minitab 17
User Guide
Minitab is a statistical package designed to perform statistical analysis (Minitab 17, 2013).
Today, Minitab is used both in academia and in industry. In Minitab, you create and open
projects to store all your data and results. A session or log of activities, a Project Manager that
summarizes the project contents, and any worksheets or graphs used are the components that
form a project. Project components are displayed in separate windows inside the Minitab
application window. By default, you see only the session and one worksheet window when
you begin a new project in Minitab. (You can bring any window to the front by selecting the
desired window in the Minitab Windows menu.) You can open and save an entire project or,
as is done in this text, open and save worksheets.
Minitab’s statistical rigor, availability for many different types of computer systems, and
commercial acceptance makes this program a great tool for using statistics for quality
improvement.
Go to www.ftpress.com/sixsigma to download the data files referenced in this chapter so you
can practice with Minitab.
Using Minitab Worksheets
You use a Minitab worksheet (see Figure 4.16 ) to enter data for variables by column. Minitab
worksheets are organized as numbered rows and columns numbered in the form
Cn in which
C1 is the first column. You enter variable labels in a special unnumbered row that precedes
row 1. Unlike worksheets in programs such as Microsoft Excel, currently Minitab worksheets
do not accept formulas and do not automatically recalculate themselves when you change the
values of the supporting data.
Figure 4.15 Mean +/- 3 standard deviations from the mean
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 71
Figure 4.16 Minitab worksheet
By default, Minitab names opened worksheets serially in the form of Worksheet1, Worksheet2, and so on. Better names are ones that reflect the content of the worksheets, such as
CHEMICAL for a worksheet that contains data for the viscosity of a chemical. To give a sheet
a descriptive name, open the Project Manager window, right-click the icon for the worksheet,
select Rename from the shortcut menu, and type in the new name.
Opening and Saving Worksheets and Other Components
You open worksheets to use data that have been created by you or others at an earlier time.
To open a Minitab worksheet, first select
File | Open Worksheet. In the Open Worksheet
dialog box that appears (see Figure 4.17 ) perform the following steps:
1. Select the appropriate folder (also known as a directory) from the Look In drop-down
list box.
2. Check, and select if necessary, the proper Files of Type value from the drop-down
list at the bottom of the dialog box. Typically, you do not need to make this selection
as the default choice Minitab lists all Minitab worksheets. However, to list all project
From the Library of Pearson HED
72 A Guide to Six Sigma and Process Improvement for Practitioners and Students
files, select Minitab Project; to list all Microsoft Excel files, select Excel (*.xls); and to
list every file in the folder, select All.
3. If necessary, change the display of files in the central files list box by clicking the rightmost (View Menu) button on the top row of buttons and selecting the appropriate
view from the drop-down list.
4. Select the file to be opened from the files list box. If the file does not appear, verify that
steps 1, 2, and 3 were done correctly.
5. Click OK.
To open a Minitab Project that can include the session, worksheets, and graphs, select
Minitab Project in the previous step 2 or select the similar
File | Open Project. Individual
graphs can be opened as well by selecting
File | Open Graph.
Figure 4.17 Minitab Open Worksheet dialog box
You can save a worksheet individually to ensure its future availability, to protect yourself
against a system failure, or to later import it into another project. To save a worksheet, select
the worksheet’s window and then select
File | Save Current Worksheet As. In the Save
Worksheet As dialog box that appears (see Figure 4.18 ), perform these steps:
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 73
1. Select the appropriate folder from the Save In drop-down list box.
2. Check, and select if necessary, the proper Save As Type value from the drop-down
list at the bottom of the dialog box. Typically, you want to accept the default choice,
Minitab, but select Minitab Portable to use the data on a different type of computer
system or select an earlier version such as Minitab 13 to use the data in that earlier
version.
3. Enter (or edit) the name of the file in the File Name box.
4. Optionally, click the Description button and in the Worksheet Description dialog box
(not shown), enter documentary information and click
OK.
5. Click OK (in the Save Worksheet As dialog box).
Figure 4.18 Save Worksheet As dialog box
To save a Minitab Project, select the similar File | Save Project As. The Save Project As dialog
box (not shown) contains an Options button that displays the Save Project – Options dialog
box in which you can indicate which project components other than worksheets (session,
dialog settings, graphs, and Project Manager content) will be saved.
Individual graphs and the session can also be saved separately by first selecting their windows
and then selecting the similar
File | Save Graph As or File | Save Session As, as appropriate.
From the Library of Pearson HED
74 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Minitab graphs can be saved in either a Minitab graph format or any one of several common
graphics formats, and Session files can be saved as simple or formatted text files.
You can repeat a save procedure and save a worksheet, project, or other component using a
second name as an easy way to create a backup copy that can be used should some problem
make your original file unusable.
Obtaining a Bar Chart
To obtain a bar chart, open the KEYBOARD worksheet. Select Graph | Bar Chart and then
follow these steps:
1. In the Bar Charts dialog box (see Figure 4.19 ), select Simple. Click OK.
Figure 4.19 Bar Charts dialog box
2. In the Bar Chart dialog box (see Figure 4.20 ), enter C2 or Cause in the Categorical
Variables edit box.
3. Select the Data Options button. In the Bar Chart: Data Options dialog box (see Figure
4.21 ), select the
Frequency tab since the data is in the form of frequencies in prespecified classes. Enter C2 or Frequency in the Frequency Variable(s) edit box. Click
OK to return to the Bar Chart dialog box. Click OK.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 75
Figure 4.21 Bar Chart: Data Options dialog box
To select colors for the bars and borders in the bar chart, follow these steps:
1. Right-click on any of the bars of the bar chart.
2. Select Edit Bars.
Figure 4.20 Bar Chart data source dialog box
From the Library of Pearson HED
76 A Guide to Six Sigma and Process Improvement for Practitioners and Students
3. In the Attributes tab of the Edit Bars dialog box, enter selections for Fill pattern and
Border and Fill Lines.
Figure 4.22 shows the output for the bar chart.
Figure 4.22 Minitab output for the bar chart
Obtaining a Pareto Diagram
To obtain a Pareto diagram, open the KEYBOARD.MTW worksheet. Note that this data set
contains the causes of the defects in column C1 and the frequency of defects in column C2.
Select
Stat | Quality Tools | Pareto Chart; see Figure 4.23 for the dialog box. Follow these
steps:
1. In the Defects or Attribute Data In edit box, enter C1 or Cause.
2. In the Frequencies In edit box, enter C2 or Frequency.
3.
In the Combine Remaining Defects into One Category After This Percent edit box,
enter
99.9.
4. Click OK.
If the variable of interest was located in a single column and is in raw form with each row
indicating a type of error, the Charts Defects Data In option button would be selected and the
appropriate column number or variable name would be entered in the Defects or Attribute
Data In edit box.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 77
To select colors for the bars and borders in the Pareto diagram, perform these steps:
1. Right-click on any of the bars of the Pareto diagram.
2. Select Edit Bars.
3. In the Attributes tab of the Edit Bars dialog box, enter selections for Fill pattern and
Border and Fill Lines.
Figure 4.24 shows the output for the Pareto diagram.
Figure 4.24 Minitab output for the Pareto diagram
Figure 4.23 Pareto Chart dialog box
From the Library of Pearson HED
78 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Obtaining a Line Graph (Time Series Plot)
To obtain a line graph, open the TRANSMIT worksheet. To obtain a run chart of the percentage of errors, follow these steps:
1. Enter the label Error% in column C3.
2. Select Calc | Calculator.
3. In the Calculator dialog box (see Figure 4.25 ), enter C3 or Error% in the Store Result
in Variable edit box. To obtain the percentage of errors, enter
‘Errors’ / 125 in the
Expression edit box. Click
OK.
Figure 4.25 Calculator dialog box
4. Select Graph | Time Series Plot.
5. Select Simple, click OK.
6.
In the Series edit box, enter C3 or ‘Error%’ (see Figure 4.26 ).
7. Click OK.
Figure 4.27 shows the output for the Time Series Plot.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 79
Figure 4.27 Minitab output for the Time Series Plot
Obtaining a Histogram
To obtain a histogram, open the CHEMICAL.MTW worksheet. Select Graph | Histogram
and follow these steps:
Figure 4.26 Time Series Plot: Simple data options dialog box
From the Library of Pearson HED
80 A Guide to Six Sigma and Process Improvement for Practitioners and Students
1. In the Histograms dialog box (see Figure 4.28 ), select Simple. Click OK.
Figure 4.28 Histograms dialog box
2. In the Histogram data source dialog box (see Figure 4.29 ), enter C2 or Viscosity in
the Graph Variables edit box. To obtain reference lines, select the
Scale button.
Figure 4.29 Histogram data source dialog box
3. In the Histogram: Scale dialog box (see Figure 4.30 ), select the Reference Lines tab.
Enter
13 and 18 in the Show Reference Lines at Y Values box. Click OK to return to
the Histogram data source dialog box. Click
OK to obtain the histogram.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 81
Figure 4.30 Histogram: Scale dialog box
To select colors for the bars and borders in the histogram, do the following:
1. Right-click on any of the bars of the histogram.
2. Select Edit Bars.
3. In the Attributes tab of the Edit Bars dialog box, enter selections for Fill pattern and
Border and Fill Lines.
Figure 4.31 shows the output for the histogram.
From the Library of Pearson HED
82 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 4.31 Minitab output for the histogram
Obtaining a Dot Plot
To obtain a dot plot using Minitab, open the CHEMICAL.MTW worksheet. Select Graph |
Dotplot
, and then do the following:
1. In the Dotplots dialog box (see Figure 4.32 ), select the One Y Simple choice. If dot
plots of more than one group are desired, select the
One Y With Groups Choice.
Figure 4.32 Dotplots dialog box
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 83
2. In the Dotplot data source dialog box (see Figure 4.33 ) in the Graph Variables edit
box enter
C2 or Viscosity. Click OK.
Figure 4.33 Dotplot data source dialog box
Figure 4.34 shows the output for the dot plot.
Figure 4.34 Minitab output for the dot plot
From the Library of Pearson HED
84 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Obtaining a Run Chart
To obtain a run chart, open the TRANSMIT worksheet. To obtain a run chart of the percentage of errors, follow these steps:
1. Enter the label Error% in column C3.
2. Select Calc | Calculator.
3. In the Calculator dialog box (see Figure 4.35 ), enter C3 or Error% in the Store Result
in Variable edit box. To obtain the percentage of errors, enter
‘Errors’ / 125 in the
Expression edit box. Click
OK.
Figure 4.35 Calculator dialog box
4. Select Stat | Quality Tools | Run Chart.
5. In the Run Chart dialog box (see Figure 4.36 ), select the Single Column option button. Enter C3 or ‘Error%’ in the edit box. Enter C1 or Day in the Subgroup Size edit
box. Click
OK.
Figure 4.37 shows the output for the run chart.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 85
Figure 4.37 Minitab output for the run chart
Obtaining Descriptive Statistics
To obtain descriptive statistics for the viscosity of the chemical data set, open the
CHEMICAL.MTW worksheet. Select Stat | Basic Statistics | Display Descriptive Statistics
and follow these steps:
Figure 4.36 Run Chart dialog box
From the Library of Pearson HED
86 A Guide to Six Sigma and Process Improvement for Practitioners and Students
1. In the Display Descriptive Statistics dialog box (see Figure 4.38 ), enter C2 or Viscosity
in the Variables edit box.
Figure 4.38 Display Descriptive Statistics dialog box
2. Select the Statistics button. In the Display Descriptive Statistics: Statistics dialog box
(see Figure 4.39 ), select the
Mean, Standard Deviation, First Quartile, Median,
Third Quartile, Minimum, Maximum, Range, Skewness, and N Total (the sample
size) check boxes (see Figure 4.39 ). Click
OK to return to the Display Descriptive
Statistics dialog box. Click
OK again to obtain the descriptive statistics.
Figure 4.39 Display Descriptive Statistics: Statistics dialog box
Figure 4.40 shows the output for Descriptive Statistics.
From the Library of Pearson HED
Chapter 4 Understanding Data: Tools and Methods 87
Figure 4.40 Minitab output for descriptive statistics
Takeaways from This Chapter
When understanding data for process improvement, you need to be familiar with two
types: attribute and measurement.
Different graphs are used for displaying each type of data.
For attribute data:
Bar charts graphically display the frequency or percentages of items that fall into
two or more categories.
Pareto diagrams help you distinguish between the critical few and the trivial
many categories data exhibits.
Line graphs are used to graphically display attribute data over time.
For measurement data:
Histograms help you understand the distribution of the data.
Dot plots are similar to histograms, but you use them when you have smaller
data sets.
Run charts are used to graphically display measurement data over time to spot
trends.
Various measures of central tendency help you find a single value that best represents
a distribution of data.
Various measures of variation help you find a single number that best represents the
spread of the data.
From the Library of Pearson HED
88 A Guide to Six Sigma and Process Improvement for Practitioners and Students
References
Minitab Version 17 (State College, PA: Minitab 2013).
Additional Readings
Berenson, M. L. and D. M. Levine (2014), Basic Business Statistics, 13th ed. (Upper Saddle
River, NJ: Prentice Hall).
Gitlow, H., A. Oppenheim, R. Oppenheim, and D. Levine (2015),
Quality Management: Tools
and Methods for Improvement
, 4th ed. (Naperville, IL: Hercher Publishing Company).
This book is free online at hercherpublishing.com.
Gitlow, H. and D. Levine (2004),
Six Sigma for Green Belts and Champions: Foundations,
DMAIC, Tools and Methods, Cases and Certification
(Upper Saddle River, NJ: Prentice
Hall).
From the Library of Pearson HED
89
5
Understanding Variation: Tools and Methods
What Are the Objectives of This Chapter?
This chapter has two objectives. The first objective is to introduce you to statistical control
charts that help you to understand when and where special causes of variation are occurring in a process, and to understand when a process is exhibiting only common variation. It
also explains how to reduce common causes of variation. The second objective is to explain
measurement systems analysis; that is, how do you know if your data has too much variation
(noise) in it to be of practical value in improving a process?
The chapter is split into two sections. The first section is a high level overview of the tools and
methods used in understanding variation by looking at what each tool and method is, why
each tool and method is used, and examples of each tool and method. The second section
shows you how to use Minitab to utilize the different tools and methods that you learned
about in the first section.
What Is Variation?
Recall from Chapter 2 , “Process and Quality Fundamentals,” that all processes have outputs
and these outputs may be measured. The distribution of these measurements varies, and the
differences between these measurements are called
process variation.
Recall also that there are two types of variation: common cause variation and special cause
variation (Gitlow et al., 2015; Gitlow and Levine, 2004).
Common Cause Variation
Common causes of variation create fluctuations or patterns in data that are due to the system
itself—for example, the fluctuations (variation) caused by hiring, training, and supervisory
policies and practices. Common causes of variation are the responsibility of management.
Only management can change the policies and procedures that define the common causes
of variation inherent in a system.
From the Library of Pearson HED
90 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Special Cause Variation
Special causes of variation create fluctuations or patterns in data that are not inherent to a
process; they come from outside the process. Special causes of variation are the responsibility of workers and engineers. Workers and engineers identify, and if possible, resolve special
causes of variation. If they cannot resolve a special cause of variation, they enlist the aid of
management.
Using Control Charts to Understand Variation
The control chart is a tool for distinguishing between the common and special causes of
variation for a variable (Gitlow et al., 2015; Gitlow and Levine, 2004). They are used to assess
and monitor the stability of a variable (presence of only common causes of variation). The
data for a control chart is obtained from a subgroup or sample of items selected at each
observation session, for example, by month, by patient, or by form.
There are two uses for control charts. The first one we discussed, that is, to distinguish special
from common causes of variation. This happens on the factory floor, in a call center, and so
on. The second one occurs once data is aggregated, say over several areas in an organization
(for example, several sales teams), the control chart loses the ability to detect special causes
of variation. Consequently, they are primarily used to stop management from overreacting
to common causes of variation and treating common causes of variation as special causes
of variation, and making the process more complex than is necessary. Think of it like this:
Overreaction to common causes of variation can make a straw look like a bowl of spaghetti.
Instead of it being a straight line from point A to point B in a process, it becomes a maze
creating problems like increased waiting times, increased cycle times, and increased defects,
to name a few problems.
The most common types of control charts can be divided into two categories determined by
the type of data used to monitor a process. These two broad categories are
attribute control
charts
and variables control charts.
Attribute Control Charts
Attribute control charts are used when you need to evaluate variables defined by attribute
data. Either classification data (for example, defective or conforming) or count data (for
example, number of defects per area of opportunity).
The attribute control charts covered in this chapter are
P charts—For classification data with either equal or unequal subgroup size
C charts—For count data with consistent areas of opportunity
U charts—For count data with variable areas of opportunity
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 91
Variables Control Charts
Variables control charts are used when you need to evaluate variables measured on a continuous scale such as height, weight, temperature, cycle time, waiting time, revenue, costs,
and so on. Variables control charts contain two sections; one studies process variability while
the other studies process central tendency. When dealing with variables control charts we
look at the subgroup size to determine which type of control chart to use. A subgroup is the
group of measurements that make up each point that is plotted on the control chart.
The variables control charts covered in this chapter are
Individuals and moving range (I-MR) charts—For subgroups of 1
X Bar and R charts—For subgroups whose size is between 2 and 9
X Bar and S charts—For subgroups whose size is greater than or equal to 10
An Anecdote of Using Control Charts

I was consulting in a factory that had just started to require statistical evidence of qual
ity from their suppliers. Top management had just completed a course in statistical
process control, but certainly were not experts in the field.
I was invited to a meeting with a supplier that presented the control charts of their
products. The top management was impressed by the display of control charts. How
ever, I pointed out that all the control charts were out of control (they showed special
causes of variation) and were not producing within specification limits.
The moral to the story is that control charts do not equal quality. No amount of train
ing can make people understand control charts; they must study them themselves, with
the help of a master, to really understand them. The proper long-term use of control
charts produces quality products.

Understanding Control Charts
Again, the principal focus of the control chart on the shop floor is the attempt to separate
special or assignable causes of variation from random noise or common causes of variation.
The distinction between the two causes of variation is crucial because special causes of variation are those that are not part of a process and are correctable or exploitable by workers, or
sometimes they need the help of managers. Common causes of variation can be reduced only
by changing the system. Only management can change the system or enlightened management can empower the associates to work on the system as long as the aim of management
and needs of customers are clear.
Control charts allow you to monitor the process and determine the presence of special causes
of variation. Control charts help prevent two types of errors: type one errors and type two
errors (Gitlow et al., 2015):
From the Library of Pearson HED
92 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Type one errors—Involve the belief that an observed value represents special cause
variation when in fact it is due to the common cause variation of the system. Treating
common causes of variation as special cause variation can result in tampering with or
overadjustment of a process with an accompanying increase in variation. This is the
action discussed previously, which causes the process flow from point A to point B to
look like a bowl of spaghetti instead of a straw, with all the ensuing problems.
Type two errors—Involve treating special cause variation as if it is common cause
variation and not taking corrective action when is necessary.
All control charts have a common structure:
Measurement of a process characteristic, which is plotted in order of time, or some or
the variable, for example, employee or location.
A central line that represents the process average.
Upper (UCL) ,and lower (LCL) control limits that provide information on the process variation. Control limits are calculated by adding and subtracting three standard
deviations of the statistic of interest from the process mean.
Figure 5.1 shows an example of a control chart that exhibits only common causes of variation.
Figure 5.1 Control chart example
Once these control limits are computed from the data, the control chart is evaluated by
discerning any nonrandom pattern that might exist in the data. Figure 5.2 illustrates three
different patterns.
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 93
Figure 5.2 Three control chart patterns
In panel A of Figure 5.2 , there does not appear to be any pattern in the ordering of values
over time, and no points fall outside the three standard deviation control limits. It appears
that the process is stable and contains only common cause variation.
Panel B, on the contrary, contains two points that fall outside the three standard deviation
control limits. Each of these points should be investigated to determine the special causes
that led to their occurrence. This process is not in control.
Although Panel C does not have any points outside the control limits, it has a series of consecutive points above the average value (the center line), as well as a series of consecutive
points below the average value. In addition, a long-term overall downward trend in the value
of the variable is clearly visible. This process is also not in control. We will explain why later
in this book.
Control limits are often called
three-sigma limits. In practice, virtually all of a process’s output is located within a three-sigma interval of the process mean, provided that the process is
stable; that is, no special causes of variation are present in the process.
Rules for Determining Out of Control Points
As you saw earlier, the simplest rule for detecting the presence of a special cause is one or
more points falling beyond the mean plus or minus 3 standard deviation limits on the chart.
The control chart can be made more sensitive and effective in detecting out of control points
by considering other signals and patterns unlikely to occur by chance alone. For example, if
only common causes are operating, you would expect the points plotted to approximately
follow a bell-shaped normal distribution.
Figure 5.3 presents a control chart in which the area between the upper and lower control limits is subdivided into bands, each of which is one standard deviation wide. These
From the Library of Pearson HED
94 A Guide to Six Sigma and Process Improvement for Practitioners and Students
additional limits or zone boundaries can be useful in detecting other unlikely patterns of
data points.
UCL
CL
LCL
Time
MEAN

+2σ to +3σ band ZONE A
+1σ to +2σ band ZONE B
Mean to +1σ band ZONE C
Mean to -1σ band ZONE C
-1σ to -2σ band ZONE B
-2σ to -3σ band ZONE A

Figure 5.3 A control chart showing bands, each of which is one standard deviation wide
In a stable process, one would expect 68.26% of all the statistics to be between the upper and
lower C bands, 95.44% of all the statistics to be between the upper and lower B bands, and
99.73% of all the statistics to be between the upper and lower A zones. This means that 0.27%
of the statistics would be beyond the upper and lower A zones; 0.135% above upper zone A
and 0.135% below lower zone A.
You can conclude that the process is out of control if any of the following events occur (Gitlow et al., 2015; Gitlow and Levine, 2004):
Rule 1—A point falls above the UCL or below the LCL. If a point falls above the
upper zone A limit then one of two things occurred: either a special cause occurred
or a common cause occurred with the likelihood of 0.135% or 1,350 times per million
opportunities. This is considered so unlikely by statisticians that they are willing to
gamble the data point is a special cause of variation and act accordingly. See Figure
5.4 for an illustration of Rule 1.
Rule 2—Nine or more consecutive points lie above the center line or nine or
more consecutive points lie below the center line. See Figure 5.5 for an illustration of
Rule 2.
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 95
Figure 5.5 Illustration of Rule 2
Rule 3—Six or more consecutive points move upward in value or six or more consecutive points move downward in value. See Figure 5.6 for an illustration of Rule 3.
Rule 4—An unusually small number of consecutive points above and below the center
line are present (a saw tooth pattern). Rule 4 is used to determine whether a process
is unusually noisy (high variability) or unusually quiet (low variability). Figure 5.7
illustrates Rule 4.
Figure 5.4 Illustration of Rule 1
From the Library of Pearson HED
96 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 5.7 Illustration of Rule 4
Rule 5—Two out of three consecutive points fall in the high Zone A or above, or in
the low Zone A or below. See Figure 5.8 for an illustration of Rule 5.
Rule 6—Four out of five consecutive points fall in the high Zone B or above, or in the
low Zone B or below. See Figure 5.9 for an illustration of Rule 6.
Figure 5.6 Illustration of Rule 3
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 97
Figure 5.9 Illustration of Rule 6
Rule 7—Fifteen consecutive points fall within Zone C on either side of the center line.
Rule 7 is used to determine whether a process is unusually noisy (high variability) or
unusually quiet (low variability). See Figure 5.10 for an illustration of Rule 7.
If only common causes are operating, each of the preceding seven patterns is statistically
unlikely to occur. The presence of one or more of these low probability events indicates
that one or more special causes may be operating, thereby resulting in a process that is out
of a state of statistical control. Other rules for special causes have been developed and are
incorporated within the control chart features of Minitab. Different rules may be considered
appropriate for specific charts.
Figure 5.8 Illustration of Rule 5
From the Library of Pearson HED
98 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 5.10 Illustration of Rule 7 (Source: minitab.com)
Control Charts for Attribute Data
Some control charts are specifically designed to operate on attribute data; they are p charts,
c charts, and u charts. Each chart has a specific purpose that will be discussed in this section,
and later in more detail.
P Charts
What: The p chart is a control chart used to study classification type attribute data, for
example, the proportion of defective (nonconforming) items by month. An item is defective
if it is not fit for use, for example, a water bottle with a hole in it. Subgroup sizes in a p chart
may remain constant or may vary. A p chart may be used to help process improvers control
defective versus conforming, go versus no-go, or acceptable versus not acceptable outputs
from a process. Using the seven rules discussed previously, the control chart can be used to
distinguish special from common causes of variation.
The p chart is used when
There are only two possible outcomes for an event. An item must be found to be either
conforming or nonconforming (defective).
The probability, p, of a nonconforming item is constant over time.
Successive items are independent over time.
Subgroups are of sufficient size to detect an out of control event. A general rule for
subgroup size for a p chart is that the subgroup size should be large enough to detect
a special cause of variation if it exists. Frequently, subgroup sizes are between 50 and
500, per subgroup, depending on the metric being studied.
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 99
Subgroup frequency, how often you draw a subgroup from the process under study,
should be often enough to detect changes in the process under study. This requires
expertise in the process under study. If the process can change very quickly, more frequent sampling is needed to detect special causes of variation. If the process changes
slowly, less frequent sampling is needed to detect a special cause of variation.
Why: P charts are used to monitor stability of a process that generates classification attribute
data; that is, for example, the percentage of process output that is defective by time period.
Example 1: P Chart with Equal Subgroup Sizes
As an illustration of the p chart, consider the case of a large health system that has had
complaints from several pathologists concerning a supposed problem with a manufacturer
sending cracked slides. Slides are pieces of glass that pathologists place biopsies on so they
can be examined under a microscope. This allows the pathologist to assist the clinician in
confirming a diagnosis of the disease from the biopsy. According to the pathologists, some
slides are cracked or broken before or during transit, rendering them useless scrap. The
fraction of cracked or broken slides is naturally of concern to the hospital administration.
Each day a sample of 100 slides is drawn from the total of all slides received from each slide
vendor. Table 5.1 presents the sample results for 30 days of incoming shipments for a particular vendor.
Table 5.1 30 Days of Incoming Slide Shipments for a Particular Vendor

Day Sample Size Number
Cracked
1 100 14
2 100 2
3 100 11
4 100 4
5 100 9
6 100 7
7 100 4
8 100 6
9 100 3
10 100 2
11 100 3
12 100 8
13 100 4
14 100 15
15 100 5

 

Day Sample Size Number
Cracked
16 100 3
17 100 8
18 100 4
19 100 2
20 100 5
21 100 5
22 100 7
23 100 9
24 100 1
25 100 3
26 100 12
27 100 9
28 100 3
29 100 6
30 100 9

From the Library of Pearson HED
100 A Guide to Six Sigma and Process Improvement for Practitioners and Students
These data are appropriate for a p chart because each slide is classified as cracked (defective)
or not cracked (conforming), the probability of a cracked slide is assumed to be constant
from slide to slide, and each slide is considered independent of the other slides.
Figure 5.11 illustrates the p chart obtained from Minitab for the cracked slides data.
Figure 5.11 P chart for cracked slides
From Figure 5.11 , we find a process that lacks control. On day 1, the proportion of cracked
slides (14/100 = 0.14) is above the upper control limit and on day 14 the proportion of
cracked slides (15/100 = 0.15) is above the upper control limit. None of the other rules for
determining whether a process is out of control seems to be violated. That is, there are no
instances when two out of three consecutive points lie in zone A on one side of the center
line; there are no instances when four out of five consecutive points lie in zone B or beyond
on one side of the center line; there are no instances when six consecutive points move
upward or downward; nor are there nine consecutive points on one side of the center line.
Upon examination of the out of control points it was found that for those two shipments the
boxes containing the slides were overpacked. Once this was realized, the vendor was contacted and the pathology department was assured that it would never happen again due to a
changed protocol. Consequently, the two out of control data points were dropped, and the
control chart was recomputed. The revised control chart is shown in Figure 5.12 .
Now that the process is stable, a Pareto analysis is done on the causes of cracked tiles. Figure
5.13 shows that the major cause of cracked tiles is improper wrapping by the vendor. The
vendor was notified and changed the wrapping protocol.
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 101
Figure 5.13 Pareto analysis of causes of cracked tiles
Figure 5.14 shows a before and after control chart of the cracked tile problem.
As, you can see from Figure 5.14 , the common causes of variation were dramatically reduced
due to good communication between the supplier and the pathology department.
Figure 5.12 Revised control chart without special cause subgroups
From the Library of Pearson HED
102 A Guide to Six Sigma and Process Improvement for Practitioners and Students
NOTE
Many statisticians believe that the Minitab zone rules are overly conservative. For
example, some statisticians would argue that eight points in a row above or below
the average is a special case of variation, while Minitab says nine points in a row
indicates a special cause of variation; see the Western Electric handbook for more
information on this point.
Example 2: P Chart with Unequal Subgroup Sizes
In many instances, unlike the example concerning the slides, the subgroup size varies from
subgroup to subgroup. Consider the case of a hospital that has seen an increase in chemotherapy turnaround times. The pharmacy director suspects the increased turnaround times
are due to various delays in the process. Based on carefully laid out operational definitions on
what constitutes a delay, each day for 60 consecutive days data is collected on the proportion
of orders with delays and is presented in Table 5.2 .
Figure 5.14 Before and after control chart
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 103
Table 5.2 Delayed Orders for 60 Consecutive Days

Day Delayed
Orders
Total
Orders
1 21 94
2 22 82
3 25 95
4 25 86
5 26 106
6 11 109
7 21 113
8 20 112
9 23 86
10 12 112
11 11 99
12 20 97
13 19 104
14 16 129
15 22 91
16 19 124
17 26 106
18 16 105
19 24 89
20 16 79
21 20 104
22 25 97
23 20 96
24 24 79
25 20 115
26 17 95
27 20 104
28 21 112
29 14 100
30 20 103

 

Day Delayed
Orders
Total
Orders
31 22 91
32 21 105
33 24 107
34 20 107
35 22 100
36 21 96
37 19 123
38 15 119
39 21 126
40 18 95
41 17 147
42 24 105
43 18 96
44 21 103
45 15 87
46 21 99
47 23 83
48 17 91
49 7 74
50 19 113
51 18 120
52 21 74
53 16 107
54 23 90
55 22 116
56 20 123
57 13 90
58 20 95
59 21 100
60 21 93

From the Library of Pearson HED
104 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 5.15 shows the p chart based on the delays data. Due to the different subgroup sizes,
the values obtained from Minitab for the upper and lower control limits and the zones are
different for each subgroup. None of the points are outside the control limits and none of
the rules concerning consecutive points have been violated. Thus, any improvement in the
number of order delays must come from management changing the system.
Figure 5.15 P chart for orders with delays
C Charts
What: When there are multiple opportunities for defects or imperfections in a given continuous unit (such as a large sheet of fabric, a hospital ward, a stretch of highway, a refrigerator,
an automobile, a factory with a constant number of employees, to name a few examples),
each unit is called an area of opportunity; each area of opportunity is a subgroup. The c chart
is used when the areas of opportunity are of constant size.
A defective item is a nonconforming unit. It must be discarded, reworked, returned, sold,
scrapped, or downgraded. It is unusable for its intended purpose in its present form. A defect,
however, is an imperfection of some type that does not necessarily render the entire item
unusable, yet is undesirable. One or more defects may not make an entire item defective. An
assembled piece of machinery such as a car, dishwasher, or air conditioner may have one or
more defects that may not render the entire item defective but may cause it to be downgraded
or may necessitate its being reworked, for example a scratch, or a dent.
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 105
The c chart is used when
Areas of opportunity are of constant size.
Subgroups should be of sufficient size to detect an out of control event. A general
rule for subgroup size for a c chart is that the subgroup size should be large enough
to detect a special cause of variation if it exists. This is operationally defined as the
average number of defects per area of opportunity being at least 20.0.
The frequency of collected subgroups should be often enough to detect changes in
the process under study. This requires expertise in the process under study. If the
process can change quickly, more frequent sampling is needed to detect special causes
of variation. If the process changes slowly, less frequent sampling is needed to detect
a special cause of variation.
Why: C charts are used to monitor stability of count attribute data where the area of opportunity is constant from observation to observation.
Example: To illustrate the c chart, consider the number of add-ons (unscheduled patients) in
an outpatient clinic in a hospital. Many times patients are added on to the regular schedule
at the last minute. This is problematic as capacity of the clinic is limited and add-ons create
problematic wait times, usually too long for patient satisfaction to be at an acceptable level.
A process improvement team collected data on the number of add-ons per day at one of its
outpatient clinics. Results of these data collections produce the results in Table 5.3 .
Table 5.3 Number of Add-Ons Per Day for 50 Consecutive Days in an Outpatient Clinic

Day Add-ons
1 13
2 19
3 9
4 21
5 18
6 13
7 18
8 19
9 19
10 15
11 11
12 9
13 16
14 12

 

Day Add-ons
15 16
16 12
17 14
18 14
19 14
20 13
21 14
22 11
23 17
24 11
25 15
26 21
27 16
28 15

 

Day Add-ons
29 18
30 19
31 10
32 14
33 14
34 12
35 16
36 12
37 8
38 12
39 20
40 24
41 11
42 10

 

Day Add-ons
43 10
44 5
45 17
46 14
47 11
48 23
49 13
50 8

From the Library of Pearson HED
106 A Guide to Six Sigma and Process Improvement for Practitioners and Students
The assumptions necessary for using the c chart are well met here, as the clinic days are
considered to be the areas of opportunity; add-ons are discrete events and seem to be independent of one another. Even if these conditions are not precisely met, the c chart is fairly
robust, or insensitive to small departures from the assumptions.
From Figure 5.16 , the number of add-ons appears to be stable around a center line or mean
of 14.32. None of the add-ons are outside the control limits, and none of the rules concerning
zones have been violated.
Figure 5.16 C chart of add-ons
U Charts
What: The u chart is similar to the c chart in that it is a control chart for the count of the
number of defects in a given area of opportunity. The fundamental difference between a
c chart and a u chart lies in the fact that during construction of a c chart, the area of opportunity remains constant from observation to observation, while this is not a requirement for
the u chart. Instead, the u chart considers the number of events (such as number of defects)
as a fraction of the total size of the area of opportunity in which these events were possible;
for example, the number of accidents in a factory where the number of employees changes
dramatically from time period to time period. The characteristic used for the control chart,
U
i, is the ratio of the number of events in the ith subgroup to the area of opportunity in the
i
th subgroup.
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 107
The u-chart is used when
You are considering the number of events (such as defects) as a fraction of the total
size of the area of opportunity in which these events were possible, thus circumventing the problem of having different areas of opportunity for different observations.
Subgroups are of sufficient size to detect an out of control event. A general rule for
subgroup size for a u chart is that the subgroup size should be large enough to detect
a special cause of variation if it exists.
The frequency with which subgroups are selected for study should be often enough
to detect changes to the process under study. This requires expertise in the process
under study. If the process can change quickly, more frequent sampling is needed to
detect special causes of variation. If the process changes slowly, less frequent sampling
is needed to detect a special cause of variation.
Why: U charts are used to monitor stability of count attribute data where the area of opportunity is not constant from subgroup to subgroup.
Example: To illustrate a u chart, consider the case of patient falls in a hospital. A large metropolitan hospital wants to examine the number of patient falls to see whether they are a
problem. Since the census (the number of patients in the hospital) is constantly changing
from day to day the u chart should be used. Table 5.4 shows the data on the number of patient
falls for the past 30 days.
Table 5.4 Patient Falls in a Hospital (Note: One patient may fall multiple times in a day.)

Day Census Falls
1 978 6
2 1040 5
3 1101 7
4 990 3
5 956 8
6 1004 3
7 1025 10
8 999 5
9 1013 7
10 1045 4
11 994 2
12 1105 4
13 1043 6
14 987 2
15 992 4

 

Day Census Falls
16 1009 7
17 976 2
18 998 1
19 1016 8
20 987 5
21 1092 3
22 1056 6
23 983 2
24 1025 5
25 987 7
26 1012 3
27 1091 7
28 1034 4
29 1024 8
30 1011 3

From the Library of Pearson HED
108 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 5.17 illustrates the u chart obtained from Minitab for the patient falls data.
Figure 5.17 U chart of patient falls in a hospital
From Figure 5.17 , the number of defects per unit (falls per day) appears to be stable around
a center line or mean of .00481. No points indicate a lack of control, so there is no reason to
believe that any special variation is present. If sources of special variation were detected, we
would identify the source or sources of the special variation, eliminate them from the system
if detrimental, or incorporate them into the system if beneficial; drop the data points from
the data set; and reconstruct and reanalyze the control chart.
Control Charts for Measurement Data
Variables control charts are used to study a metric from a process when the metric is measurement data, for example, height, area, temperature, cost, revenue, cycle time, service time,
or waiting time (Gitlow et al., 2015; Gitlow and Levine, 2004). Variables charts are typically
used in pairs. One chart studies the variation in a process, while the other chart studies the
process average. The chart that studies process variability must be examined before the chart
that studies the process average. This is because the chart that studies the process average
assumes that the process variability is stable.
It is important to note that once the percentage of defective items or the number of defects
per item gets low, attribute charts become increasingly less useful in improving a process.
This is because as the percentage of defectives or the number of defects gets small, the sample size required to find a defective item or items with almost no defects gets so large that
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 109
attribute charts become not useful. For example, if a p chart is used on a metric that has one
defective per 10,000 items, the sample size necessary to find one defect on average is 10,000.
As you can see, this creates a situation that is not practical. In this case process improvers
switch from attribute charts to variables control charts that measure how far off nominal an
item is, even if it is conforming. This means moving from the goal post view of quality to the
Taguchi Loss Function view of quality.
Individuals and Moving Range (I-MR) Charts
What: Often a situation occurs where only a single value is observed per subgroup. Perhaps measurements are destructive and/or expensive; or perhaps they represent a single
batch where only one measurement is appropriate, such as the total yield of a homogeneous
chemical batch process; or the measurements are monthly or quarterly revenue or cost data.
Whatever the case, there are circumstances when data must be taken as individual units that
cannot be conveniently divided into subgroups. In this case I-MR charts are used.
Individuals and moving range charts have two parts, one that charts the process variability
and the other that charts the process average. The two parts are used in tandem. Stability
must first be established in the portion charting the variability because the estimate of the
process variability provides the basis for the control limits of the portion charting the process
average.
Single measurements of variables are considered a subgroup of size one (n=1 per subgroup).
Hence, there is no variability within the subgroups themselves. An estimate of the process
variability must be made in some other way. The estimate of variability used for individual
value charts is based on the point-to-point variation in the sequence of single values, measured by the moving range (the absolute value of the difference between each data point and
the one that immediately preceded it). For example, the first moving range is the absolute
value of the difference between the first subgroup and the second subgroup, the second
moving range is the absolute value of the difference between the second subgroup and the
third subgroup, and so on. Consequently, there will always be one less moving range than
individual values.
As before, subgroup frequency should be often enough to detect changes in the process under
study. This requires expertise in the process under study. If the process can change quickly,
more frequent sampling is needed to detect special causes of variation. If the process changes
slowly, less frequent sampling is needed to detect a special cause of variation.
Why: I-MR charts are used to monitor process control and stability of measurement data
where subgroup size is equal to 1.
Example: To illustrate the individual value chart, consider a pathology department of a
hospital that needs to provide diagnosis of GI biopsies in a timely manner. There are many
steps in the process including extracting the biopsy from the patient, lab processing, reading
and sign-out by the faculty. The hospital administration is concerned with the turnaround
times for GI biopsies taking too long. Table 5.5 shows turnaround times for 100 consecutive
GI biopsy cases.
From the Library of Pearson HED
110 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Table 5.5 Turnaround Times for 100 Consecutive GI Biopsy Cases

Case GI Biopsy Turnaround
Times in Days
1 4.3
2 5.1
3 2.9
4 3.3
5 4.4
6 3.8
7 5.8
8 3.7
9 4.3
10 5.2
11 4.2
12 3.6
13 4
14 2.5
15 5.3
16 4.9
17 4.8
18 6.4
19 4.2
20 3.4
21 4
22 4.2
23 3.9
24 2.2
25 5.5
26 4.6
27 2.8
28 2.6
29 3.6
30 4
31 6

 

Case GI Biopsy Turnaround
Times in Days
32 5.4
33 2.6
34 4.5
35 3.1
36 4.8
37 4.1
38 3.1
39 3.3
40 4.3
41 2.9
42 2.4
43 2.4
44 4.5
45 3.4
46 5
47 2.9
48 4.7
49 3.9
50 4.2
51 2.2
52 4.6
53 2.4
54 5.5
55 4.1
56 4.6
57 4.7
58 3.7
59 4.6
60 3.6
61 3.8
62 4.2

From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 111

Case GI Biopsy Turnaround
Times in Days
63 5.4
64 4.8
65 5.9
66 3.9
67 3.9
68 4
69 4.9
70 3.6
71 4
72 2.6
73 3.2
74 5.6
75 4.4
76 5.1
77 5.8
78 4.4
79 4.1
80 2.5
81 1.9
82 3.6
83 4.9
84 3.8
85 4.5
86 5.2
87 5
88 1.8
89 4.8
90 3.8
91 5.1
92 3.6
93 5.3
94 3.6

 

Case GI Biopsy Turnaround
Times in Days
95 4.3
96 4.3
97 3.7
98 2.9
99 2
100 4.5

From the Library of Pearson HED
112 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 5.18 illustrates the moving range and individual value charts obtained from Minitab
for the GI biopsy turnaround time data.
Figure 5.18 I-MR chart of GI biopsy turnaround times
In Figure 5.18 , the bottom portion is the moving range chart, and the top portion is the individual value chart. First, the moving range chart is examined for signs of special variation.
None of the points on the moving range chart are outside the control limits, and there are no
other signals indicating a lack of control. Thus, there are no indications of special sources of
variation on the moving range chart. Now the individual value chart can be examined. There
are no indications of a lack of control, so the process can be considered to be stable and the
output predictable with respect to time as long as conditions remain the same.
After the team brainstormed what could be changed in the process flowchart to decrease the
number of days to turn around biopsies (TAT or turnaround time), several changes become
obvious; for example, changing the frequency with which slides are delivered to pathology to
do the biopsies. The changes are implemented in the flowchart, and the results of the revised
flowchart are shown in the right panel of Figure 5.19 .
X Bar and R Charts
What: X Bar and R charts are used when the subgroup size is between 2 and 9. The subgroup
range,
R, is plotted on the R chart, which monitors variability of the process, while the subgroup average, X Bar, is plotted on the X Bar chart, which monitors the central tendency of
the process.
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 113
Why: X Bar and R charts are used if the subgroup size is small (2 to 9) because it is unlikely
that an extreme value will be selected into a subgroup. If an extreme value is accepted into a
subgroup, the range gives a distorted view of the variability in the process because the range
is computed solely from extreme values.
Example: To illustrate the X Bar and R charts, consider a large pharmaceutical firm that
provides vials filled to a specification of 52.0 grams. The firm’s management has embarked
on a program of statistical process control and has decided to use variables control charts for
this filling process to detect special causes of variation. Samples of six vials are selected every
5 minutes during a 105-minute period. Each set of six measurements makes up a subgroup.
Table 5.6 lists the vial weights for 22 subgroups.
Table 5.6 Vial Weights for Six Vials Selected During 22 Time Periods

Observation
1
Time
9:30
1
52.22
2
52.85
3
52.41
4
52.55
5
53.10
6
52.47
2 9:35 52.25 52.14 51.79 52.18 52.26 51.94
3 9:40 52.37 52.69 52.26 52.53 52.34 52.81
4 9:45 52.46 52.32 52.34 52.08 52.07 52.07
5 9:50 52.06 52.35 51.85 52.02 52.30 52.20
6 9:55 52.59 51.79 52.20 51.90 51.88 52.83
7 10:00 51.82 52.12 52.47 51.82 52.49 52.60
8 10:05 52.51 52.80 52.00 52.47 51.91 51.74
9 10:10 52.13 52.26 52.00 51.89 52.11 52.27
10 10:15 51.18 52.31 51.24 51.59 51.46 51.47

Figure 5.19 Before and after change concept control chart
From the Library of Pearson HED
114 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Observation
11
Time
10:20
1
51.74
2
52.23
3
52.23
4
51.70
5
52.12
6
52.12
12 10:25 52.38 52.20 52.06 52.08 52.10 52.01
13 10:30 51.68 52.06 51.90 51.78 51.85 51.40
14 10:35 51.84 52.15 52.18 52.07 52.22 51.78
15 10:40 51.98 52.31 51.71 51.97 52.11 52.10
16 10:45 52.32 52.43 53.00 52.26 52.15 52.36
17 10:50 51.92 52.67 52.80 52.89 52.56 52.23
18 10:55 51.94 51.96 52.73 52.72 51.94 52.99
19 11:00 51.39 51.59 52.44 51.94 51.39 51.67
20 11:05 51.55 51.77 52.41 52.32 51.22 52.04
21 11:10 51.97 51.52 51.48 52.35 51.45 52.19
22 11:15 52.15 51.67 51.67 52.16 52.07 51.81

X Bar and R charts obtained from Minitab for the vials filled data are shown in Figure 5.20 .
Figure 5.20 X Bar and R chart for filling operations
In Figure 5.20 , the bottom portion is the R chart, and the top portion is the X Bar chart. First,
the R chart is examined for signs of special variation. None of the points on the R chart is
outside the control limits, and there are no other signals indicating a lack of control. Thus,
there are no indications of special sources of variation on the R chart.
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 115
Now the X Bar chart can be examined. Notice that a total of five points on the X Bar chart
are outside the control limits (1, 3, 10, 17, and 19), and points 16, 17, and 18 are above +2
standard deviations (2 out of 3 points 2 sigma or beyond rule). Also notice that 18 through
22 show the same type of pattern as points 16 through 18. This indicates a lack of control.
An interesting point comes up here. If the high points are resolved, the low points may be in
control and vice versa. This type of analysis requires a process expert.
Further investigation is warranted to determine the source(s) of these special variations. The
next step is to eliminate the bad special causes of variation and instill the good special causes
of variation. Once this is done, the next step is to reduce the common causes of variation
in the process. This is covered in Chapters 10 through 14 , which are all about the Six Sigma
DMAIC model that is used to improve a process.
X Bar and S Charts
What: As the sample size (subgroup size) n increases, the range becomes increasingly less
efficient as a measure of variability. Since the range ignores all information between the two
most extreme values, as the sample size increases, the range will use a smaller proportion of
the information available in a sample. In addition, the probability of observing an extreme
value in a sample increases as
n gets larger. A single extreme value will result in an unduly
large value for the sample range and will inflate the estimate of process variability. Thus, as
the subgroup size increases (
n is equal to or greater than 10), the individual subgroup standard deviations provide a better estimate of the process standard deviation than the range.
Subgroups should be of sufficient size to detect an out of control event, as with X Bar and R
charts. The common subgroup sizes for X Bar and S charts are 10 or more items. Additionally, subgroup frequency should be often enough to detect changes in the process under
study. This requires expertise in the process under study. If the process can change quickly,
more frequent sampling is needed to detect special causes of variation. If the process changes
slowly, less frequent sampling is needed to detect a special cause of variation.
Why: X Bar and S charts are used to monitor process control and stability of measurement
data where subgroup size is greater than or equal to 10.
Example: To illustrate X Bar and S charts, consider a hospital studying the length of time
patients spend in the admitting process. Samples of 12 patients are selected each day for a 20
day period. The first patient to arrive on the hour is sampled, and the hospital clinic is open
12 hours per day. Admitting time has been operationally defined to all stakeholders’ satisfaction. It is measured in seconds and is seen in Table 5.7 .
From the Library of Pearson HED
116 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Table 5.7 Admitting Process Time in Seconds

Day Patient Time
1 362
1 468
1 553
1 390
1 460
1 910
1 707
1 829
1 955
1 705
1 884
1 904
2 611
2 873
2 768
2 807
2 476
2 816
2 567
2 833
2 521
2 959
2 315
2 414
3 320
3 944
3 593
3 857
3 710
3 724
3 545
3 526
3 348

 

Day Patient Time
3 456
3 576
3 855
4 621
4 927
4 948
4 817
4 641
4 764
4 986
4 430
4 743
4 451
4 645
4 996
5 680
5 794
5 650
5 780
5 442
5 372
5 627
5 882
5 756
5 548
5 767
5 745
6 759
6 665
6 730
6 930
6 369
6 635

 

Day Patient Time
6 313
6 843
6 264
6 663
6 991
6 431
7 372
7 835
7 884
7 930
7 667
7 747
7 390
7 644
7 339
7 664
7 245
7 893
8 370
8 294
8 480
8 558
8 502
8 595
8 847
8 544
8 853
8 876
8 744
8 816
9 530
9 881
9 943

From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 117

Day Patient Time
9 383
9 316
9 611
9 778
9 531
9 896
9 772
9 719
9 670
10 494
10 914
10 870
10 272
10 662
10 348
10 447
10 306
10 751
10 445
10 717
10 387
11 659
11 919
11 603
11 897
11 319
11 499
11 799
11 482
11 615
11 497
11 430
11 765
12 274

 

Day Patient Time
12 754
12 428
12 811
12 916
12 332
12 765
12 961
12 437
12 692
12 380
12 566
13 797
13 253
13 829
13 857
13 898
13 387
13 918
13 900
13 691
13 600
13 450
13 775
14 678
14 679
14 351
14 663
14 638
14 928
14 258
14 338
14 446
14 936
14 584

 

Day Patient Time
14 535
15 997
15 205
15 893
15 734
15 474
15 631
15 746
15 642
15 484
15 525
15 685
15 358
16 242
16 474
16 966
16 823
16 515
16 617
16 894
16 519
16 636
16 547
16 993
16 858
17 594
17 817
17 381
17 462
17 429
17 786
17 901
17 278
17 472

From the Library of Pearson HED
118 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Day Patient Time
17 885
17 991
17 557
18 368
18 850
18 510
18 688
18 201
18 795
18 977
18 715
18 253
18 310

 

Day Patient Time
18 412
18 813
19 806
19 575
19 348
19 298
19 487
19 697
19 249
19 668
19 533
19 985
19 284

 

Day Patient Time
19 707
20 497
20 785
20 806
20 263
20 435
20 337
20 659
20 537
20 786
20 607
20 466
20 564

In Figure 5.21 , the bottom portion is the S chart, and the top portion is the X Bar chart. First,
the S chart is examined for signs of special variation. None of the points on the S chart is outside the control limits, and there are no other signals indicating a lack of control. Thus, there
are no indications of special sources of variation on the S chart. Now the X Bar chart can be
examined. There are no indications of a lack of control, so the process can be considered to be
stable and the output predictable with respect to time as long as conditions remain the same.
Figure 5.21 X bar S chart for admitting time data
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 119
The next step is to reduce the common causes of variation in the process. This is covered in
Chapters 10 through 14 , which are all about the Six Sigma DMAIC model, which is used to
improve a process.
Which Control Chart Should I Use?
Often the most difficult part of using control charts is knowing which control chart to use in
which situation. Figure 5.22 is a flow diagram that can help in deciding which type of chart
is most appropriate for a given situation.

Are the
subgroup
sizes
< 10?
Do you
have attribute
data?

Yes

Yes
Yes No
Count

No Use individuals
and moving
range charts.

Classification
Yes
Use c
charts.
Use X Bar R
charts.
Use X Bar S
charts.

Are the areas
of opportunity
equal?
Classification
or count
data?
Are the
subgroups
> 1?
Figure 5.22 A flow diagram to help decide which type of control chart is most appropriate for a given
situation
Control Chart Case Study
A manufacturer of surgical screws wants to improve the capability of his process. Each day,
for 120 days, he counts the number of defective surgical screws from the defective bin. The
company manufactures 100,000 surgical screws of a given type per day. In this case study we
are examining one of many types of surgical screws. The different classifications of defective
surgical screws for the type of surgical screws under study are improper threads, non-round
From the Library of Pearson HED
120 A Guide to Six Sigma and Process Improvement for Practitioners and Students
heads, improper Philips head indentation, burr(s), too short, too long, and bent. Table 5.8
shows the number of defective surgical screws in each classification by day.
Table 5.8 Defective Surgical Screws in Each Classification by Day

Row
1
Improper
Threads
8
Non
Round
Head
14
Improper
Philips Head
Indentation
9
Burr(s)
6
Too
Long
22
Too
Short
10
Bent
7
Total
Defectives
76
2 18 18 9 7 32 16 14 114
3 12 13 10 13 35 7 9 99
4 12 7 12 5 27 12 9 84
5 12 8 11 5 27 13 19 95
6 8 11 16 11 26 13 18 103
7 5 12 14 14 22 16 10 93
8 16 10 11 9 27 15 15 103
9 14 9 10 11 25 3 7 79
10 12 7 13 13 17 15 9 86
11 11 6 15 10 27 9 12 90
12 13 8 14 8 27 11 13 94
13 10 12 8 10 25 9 4 78
14 14 8 15 8 23 9 9 86
15 8 6 10 10 24 7 8 73
16 11 13 10 11 21 12 9 87
17 16 10 14 11 26 11 7 95
18 12 13 17 15 36 31* 13 137**opera
tor sick on
the machine
that cuts
length
19 12 11 10 12 20 10 8 83
20 7 6 14 3 16 14 10 70
21 10 14 12 8 28 18 13 103
22 7 12 9 5 21 12 8 74
23 7 13 16 11 24 7 16 94
24 9 4 8 9 32 8 7 77
25 8 12 12 13 17 7 10 79
26 8 17 10 10 23 16 10 94
27 7 12 19 13 28 14 6 99

From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 121

Row
28
Improper
Threads
16
Non
Round
Head
11
Improper
Philips Head
Indentation
16
Burr(s)
12
Too
Long
23
Too
Short
15
Bent
7
Total
Defectives
100
29 12 9 11 9 27 14 10 92
30 10 11 9 10 33 14 6 93
31 10 12 7 16 33 18 15 111
32 15 10 12 4 22 15 8 86
33 12 18 14 4 24 6 14 92
34 8 11 15 8 23 11 16 92
35 15 14 11 13 30 16 7 106
36 10 10 11 7 25 9 11 83
37 11 15 10 10 32 11 13 102
38 11 12 9 5 30 13 4 84
39 7 14 18 9 28 7 10 93
40 2 7 12 4 26 11 14 76
41 11 13 18 13 26 10 7 98
42 5 12 19 3 27 13 6 85
43 10 11 13 13 24 9 12 92
44 9 11 11 4 22 13 17 87
45 7 6 7 10 31 12 17 90
46 7 11 14 11 32 13 13 101
47 15 10 10 9 22 13 8 87
48 5 15 7 9 33 13 8 90
49 11 10 10 11 26 8 10 86
50 13 8 17 6 20 10 10 84
51 12 10 6 13 23 5 18 87
52 4 18 14 9 20 8 6 79
53 10 10 16 9 24 11 10 90
54 9 8 13 5 28 16 9 88
55 14 9 21 10 26 11 14 105
56 11 9 12 10 17 7 17 83
57 5 20 10 12 26 8 17 98
58 12 15 12 11 34 13 7 104
59 15 16 14 7 19 15 5 91
60 14 13 14 9 21 14 14 99
61 11 11 6 6 20 13 11 78

From the Library of Pearson HED
122 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Row
62
Improper
Threads
2
Non
Round
Head
13
Improper
Philips Head
Indentation
14
Burr(s)
11
Too
Long
22
Too
Short
15
Bent
6
Total
Defectives
83
63 6 6 16 4 20 13 9 74
64 12 16 9 8 17 9 17 88
65 6 12 14 7 20 14 10 83
66 9 12 14 5 18 12 5 75
67 10 17 14 11 30 20 7 109
68 13 14 7 8 27 13 11 93
69 11 8 14 3 23 10 17 86
70 6 11 16 9 26 14 16 98
71 10 12 12 6 30 16 7 93
72 8 13 16 6 24 8 11 86
73 8 9 17 9 27 9 12 91
74 6 20 11 13 23 15 10 98
75 10 19 5 8 22 10 7 81
76 8 9 13 8 26 15 9 88
77 6 6 12 7 23 11 15 80
78 11 10 16 12 27 5 10 91
79 10 11 8 8 26 11 13 87
80 8 15 15 7 35 18 5 103
81 4 13 16 4 29 5 11 82
82 13 9 12 8 33 14 8 97
83 16 15 5 13 20 11 16 96
84 8 9 13 10 30 13 12 95
85 11 13 14 13 27 9 7 94
86 10 14 10 10 13 14 7 78
87 10 14 16 8 27 14 12 101
88 10 8 10 10 21 6 6 71
89 9 15 9 9 26 15 14 97
90 9 9 8 11 23 16 18 94
91 9 11 13 17 27 9 19 105
92 8 14 15 9 27 11 7 91
93 10 12 13 12 12 12 8 79
94 13 9 13 12 25 18 14 104
95 8 17 10 15 27 6 16 99

From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 123

Row
96
Improper
Threads
9
Non
Round
Head
13
Improper
Philips Head
Indentation
11
Burr(s)
8
Too
Long
21
Too
Short
10
Bent
20
Total
Defectives
92
97 10 13 14 12 26 14 8 97
98 11 13 19 10 28 16 14 111
99 14 9 10 12 25 5 12 87
100 15 8 13 8 27 7 7 85
101 4 11 13 9 14 11 17 79
102 8 7 19 14 21 8 8 85
103 8 3 6 7 25 17 10 76
104 11 11 17 11 25 12 11 98
105 6 11 19 9 21 8 15 89
106 13 12 12 8 26 16 14 101
107 13 10 11 13 23 6 13 89
108 9 11 13 6 24 19 15 97
109 12 13 12 3 22 8 9 79
110 10 12 13 12 27 5 6 85
111 6 9 8 7 22 6 17 75
112 11 19 12 19 28 12 9 110
113 18 12 12 12 21 8 9 92
114 11 10 6 6 24 11 8 76
115 8 9 11 6 21 18 10 83
116 7 11 13 12 31 14 13 101
117 6 15 7 12 35 8 11 94
118 13 12 13 7 21 15 11 92
119 7 11 13 7 29 6 9 82
120 4 11 14 9 19 10 8 75

As you can see from the p chart in Figure 5.23 there is a special cause of variation on the 18th
subgroup. Team members went back to the log sheet for that day and saw that the operator
who cuts the length of the surgical screws was out sick. The team members assumed that this
was the cause of the special variation. Consequently, they added into the flowchart of the job,
to have a trained backup operator for each machine in case this special cause happened again.
Next time they would be prepared. Consequently, they dropped the out of control point and
recalculated the control chart. Figure 5.24 shows the result.
From the Library of Pearson HED
124 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 5.23 P chart of total defectives
As you can see in Figure 5.24 , the process is now in statistical control. Since the process is
stable the team members took all the data points and put them into a Pareto diagram, shown
in Figure 5.25 .
Figure 5.24 P chart after special cause removed
Team members realized that the same problem caused too short and too long screws. Consequently, they combined the categories; see the Pareto diagram in Figure 5.26 .
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 125
Figure 5.26 Pareto diagram with combined reasons
Team members studied the problem and discovered that a setting on the surgical screw
machine was off. They corrected it, and the problem of too short and too long screws completely disappeared; see the control chart in Figure 5.27 .
The preceding case study is an example of how to use control charts and change concepts to
resolve special causes in a process and how to remove common causes from a process.
Figure 5.25 Pareto diagram of defect reasons
From the Library of Pearson HED
126 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Measurement Systems Analysis
Measurement systems analysis studies are used to calculate the capability of a measurement
system for a variable to determine whether it can deliver accurate data to team members.
There are three parts to a measurement systems study:
1. Measurement systems analysis (MSA) checklist
2. Test-retest study
3. Gage Repeatability and Reproducibility (R&R) study
We want to make sure that when we are measuring and collecting data that our measurement
methods, instruments, and the process of collecting data are correct to ensure the integrity of
the data we are using in our process improvement efforts.
We only cover the basics of measurement systems analysis in this book. For a more advanced
understanding see Gitlow and Levine (2004). We discuss measurement systems analysis
checklists and Gage R&R studies.
R&R stands for repeatability and reproducibility.
Measurement System Analysis Checklist
A measurement system analysis checklist involves determining whether the following tasks
have been completed (Gitlow and Levine, 2004):
1. Description of the ideal measurement system (flowchart the process)
2. Description of the actual measurement system (flowchart the process)
3. Identification of the causes of the differences between the ideal and actual measurement systems
Figure 5.27 P chart after correcting problem with too short and too long screws
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 127
4. Identification of the accuracy (bias) and precision (repeatability) of the measurement
system
5. Estimation of the proportion of observed variation due to unit-to-unit variation and
R&R variation using a Gage R&R study
Gage R&R Study
A Gage R&R study is used to estimate the proportion of observed total variation due to unitto-unit variation and R&R variation. R&R variation includes repeatability, reproducibility,
and operator-part interaction (different people measure different units in different ways). If
R&R variation is large relative to unit-to-unit variation, the measurement system must be
improved before collecting data.
The data required by a Gage R&R study should be collected so that it represents the full
range of conditions experienced by the measurement system. For example, the most senior
inspector and the most junior inspector should repeatedly measure each item selected in the
study. The data should be collected in random order to prevent inspectors from influencing
each other.
An example of a Gage R&R study follows. Two inspectors independently use a light gauge
to measure each of five units four separate times; this results in 40 measurements. The data
to be collected by the two inspectors is shown in Table 5.9 . Table 5.9 presents the standard
order, or the logical pattern, for the data to be collected by the team members.
Table 5.9 Standard Order for Collecting Data for the Gage R&R Study

Row Unit Inspector Measurement
1 1 Enya To be collected
2 1 Enya To be collected
3 1 Enya To be collected
4 1 Enya To be collected
5 1 Lucy To be collected
6 1 Lucy To be collected
7 1 Lucy To be collected
8 1 Lucy To be collected
9 2 Enya To be collected
10 2 Enya To be collected
11 2 Enya To be collected
12 2 Enya To be collected
13 2 Lucy To be collected

 

Row Unit Inspector Measurement
14 2 Lucy To be collected
15 2 Lucy To be collected
16 2 Lucy To be collected
17 3 Enya To be collected
18 3 Enya To be collected
19 3 Enya To be collected
20 3 Enya To be collected
21 3 Lucy To be collected
22 3 Lucy To be collected
23 3 Lucy To be collected
24 3 Lucy To be collected
25 4 Enya To be collected
26 4 Enya To be collected

From the Library of Pearson HED
128 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Row Unit Inspector Measurement
27 4 Enya To be collected
28 4 Enya To be collected
29 4 Lucy To be collected
30 4 Lucy To be collected
31 4 Lucy To be collected
32 4 Lucy To be collected
33 5 Enya To be collected

 

Row Unit Inspector Measurement
34 5 Enya To be collected
35 5 Enya To be collected
36 5 Enya To be collected
37 5 Lucy To be collected
38 5 Lucy To be collected
39 5 Lucy To be collected
40 5 Lucy To be collected

Table 5.10 shows the random order used by team members to collect the measurement data
required for the measurement study. Random order is important because it removes any
problems induced by the structure of the standard order. Table 5.10 is an instruction sheet
to the team members actually collecting the data.
Table 5.10 Random Order for Collecting Data for the Gage R&R Study

Random
Order
Standard
Order
Unit Inspector
1 36 5 Enya
2 5 1 Lucy
3 30 4 Lucy
4 29 4 Lucy
5 26 4 Enya
6 28 4 Enya
7 6 1 Lucy
8 8 1 Lucy
9 4 1 Enya
10 3 1 Enya
11 18 3 Enya
12 20 3 Enya
13 40 5 Lucy
14 9 2 Enya
15 31 4 Lucy
16 24 3 Lucy
17 38 5 Lucy
18 17 3 Enya
19 32 4 Lucy
20 11 2 Enya

 

Random
Order
Standard
Order
Unit Inspector
21 27 4 Enya
22 19 3 Enya
23 10 2 Enya
24 33 5 Enya
25 37 5 Lucy
26 2 1 Enya
27 35 5 Enya
28 23 3 Lucy
29 13 2 Lucy
30 7 1 Lucy
31 15 2 Lucy
32 22 3 Lucy
33 14 2 Lucy
34 34 5 Enya
35 1 1 Enya
36 12 2 Enya
37 16 2 Lucy
38 39 5 Lucy
39 21 3 Lucy
40 25 4 Enya

From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 129
Table 5.11 shows the data collected in the Gage R&R study in random order.
Table 5.11 Data for Gage R&R Study

Random Order Standard Order Unit Inspector Measure
1 36 5 Enya 21.85
2 5 1 Lucy 21.19
3 30 4 Lucy 23.14
4 29 4 Lucy 23.09
5 26 4 Enya 23.28
6 28 4 Enya 23.23
7 6 1 Lucy 21.29
8 8 1 Lucy 21.24
9 4 1 Enya 21.24
10 3 1 Enya 21.33
11 18 3 Enya 22.28
12 20 3 Enya 22.34
13 40 5 Lucy 21.78
14 9 2 Enya 21.65
15 31 4 Lucy 23.02
16 24 3 Lucy 22.17
17 38 5 Lucy 21.84
18 17 3 Enya 22.31
19 32 4 Lucy 23.19
20 11 2 Enya 21.67
21 27 4 Enya 23.24
22 19 3 Enya 22.31
23 10 2 Enya 21.60
24 33 5 Enya 21.84
25 37 5 Lucy 21.76
26 2 1 Enya 21.29
27 35 5 Enya 21.93
28 23 3 Lucy 22.14
29 13 2 Lucy 21.50
30 7 1 Lucy 21.21
31 15 2 Lucy 21.51
32 22 3 Lucy 22.23
33 14 2 Lucy 21.55

From the Library of Pearson HED
130 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Random Order Standard Order Unit Inspector Measure
34 34 5 Enya 21.89
35 1 1 Enya 21.34
36 12 2 Enya 21.56
37 16 2 Lucy 21.55
38 39 5 Lucy 21.81
39 21 3 Lucy 22.18
40 25 4 Enya 23.27

A visual analysis of the data in Table 5.11 using a Gage R&R run chart from Minitab reveals
the results shown in Figure 5.28 . In Figure 5.28 , each dot represents Enya’s measurements
and each square represents Lucy’s measurements. Multiple measurements by each inspector are connected with lines. Good repeatability is demonstrated by the low variation in
the squares connected by lines and the dots connected by lines, for each unit. Figure 5.28
indicates that repeatability (within group variation) is good. Good reproducibility is demonstrated by the similarity of the squares connected by lines and the dots connected by lines,
for each unit. Reproducibility (one form of between group variation) is good. The gage run
chart shows that most of the observed total variation in light gage readings is due to differences between units. This data looks good.
Figure 5.28 Gage R&R run chart obtained from Minitab
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 131
How-To Guide for Understanding Variation: Minitab User
Guide (Minitab Version 17, 2013)
The following section lists the steps necessary to create the statistical tools and methods
described in this chapter using Minitab 17. It is important that you get comfortable with
Minitab so it becomes part of your regular routine as you do your job and improve your job
using process improvement theory and practices.
Go to www.ftpress.com/sixsigma to download the data files referenced in this chapter so you
can practice with Minitab.
Using Minitab to Obtain Zone Limits
To plot zone limits on any of the control charts discussed in this chapter, open to the Data
Source dialog box for the control chart being developed and do the following:
1. Click the Scale button. Click the Gridlines tab. Select the Y major ticks, Y minor
ticks
, and X major ticks check boxes. Click OK to return to the Data Source dialog
box.
2. Select the Options button. Select the S limits tab. In the Standard Deviation Limit
Positions edit box, select
Constants in the drop-down list box and enter 1 2 3 in the
edit box. Click
OK to return to the Data Source dialog box.
Using Minitab for the P Chart
To illustrate how to obtain a p chart, refer to the data in Table 5.1 concerning the number of
broken slides. Open the
SLIDES.MPJ worksheet and do the following:
An Anecdote of Measurement Systems

Several years ago I visited a chocolate factory. We began the tour, and when we got to
the chocolate bean roasting operation, I asked the owner: “How do you know when
the beans are properly roasted?” He said: “I have two roaster tasters, and each has at
least ten years of experience. No problem!” I asked how he knew that each roaster was
consistent with himself over time and consistent with the other roaster over time. He
got angry with me because he thought my question was stupid. I asked if he would
consider a test where ten beans were sliced into four sections and randomly assigned to
the roasters. My idea was for each roaster to taste each of the ten beans twice in random
order, and then to use a gage run chart to determine whether they were consistent with
themselves for each bean, and consistent with each other for each bean. He got furious
and kicked me out of the factory. I don’t know if his measurement system for roasting
beans was effective in providing quality information, but his chocolate did taste pretty
good to me. But then again, I am a chocoholic.

From the Library of Pearson HED
132 A Guide to Six Sigma and Process Improvement for Practitioners and Students
1. Select Stat | Control Charts | Attribute Charts | P. In the P Chart dialog box (see Figure 5.29 ) enter C3 or ‘Number Cracked’ in the Variables edit box. Since the subgroup
sizes are equal, select
Size in the Subgroup Sizes drop-down list box and enter 100 in
the edit box. Click the
P Chart Options button.
Figure 5.29 Minitab P Chart dialog box
2. In the P Chart: Options dialog box, click the Tests tab (see Figure 5.30 ). Select
Perform All Tests for Special Causes from the drop-down list. Click OK to return to
the P Chart dialog box. (These values stay intact until Minitab is restarted.)
Figure 5.30 Minitab P Chart: Options dialog box, Tests tab
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 133
3. If there are points that should be omitted when estimating the center line and control
limits, click the
Estimate tab in the P Chart: Options dialog box (see Figure 5.31 ).
Enter the points to be omitted in the edit box shown. Click
OK to return to the
P Chart dialog box. In the P Chart dialog box, click
OK to obtain the p chart.
Figure 5.31 Minitab P Chart: Options dialog box, Estimate tab
Figure 5.32 shows the output for the p chart.
Figure 5.32 Minitab output for the p chart
From the Library of Pearson HED
134 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Using Minitab for the C Chart
To illustrate how to obtain a c chart, refer to the data in Table 5.3 concerning the number of
add-ons in an outpatient clinic. Open the
ADDONS.MPJ worksheet and follow these steps:
1. Select Stat | Control Charts | Attribute Charts | C. In the C Chart dialog box (see
Figure 5.33 ), enter
C2 or ADDONS in the Variables edit box.
Figure 5.33 Minitab C Chart dialog box
2. Click the C Chart Options button. In the C Chart: Options dialog box, click the
Tests tab. Select Perform All Tests for Special Causes from the drop-down list.
Click
OK to return to the C Chart dialog box. (These values stay intact until Minitab
is restarted.) Click
OK to obtain the c chart.
3. If there are points that should be omitted when estimating the center line and control
limits, click the
Estimate tab in the C Chart: Options dialog box. Enter the points to
be omitted in the edit box shown. Click
OK to return to the C Chart dialog box.
Figure 5.34 shows the output for the c chart.
Using Minitab for the U Chart
To illustrate how to obtain a u chart, refer to the data in Table 5.4 concerning the number of
patient falls in a hospital. Open the
FALLS.MPJ worksheet and perform these steps:
1. Select Stat | Control Charts | Attribute Charts | U. In the U Chart dialog box
(see Figure 5.35 ), enter
C3 or FALLS in the Variables edit box. In the Subgroup
Sizes drop-down list box, select
Indicator Column and enter C2 or CENSUS in the
edit box.
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 135
2. In the U Chart: Options dialog box, click the Tests tab. Select the Perform All Tests
for Special Causes
from the drop-down list. Click OK to return to the U Chart dialog box. (These values stay intact until Minitab is restarted.) Click OK to obtain the
U chart.
Figure 5.34 Minitab output for the c chart
Figure 5.35 Minitab U Chart dialog box
From the Library of Pearson HED
136 A Guide to Six Sigma and Process Improvement for Practitioners and Students
3. If there are points that should be omitted when estimating the center line and control
limits, click the
Estimate tab in the U Chart: Options dialog box. Enter the points to
be omitted in the edit box shown. Click
OK to return to the U Chart dialog box.
Figure 5.36 shows the output for the u chart.
Figure 5.36 Minitab output for the u chart
Using Minitab for the Individual Value and Moving Range Charts
Individual Value and Moving Range charts can be obtained from Minitab by selecting Stat
| Control Charts | Variable Charts for Individuals | I-MR
from the menu bar. To illustrate
how to obtain Individual Value and Moving Range charts, refer to the data in Table 5.5 concerning the turnaround times of GI biopsies. Open the
TURNAROUND.MPJ worksheet
and follow these steps:
1. Select Stat | Control Charts | Variable Charts for Individuals | I-MR. In the
Individuals-Moving Range Chart dialog box (see Figure 5.37 ), enter
‘GI BIOPSY
TURNAROUND TIMES’
in the Variables edit box. Click the I-MR Options button.
2. In the I-MR Chart: Options dialog box, click the Tests tab. Select Perform All Tests
for Special Causes
from the drop-down list. Click OK to return to the I-MR Chart
dialog box. (These values stay intact until Minitab is restarted.)
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 137
3. If there are points that should be omitted when estimating the center line and control
limits, click the
Estimate tab in the I-MR Chart: Options dialog box. Enter the points
to be omitted in the edit box shown. Click
OK to return to the I-MR Chart dialog box.
(Note: When obtaining more than one set of Individual Value and Moving Range
charts in the same session, be sure to reset the values of the points to be omitted before
obtaining new charts.)
4. In the I-MR Chart dialog box, click OK to obtain the individual value and moving
range charts.
Figure 5.38 shows the output for the I-MR chart.
Using Minitab for the X Bar and R Charts
X Bar and R charts can be obtained from Minitab by selecting Stat | Control Charts | Variable
Charts for Subgroups | Xbar-R from the menu bar. The format for entering the variable name
is different, depending on whether the data are stacked down a single column or unstacked
across a set of columns with the data for each time period located in a single row. If the data
for the variable of interest are stacked down a single column, choose All Observations for a
Chart Are in One Column from the drop-down list and enter the variable name in the edit
box below. If the subgroups are unstacked with each row representing the data for a single
time period, choose Observations for a Subgroup Are in One Row of Columns from the
drop-down list and enter the variable names for the data in the edit box below.
Figure 5.37 Minitab I-MR Chart dialog box
From the Library of Pearson HED
138 A Guide to Six Sigma and Process Improvement for Practitioners and Students
To illustrate how to obtain X Bar and R charts, refer to the data in Table 5.6 concerning the
weight of vials. Open the
VIALS.MPJ worksheet and follow these steps:
1. Select Stat | Control Charts | Variable Charts for Subgroups | Xbar-R. In the Xbar-R
Chart dialog box (see Figure 5.39 ) enter
C3 or ‘1’, C4 or ‘2’, C5 or ‘3’, C6 or ‘4’, C7 or
‘5’, and C8 or ‘6’, in the edit box. Click the Xbar-R Options button.
Figure 5.38 Minitab output for the I-MR chart
Figure 5.39 Minitab Xbar-R Chart dialog box
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 139
2. In the Xbar-R Chart: Options dialog box (see Figure 5.40 ), click the Tests tab. Select
the
Perform All Tests for Special Causes from the drop-down list. Click OK to
return to the Xbar-R Chart dialog box. (These values stay intact until Minitab is
restarted.)
Figure 5.40 Minitab Xbar-R Chart: Options dialog box, Tests tab
3. If there are points that should be omitted when estimating the center line and control
limits, click the
Estimate tab in the Xbar-R Chart: Options dialog box (see Figure
5.41 ). Enter the points to be omitted in the edit box shown. Click
OK to return to
the Xbar-R Chart dialog box. (Note: When obtaining more than one set of X Bar and
R charts in the same session, be sure to reset the values of the points to be omitted
before obtaining new charts.)
4. In the Xbar-R Chart dialog box, click OK to obtain the X Bar and R charts.
Figure 5.42 shows the output for the X Bar and R chart.
Using Minitab for the X Bar and S Charts
X Bar and S charts can be obtained from Minitab by selecting Stat | Control Charts | Variable Charts for Subgroups | Xbar-S from the menu bar. The format for entering the variable name is different, depending on whether the data are stacked down a single column or
unstacked across a set of columns with the data for each time period located in a single row.
If the data for the variable of interest are stacked down a single column, choose
All Observations for a Chart Are in One Column from the drop-down list and enter the variable name
in the edit box below. If the subgroups are unstacked with each row representing the data
for a single time period, choose
Observations for a Subgroup Are in One Row of Columns
from the drop-down list and enter the variable names for the data in the edit box below.
From the Library of Pearson HED
140 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 5.41 Minitab Xbar-R Chart: Options dialog box, Estimate tab
Figure 5.42 Minitab output for the X Bar and R chart
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 141
To illustrate how to obtain X Bar and S charts, refer to the data in Table 5.7 concerning
admitting processing time. Open the
ADMITTING.MPJ worksheet and follow these steps:
1. Select Stat | Control Charts | Variable Charts for Subgroups | Xbar-S. In the Xbar-S
Chart dialog box (see Figure 5.43 ) enter
TIME in the edit box. In the Subgroup Sizes
drop-down list box, select
All Observations for a Chart Are in One Column. Enter
10 in the edit box. Click the Xbar-S Options button.
2. In the Xbar-S Chart: Options dialog box, click the Tests tab. Select Perform All Tests
for Special Causes
from the drop-down list. Click OK to return to the Xbar-S Chart
dialog box. (These values stay intact until Minitab is restarted.)
3. If there are points that should be omitted when estimating the center line and control limits, click the Estimate tab in the Xbar-S Chart: Options dialog box. Enter the
points to be omitted in the edit box shown. Click
OK to return to the Xbar-S Chart
dialog box. (Note: When obtaining more than one set of X Bar and S charts in the
same session, be sure to reset the values of the points to be omitted before obtaining
new charts.)
Figure 5.43 Minitab Xbar-S Chart dialog box
4. In the Xbar-S Chart dialog box, click OK to obtain the X Bar and S charts.
Figure 5.44 shows the output for the X Bar and S chart.
From the Library of Pearson HED
142 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Figure 5.44 Minitab output for the X Bar and S chart
Takeaways from This Chapter
Control charts are used to distinguish between common causes of variation and special causes of variation, which lets you know whether your process is stable and under
control.
The type of control chart you use depends on the type of data you have.
Different rules help identify special causes of variation.
If you have attribute data you use attribute control charts as follows:
P charts—If you have attribute classification data
C charts—If you have attribute count data with constant areas of opportunity
U charts—If you have attribute count data with nonconstant areas of opportunity
If you have measurement data you use variables control charts as follows:
X bar R charts—If your subgroup size is between 2 and 9
X bar S charts—If your subgroup size is 10 or greater
Individuals and moving range (I-MR) charts—If your subgroup size is 1
From the Library of Pearson HED
Chapter 5 Understanding Variation: Tools and Methods 143
References
Gitlow, H. G., A. Oppenheim, R. Oppenheim, and D. M. Levine (2015), Quality Management: Tools and Methods for Improvement, 4th ed. (Naperville, IL: Hercher Publishing
Company). This book is free online at hercherpublishing.com.
Gitlow, H. and D. Levine (2004),
Six Sigma for Green Belts and Champions: Foundations, DMAIC, Tools and Methods, Cases and Certification (Upper Saddle River, NJ:
Prentice-Hall).
Minitab Version 17 (State College, PA: Minitab, Inc., 2013).
Additional Readings
Montgomery, D. C. (2000), Introduction to Statistical Quality Control, 4th ed. (New York:
John Wiley).
Shewhart, W. A. (1931),
Economic Control of Quality of Manufactured Product (New York:
Van Nostrand-Reinhard), reprinted by the American Society for Quality Control, Milwaukee, 1980.
Western Electric (1956),
Statistical Quality Control Handbook (Indianapolis, IN: Western
Electric).
From the Library of Pearson HED
This page intentionally left blank
From the Library of Pearson HED
145
6
Non-Quantitative Techniques:
Tools and Methods
What Is the Objective of This Chapter?
The objective of this chapter is to introduce you to the non-quantitative tools and methods
used in process improvement. The chapter is split into two sections. The first section is a high
level overview of the non-quantitative tools and methods used in process improvement with
examples, and the second section is a more in-depth step-by-step how-to guide on using the
different non-quantitative tools and methods.
The idea is for you to read the first part now so you understand the tools and methods utilized in the Six Sigma DMAIC model discussed in Chapters 10 through 14 . You can either
read the detailed how-to section now, or you can come back and read it later while working
on your project.
Before we proceed any further, we need to define some notation. CTQ is an acronym for
Critical to Quality characteristic. It is the problematic metric that your process improvement project is aimed at improving; it is called Y in Statistics, or the dependent variable. X
refers to the steps in the process that can be manipulated to get better output for the CTQ.
For example, suppose the total cycle time to process a student in the Accounts Receivable
department at a university (CTQ) is a function of the time the student waits for service (X
1),
the time it takes the clerk to pull the student’s records (X
2), the time it takes to process the
student’s material (X
3), and the time it takes to refile the student’s records and to call on the
next student waiting in line (X
4). In this case we can say that the CTQ is an additive function
of X
1, X2, X3, and X4; in mathematical terms: CTQ = f(X1,X2,X3,X4), or more specifically CTQ
= X
1+ X2+ X3 + X4.
High Level Overview and Examples of Non-Quantitative
Tools and Methods
This section introduces you to non-quantitative tools and methods used in process improvement with examples to ensure understanding of each tool. Have fun!
From the Library of Pearson HED
146 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Flowcharting
What: A flowchart is a tool used to map out (draw a picture of) a process (Gitlow et al.,
2015). The two types of flowcharts introduced in Chapter 3 , “Defining and Documenting a
Process,” are the following:
Process flowcharts—Process flowcharts describe the steps and decision points of a
process in a downward direction from the start (top of the page) to the stop (bottom
of the page) of the process.
Deployment flowcharts—Deployment flowcharts are organized into “lanes” that
show processes that involve various departments, people, stages, or other categories.
Why: Process flowcharts are used when you want to depict a process at a high level or when
you want to drill down into a detailed portion of a process. Deployment flowcharts are used
when you want to show who is responsible for different parts of a process, as well as track the
number and location of handoffs within the process.
Example: Figure 6.1 shows an example of a process flowchart for a typical inpatient cardiology consult process.
Figure 6.2 is an example of a deployment flowchart for a surgical biopsy process.
The starbursts in Figure 6.2 show the locations of problematic areas in the process. These
areas may be examined at a later point in time as possible areas for improving the process.
The starbursts indicate steps in the process that may become potential Xs for improving the
problematic CTQ.
Voice of the Customer (VoC)
What: Voice of the Customer analysis involves surveying stakeholders of a process to understand their requirements and needs (Gitlow and Levine, 2004). Stakeholders can include
customers, employees, investors, regulatory agencies, building and grounds, the legal system,
and the environment, to name a few possible stakeholders of a process.
There are typically four stages when conducting the VoC analysis:
1. Define/segment the market. (Identify the stakeholder groups.)
2. Plan the VoC. (Identify who will be interviewed and who will interview them, establish a time schedule, and prepare the questions.)
3. Collect the data by stakeholder group.
4. Organize and interpret the data.
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 147

No Sr. patient rep
ships forms to
billing office.

Yes
START
Inpatient Cardiology Consult Process
STOP

Physician
prepares
handwritten
note.

 

Sr. patient rep
prepares a face
sheet with name
and DOB of
patient.

 

Physician
completes
encounter form
after consult.

 

All forms are
batched
(facesheets and
encounters).

 

Administrator
collects all forms
every Thursday
morning in
person.

Did
Administrator
collect form?
Figure 6.1 Process flowchart
From the Library of Pearson HED
148 A Guide to Six Sigma and Process Improvement for Practitioners and Students

START
1
GI Clinic
GI Pathology Biopsy-Process Map–After Change
Pathology Department END
Slides of new
cases distributed
by secretary.
Cases
reviewed by
fellow daily.
Cases
reviewed by
resident daily.
Cases with
odd numbers??
Request slides
for extra staining.
Put case
on hold.
Case diagnosed
by pathologist 1
next day from
8:00-12:00.
6
Transcription
of final diagnosis
by transcriptionist.
Case diagnosed
by pathologist 2
next day from
8:00-12:00.
Pending cases
diagnosed by
pathologist and
resident/fellow
daily in second shift.
Transcription of
final diagnosis with
ancillary results by
transcriptionist.
Sign off final
diagnosis
electronically by
pathologist.
Final diagnosis
with ancillary results
by pathologists.
Does case
need ancillary
testing?
Is the staining
critical?
Y
N
Y
Y
N
GI Path Lab Received cases
signed off in lab
notebook by lab
asst.
Accessioning
conducted by
lab asst.
3
Macroscopic
evaluation by
path asst.
Macroscopic
evaluation
transcribed by
transcriptionist.
Tissue
processing using
short cycle
by histo tech.
4
Histologic
preparation and
QC of slides
by histo tech.
5

Case obtained
by medical service.
Tissue along with
pathology request form is
placed at sample collection
room and case logged in
log book by GI nurse.
Case transported
to lab by hospital courier
once a day.
Dropped cases logged
in path lab notebook the
by courier.
2
Slides
sent to case
manager by
histo tech.
Reconciliation of
slides with macroscopic
evaluation by histo tech.
Slides
transported from
hospital path lab
by AP courier.
Transcript on
of final diagnosis
without ancillary testing
by transcriptionist.
Slides
received by
secretary.
Slides of
pending cases
distributed by
secretary.
Sign off the
addendum
electronically by
pathologist.
N
Figure 6.2 Deployment flowchart
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 149
Supplier-Input-Process-Output-Customer (SIPOC) Analysis
What: A SIPOC analysis is a simple tool for identifying the suppliers and their inputs into
a process, the high level steps of a process, the outputs of the process, and the customer segments interested in the outputs of the process (Gitlow and Levine, 2004).
Why: The SIPOC analysis helps team members define and give scope to a project. It helps
the team understand the process at a high level, and it helps identify customer (stakeholder)
segments and their needs and wants (outputs: both intended and unintended), and suppliers and their inputs. It also helps to clarify who will be interviewed during the VoC analysis.
Example: Table 6.1 shows an example of a SIPOC analysis for the patient scheduling process
at an outpatient clinic in a hospital.
An Anecdote on Voice of the Customer Analysis

A few years ago a friend of mine decided to perform a personal Voice of the Customer
analysis. At the time his main customers were his wife, his son, his mother (his dad
had passed), and his friends. He decided to begin the VoC data collection with his
wife. He was surprised that she quickly listed out 20 items that she thought he needed
to attend to, his CTQs if you will. Well, he realized that it was impossible for him to
do all 20 items. So, he created a 1 to 10 dynamite scale that his wife used to rate the
importance of each of the 20 items. One stick of dynamite meant that the item was not
so important, while 10 sticks of dynamite meant the item was very important. Three
of the 30 items involved spending more time with the children; this accounted for 70%
of his wife’s sticks of dynamite. Therefore, “spending time with the children” was his
wife’s critical area of concern (CTQ). He decided to play with the children for at least
two hours every Saturday. His wife was happy with the results. His new best practice
method saved many fights with his wife by allowing him to prioritize his wife’s issues;
think Pareto diagram.

Why: The data we collect from the VoC analysis helps us take the needs of our stakeholders
into consideration when defining the metrics (variables) of concern to them.
Example: Due to the complex nature and length of VoC analysis, an example is given at the
same time we show you how to conduct VoC analysis later in the second part of this chapter.
From the Library of Pearson HED
150 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Table 6.1 SIPOC Analysis

Process Name: Patient Scheduling Process in the Outpatient Clinic at XYZ Hospital
Process Owner: Assistant Vice President of Outpatient Services
SUPPLIERS INPUTS PROCESS OUTPUTS CUSTOMERS
(providers
of required
resources)
(resources required
by the process)
(deliverables from
the process; these
can be CTQs or Xs)
(stakeholders who
put requirements
on the outputs)
Referring physicians
Patients /families
Appointment
schedulers
Insurance providers
Referral
Phone call
Patient information
Patient and physician
availability
Insurance plan
Call comes in

Patient entered
into electronic
medical record
system

Appointment
scheduled

Insurance verified

Reminder call
if time

New patient record
in electronic medical
record system
Scheduled
appointment
Verified insurance
Reminder
Arrived visit
Patient no show
Patient registration
reps in clinics
Nurses
Physicians
Insurance providers
Finance office
Patients /families

An Anecdote about a SIPOC Analysis

A SIPOC analysis is the first real opportunity that team members have to look at their
process from a 360-degree point of view. They may learn some surprising things about
dysfunctional suppliers or customers they didn’t know existed. For example, I was
consulting at a factory that had a problem with down time on machines. So, the pro
cess of concern was the request for maintenance process. One of the customers of
the request for maintenance process was the employees who operated machines that
needed maintenance. One of the suppliers to the request for maintenance process was
the maintenance personnel that fixed the machines.
The request for maintenance protocol had a 1 to 4 scale for prioritizing requests for
maintenance:
1 = low priority, get to it as soon as is convenient.
2 = moderate priority, get to it within the next day or two, or else it might turn
into a bigger problem.
3 = high priority, get to it sometime today because the machine is not working at
the necessary speed.

From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 151
Operational Definitions
What: An operational definition promotes understanding between people by putting communicable meaning into words (Gitlow et al., 2015; Gitlow and Levine, 2004). An operational
definition contains three parts: a criterion to be applied to an object or group, a test of the
object or group, and a decision as to whether the object or group meets the criterion.
Criteria—Operational definitions establish VoC specifications for each CTQ that
are compared with the Voice of the Process outputs in the test step of an operational
definition.
Test—A test involves comparing Voice of the Process data with VoC specifications
for each CTQ for a given unit of output.
Decision—A decision involves making a determination whether a given unit of output meets VoC specifications.
Why: Operational definitions are required to give communicable meaning to terms such as
late, clean, good, red, round, 15 minutes, or 3:00 p.m. Problems, such as endless bickering and
ill-will, can arise from the lack of an operational definition. A definition is operational if all
relevant users of the definition agree on the definition.
Example: A firm produces washers. One of the critical quality characteristics is roundness.
The following procedure is one way to arrive at an operational definition of roundness, as
long as the buyer and seller agree on it.
Step 1: Criterion for roundness.
Buyer: “Use calipers (a measuring device) that are in reasonably good order.” (You
perceive at once the need to question every word.)
Seller: “What is ‘reasonably good order’?”
4 = emergency priority, drop whatever you are doing and get to the machine
because it is not running and is essentially taking the factory down until it is
fixed.
The problem was that all requests for maintenance were coded as “4” by the machine
operators. They did this because they had quotas to meet so they would not get in
trouble. This meant that the maintenance personnel were running to fix machines that
were low priority, while high priority machines were not attended too in a timely fashion. So, here you see how a SIPOC analysis helps identify a dysfunction in the request
for maintenance process.
As a side bar, resolving this problem is difficult in the traditional paradigm of management that relies on quotas to reward and punish employees. This remained a problem
after I stopped consulting for the factory because they never really adopted a process
orientation to their business.
From the Library of Pearson HED
152 A Guide to Six Sigma and Process Improvement for Practitioners and Students
(We settle the question by letting you use your calipers.)
Seller: “But how should I use them?”
Buyer: “We’ll be satisfied if you just use them in the usual way.”
Seller: “At what temperature?”
Buyer: “The temperature of this room.”
Buyer: “Take six measures of the diameter about 30 degrees apart. Record the results.”
Seller: “But what is ‘about 30 degrees apart’? Don’t you mean exactly 30 degrees?”
Buyer: “No, there’s no such thing as exactly 30 degrees in the physical world. So try
for 30 degrees. We’ll be satisfied.”
Buyer: “If the range between the six diameters doesn’t exceed .007 centimeters, we’ll
declare the washer to be round.”
(They have determined the criterion for roundness.)
Step 2: Test of roundness.
a. Select a particular washer.
b. Take the six measurements and record the results in centimeters: 3.365, 3.363, 3.368,
3.366, 3.366, and 3.369.
c. The range is 3.369 to 3.363, or a 0.006 difference. They test for conformance by comparing the range of 0.006 with the criterion range of 0.007 (Step 1).
Step 3: Decision on roundness.
Because the washer passed the prescribed test for roundness, it is declared to be round.
If a seller has employees who understand what round means and a buyer who agrees, many
of the problems the company may have had satisfying the customer will disappear.
An Anecdote about Operational Definitions

If you think about the times in your life when you got the angriest at someone (or your
self), it may have involved a failure to operationally define a critical item in a plan. For
example, 50 years ago (before cell phones) my mother was picking up my father at the
airport. At that time it was a 2-hour drive to the airport each way. My parents thought
they agreed on the time and place for the pickup, but the place wasn’t specific enough,
and they missed each other. My mother drove home and my father had to take a 2-hour
taxi ride. They didn’t speak for a week because they were so angry at each other for
messing up what seemed like a simple situation. This certainly made the situation much
worse. Beware of poorly defined operational definitions.

From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 153
Anecdote on the Importance of Operational Definitions

A philosophy teacher gave his students a test. He took his chair and put it onto the
desk and said simply, “your assignment is to prove that this chair does not exist.” The
students worked furiously writing long, detailed explanations, except for one student
who completed his test in 1 minute before handing it in to the surprised looks of his
classmates and teacher. The next week when the teacher gave the tests back, the student
who finished the test in 1 minute got the highest grade. His answer? “What chair?”

Failure Modes and Effects Analysis (FMEA)
What: FMEA is a tool used to identify, estimate, prioritize, and reduce risk of damage being
caused by potential CTQs and Xs or to identify areas of risk to the success of a project (Gitlow
and Levine, 2004).
Why: To study systems to identify what may go wrong with a product, service, or process
and to mitigate the risks of potential problems.
Example: Table 6.2 shows the structure of an FMEA.
Check Sheets
What: Check sheets are used for collecting or gathering data on CTQs or Xs in a logical format, which will be analyzed by the tools and methods discussed in this book (Gitlow et al.,
2015; Gitlow and Levine, 2004).
Why: Check sheets have several purposes, the most important being to enable the user(s)
to gather and organize data in a format that permits efficient and easy analysis. The check
sheet’s design should facilitate data gathering.
Example: Table 6.3 shows an attribute check sheet of pharmacy delay reasons in a chemotherapy treatment unit lab.
From the Library of Pearson HED
154 A Guide to Six Sigma and Process Improvement for Practitioners and Students

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Critical
Parameter
Potential
Failure
mode
Potential
Failure
Effect
Potential
Causes
Current
Controls
Recommended
Action
for an X
(Alternative
setting of an X)
Responsibility,
Contingency
Plan if
Alternative
Setting for
X Fails, and
Target Date
Date
Action
Taken
Severity Occurrence Detection RPN Severity Occurrence Detection RPN
Before RPN =
After RPN =

Table 6.2 Format for an FMEA
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 155
Table 6.3 Attribute Check Sheet

Monday
8-10
a.m.
10-12
a.m.
12-2
p.m.
2-4
p.m.
Tuesday
10-12
a.m.
12-2
p.m.
2-4
p.m.
Total
Type of Defect 8-10
a.m.
Missing dose || ||| | ||| | | 12
Wrong date of
service
| ||| ||| ||| ||| | ||| | 19
Missing height
and weight
||| ||| || ||| | || || 17
Order expired ||| ||| | |||| || || ||| ||| |
||||
|| 32
Total 9 3 18 11 12 6 14 7 80

As you can see from Table 6.3 , order expired is the most common defect.
Brainstorming
What: Brainstorming is a process used to elicit a large number of ideas from a group of people in a brief amount of time (Gitlow et al., 2015; Gitlow and Levine, 2004). Team members
use their collaborative thinking power to generate unlimited ideas and thoughts.
Why: Brainstorming is used to
Identify relevant problems to address.
Identify causes of a particular problem.
Identify solutions to a particular problem.
Identify strategies to implement solutions.
Example: Consider a group of eight people, one from each department of a hospital, who
brainstorm about the problem of excessive employee absenteeism. They have already decided
on the topic to be discussed, so they can proceed to making their lists of causes. After completing their lists, they take turns reading their ideas, sequentially, one at a time. The designated leader records the ideas on a flip chart.
The first person’s list of possible causes of excessive employee absenteeism is
1. Low pay.
2. No repercussions for missing work.
3. Job is boring.
4. Family problems.
From the Library of Pearson HED
156 A Guide to Six Sigma and Process Improvement for Practitioners and Students
The second person’s list is
1. Supervisor is a jerk.
2. Drug problems in the organization.
3. Don’t like their job.
4. Unsafe environment.
5. Don’t like coworkers.
The rest of the team have similar lists. Everyone reads their lists aloud and the causes are
recorded. The leader then asks whether any new ideas have been sparked by piggybacking
on one of the first person’s causes. “Family problems” might elicit another cause, “personal
problems.” Asking for wild ideas might generate a response such as “addiction to the Internet” or “crappy food in the cafeteria.”
Once all the ideas have been discussed, each member of the team receives a copy of the list
of ideas to study. The team meets again and evaluates the ideas by adding or dropping ideas.
Affinity Diagrams
What: An affinity diagram is a tool used by teams to organize and consolidate a substantial
and unorganized amount of verbal and/or pictorial data relating to a problem; frequently
from a brainstorming session or from VoC interviews (Gitlow et al., 2015).
Why: Affinity diagrams help organize data that comes from brainstorming sessions, VoC
interviews, or other sources into natural clusters that help expose the latent structure of the
problem being studied.
Example: A team of employees in an organization addressed the question: “Why are call
center employees leaving the job?” The team members recorded their ideas on cards. They
then placed the cards on a table and everyone simultaneously moves the cards into clusters
that are thematically similar:
in silence! Silence is important so the team members don’t exert
influence over which card should go in which cluster. Figure 6.3 shows the resulting affinity
diagram.
In Figure 6.3 the team’s view of the problems in call center employees leaving the job are
given by the header cards:
1. Chairs not comfortable.
2. Not satisfied with job.
3. Poor management.
A detailed study of these three categories will help the group understand why call center
employees keep leaving the job.
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 157
Cause and Effect (Fishbone) Diagrams
What: A cause and effect (C&E) diagram (Gitlow et al., 2015), also known as a fishbone diagram, is a tool that is used to
Organize the potential causes of a problem (called an effect), usually because there are
so many potential causes of a given problem.
Select the most probable causes of the problem (effect).
Verify the cause and effect relationship between the problem (effect) and the most
probable cause(s).
Why: People are frequently overwhelmed by the number of causes related to a problem, a
cause and effect diagram helps a team identify the various causes and narrow them down to
the most probable cause(s).
Example: Figure 6.4 is an example of a cause and effect (fishbone) diagram to understand
reasons that patients are not showing up to their appointments in a hospital clinic. The
causes in the fishbone diagram could have come from a brainstorming session on the causes
of no shows to a hospital clinic. The team studies the C&E diagram and circles the most
likely causes (Xs) of the effect/problem (CTQ); see Figure 6.4. Figure 6.4 shows that the team
believes that “transportation to the clinic,” “physician,” and “new vs. established patients”
are the most likely causes (Xs) of “no shows at the clinic” (CTQ).
Brainstorming: Why are call center employees leaving the job?
CHAIRS NOT
COMFORTABLE
My back was
always sore.
I had to take
frequent breaks to
get up and walk
around.
I was always stiff.
I was always
readjusting my chair.
NOT SATISFIED
WITH JOB
I was always
checking the clock.
One word: boring!
I had no intrinsic
motivation to do
better.
My coworkers were a
bunch of weirdos.
POOR MANAGEMENT
I never know what is
expected of me.
They always want more
and more from me.
I am doing my best but it
never seems like enough.
My boss couldn’t
motivate a bear to
do #2 in the woods.
Figure 6.3 Affinity diagram
From the Library of Pearson HED
158 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Mother nature
Machines
Materials
Methods
Man (people) Measurements
Physician
Scheduler
Reminders
Transportation
to clinic.
Age and time
New vs. established
patients
Fishbone Diagram to Understand Reasons for Patient No Shows to a Hospital Clinic

Season
Day of week, Diagnosis
month of year
Weather How far out?
Insurance
authorization
received.
EFFECT: Too many no shows
in outpatient clinic at XYZ Hospital.
Figure 6.4 Cause and effect (fishbone) diagram
Another example: Figure 6.5 is an example of a cause and effect (fishbone) diagram to
understand why our friend Joe is still single:
Machines Materials
Methods
Man
EFFECT: Joe is single.
Refers to himself in
third person.

Has “lucky garter” hanging
from rearview mirror.

 

Didn’t forward chain
email back in 2008.

Two words: monthly
showers

Plays bagpipes for
his dates.

Wears a tux on first
dates.

Does not remove Bluetooth
device while making love.

 

Puts emoji in
handwritten letters.

 

His go-to pick-up line is,
“What’s up with the long face?”

 

Uses riddles to break
the ice.

 

Sleeps on Superman sheets.

 

Has stuffed animals
on his bed.

Figure 6.5 Cause and effect (fishbone) diagram on why Joe is still single
Please note that Figures 6.4 and 6.5 use materials, machines, mother nature, methods, man
(people), and measurements as the names of the major fishbones. Another method for naming the major fishbones is to use the names of the clusters from an affinity diagram. If you
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 159
refer back to Figure 6.3, you could construct a C&E diagram with the major fishbone categories of “chairs not comfortable,” “not satisfied with job,” and “poor management.”
Pareto Diagrams
What: Pareto diagrams are used to identify and prioritize issues (Xs) that contribute to a
problem (CTQs); it is mainly a prioritization tool. The Pareto principle is frequently called
the 80-20 rule; 80% of your problems (CTQs) come from 20% of your potential causes (Xs).
Pareto analysis focuses on distinguishing the vital few significant causes of problems from the
trivial many causes of problems. The vital few are the few causes that account for the largest
percentage of the problem (80%), while the trivial many are the myriad of causes that account
for a small percentage of the problem (20%). The causes of the problem become Xs, and the
problem is the CTQ(s) (Gitlow et al., 2015; Gitlow and Levine, 2004).
Why: The Pareto principle focuses attention on the significant few causes (Xs) instead of the
trivial many causes. Consequently, we are able to prioritize efforts on the most important
causes (Xs) of the problem to make the most efficient use of our time and resources.
Example: The director of a hospital pharmacy is interested in learning about the reasons on
delays to orders for the hospitals chemotherapy treatment unit. He collects data for a month
on the reasons for delays that are seen in Table 6.4 .
Table 6.4 Pharmacy Delay Reasons

Delay Reason Number
Missing D.O.S. 74
Missing height and weight 66
Dose change 15
Order clarification 9
No consent form 7
Labs pending 6
Labs high 2

After collecting data and creating a Pareto diagram in Figure 6.6 , he sees that missing date
of service (D.O.S.) and missing height and weight on the orders account for 78.2% of delays.
He has the IT folks put a “hard stop” on the order entry system that prohibits orders being
placed without a date of service or height and weight. He collects more data the next month
and as expected sees a drastic reduction in order delays.
Gantt Charts
What: A Gantt chart is a simple scheduling tool. It is a bar chart that plots tasks and subtasks
against time. It clarifies which tasks can be done in parallel and which tasks must be done
serially (Gitlow et al., 2015).
From the Library of Pearson HED
160 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Why: Once a list of tasks and subtasks has been created for a project, responsibilities can be
assigned for each. Next, beginning and finishing dates can be scheduled for each task and
subtask. Finally, any comments relevant to a task or subtask are indicated on the Gantt chart;
see Table 6.5 .
Example: Table 6.5 shows task 3 begins in March and ends in April, while task 4 begins May
and ends in November. This Gantt chart shows that three tasks begin in May.
Change Concepts
What: Change concepts are approaches to changing the Xs (steps in a process flowchart) that
have been found to be useful in developing solutions that lead to improvements in processes.
By creatively combining these change concepts with process knowledge, teams can develop
changes that lead to substantial improvement. The change concepts were developed by Associates in Process Improvement (Langley et al., 1996; Gitlow et al., 2015).
Figure 6.6 Delay reasons Pareto diagram
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 161
Table 6.5 Gantt Chart

Time Line (Month)
A M J J A S O N D J F M A M J
Comments
Tasks
1
Responsibility
HG
2 BJ B E
3 RM B E
4 HG B E E
5 RM B E
6 BJ B B
7 HG E
8 RM B E

J F M B E
From the Library of Pearson HED
162 A Guide to Six Sigma and Process Improvement for Practitioners and Students
The 70 change concepts listed here are organized into the following nine general groupings:
Number of Concepts in

Grouping Grouping
A. Eliminate Waste 11

B. Improve Work Flow 11
C. Optimize Inventory 4
D. Change the Work Environment 11
E. Enhance the Producer/Customer Relationships 8
F. Manage Time 5
G. Manage Variation 8
H. Design Systems to Avoid Mistakes 4
I. Focus on the Product/Service 8
All 70 change concepts are listed and described in the second part of this chapter.
Why: Change concepts allow process improvers to develop, test, and implement changes
(new Xs or new settings for Xs). This ability is paramount to the success of anyone who wants
to constantly improve a process. Typically a specific and unique change is needed to obtain
improvement in a specific set of circumstances.
Example: College students need access to accurate and timely information on their accounts
payable with the university they are attending.
The goals of Student Account Services are to first satisfy student account information needs
electronically, then by phone, and finally, if necessary, in person. This approach enables the
university to improve both customer service quality and employee productivity.
A quality improvement team analyzed the total number of calls and corresponding abandoned calls, which were at an all-time high, through the call center. The team studied the list
of 70 change concepts and concluded that the 13th change concept, “schedule into multiple
processes,” could be helpful in reducing the percentage of abandoned calls.
The team developed this change concept into hiring inexpensive undergraduate student
assistants who could answer basic phone questions and perform triage for complex inquiries.
If an inquiry required a higher level of account expertise than that possessed by the student
assistant, the assistant would forward the call to a full-time customer representative. The
change concept was put into place in August, and the number of abandoned calls dropped
substantially as can be seen in Table 6.6 and the line graph in Figure 6.7 .
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 163
Table 6.6 Abandoned Calls Per Month

Month Abandoned Calls
January 132
February 150
March 146
April 166
May 141
June 139

 

Month Abandoned Calls
July 173
August 35
September 29
October 31
November 22
December 19

Figure 6.7 Line graph of abandoned calls per month
Communication Plans
What: A communication plan is created for a project to identify, inform, and appease concerns regarding a process improvement event because many people typically are affected
by it.
Why: Implementing new processes typically involves the need for employees to change their
behaviors. Having a communication plan allows the team to be proactive in informing people
in the organization as to what is going on so there are no surprises and so no one is caught
off guard when it is time to implement solutions.
Example: Table 6.7 shows a communication plan.
From the Library of Pearson HED
164 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Table 6.7 No Shows Project Communication Plan

#
1
Event/
Communication
General announce
ment memo to staff
Participants/
Audience
All staff in
Department
of Behavioral
Health
Medium
Email and
fax blast
Frequency
Once
When
TBD
Lead
AVP, BH
Scheduled?
n/a
Status
TBD
Notes
Draft communication to
inform staff of project, sent
by AVP and CEO
2 Weekly email to
stakeholders
All project
stakeholders
Email Weekly Every
Friday
Black Belt n/a Ongoing Black Belt to delegate to team
member
3 Meetings with opin
ion leaders
Opinion
leaders
Face to face Once TBD Black Belt N TBD Identify opinion leaders and
work with them to promote
the project
4 Poster placed in
clinic lunch room
All staff in
Department
of Behavioral
Health
Printed
poster
Once TBD Communications
Director
N TBD Need to create and place
poster in clinic lunch room
5 Presentation at phy
sician staff meeting
All physicians
in Department
of Behavioral
Health
Face to face Once TBD Black Belt N TBD Need to create presentation
and schedule time to present
6 Update hospital
intranet site
All staff in
Department
of Behavioral
Health
Intranet
website
Once TBD Black Belt N TBD Create content and meet with
webmaster to post to intranet

From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 165
How-To Guide for Using Non-Quantitative Tools and
Methods
This section explains in detail how to use in practice the tools and methods described briefly
in the first section of this chapter. Section one of this chapter is the “what” and “why,” with
an example, for each tool and method, whereas section two is the “how to” for each of the
tools and methods discussed in this chapter.
How to Do Flowcharting
As discussed in Chapter 3 , the steps in creating a flowchart are described in Table 6.8 . The
American National Standards Institute, Inc. (ANSI) approved a standard set of flowchart
symbols used for defining and documenting a process. The shape of the symbol and the
information written within the symbol provide information about that particular step or
decision in a process.
Table 6.8 Flowchart Symbols and Functions

Symbol Function
Start/stop symbol The general symbol used to indicate the beginning and end of a process is
an oval.
Basic processing symbol The general symbol used to depict a processing operation is a rectangle.
Decision symbol A diamond is the symbol that denotes a decision point in the process. This
includes attribute type decisions such as pass-fail, yes-no. It also includes
variable types of decisions such as into which of several categories a process
measurement falls.
Flowline symbol A line with an arrowhead is the symbol that shows the direction of the stages
in a process. The flowline connects the elements of the system.

Successful flowcharting efforts have the commitment (not only support) of top
management, the process owner, the employees who work in the process, and any
other key process stakeholders. Commitment is required because if you are going
to flowchart a process there must be a reason for the effort, such as documentation
or improvement of the process. Both documentation and improvement require the
commitment of all the stakeholders of the process.
From the Library of Pearson HED
166 A Guide to Six Sigma and Process Improvement for Practitioners and Students
There must be a rationale for defining, documenting, or improving a process. The
employees who work in the process and the process owner must understand that
flowcharting the process is a positive exercise aimed at making things better and that
no one will be blamed for problems in the process.
Stakeholders of a process must agree on the starting/stopping points of the process,
as well as the objectives and metrics used to measure success.
The employees who are flowcharting the process must have enough time so that they
are not rushed and can do a thorough job.
The employees who are flowcharting the process start by having the whole team walk
the process from front to back (process owner’s point of view) and back to front (customer’s point of view).
The employees who are flowcharting the process need to determine how detailed the
flowchart should be, as well as the type of flowchart to be used by the employees. A
flowchart needs to be detailed enough to be able to uncover problems. This means
that some sections of a flowchart will be at a high level and others will be at a detailed
level.
One possible method for flowcharting a process is to write the process steps on Postit notes and then place them on large sheets of paper taped to the wall for all to see.
Next, the steps are arranged in the order that reflects the flow of the process, and more
detail and decision symbols are added until the “as is” process is reflected on the wall.
Finally, connector lines showing the direction of flows in the process are added onto
the large sheet of paper. Next, the flowchart on the wall is replicated in an electronic
format such as Visio. Finally, all the stakeholders of the process verify and validate to
obtain full agreement that the flowchart captures the essence of the process.
Refer to Figure 6.1, which shows an example of a process flowchart created using the symbols
in Table 6.8 .
Deployment flowcharts differ from process flowcharts because they add either rows or columns with headings listing departments, employees, or stages of the process. Refer back to
Figure 6.2 ; it shows that the different steps in the process that take place in different areas of
the organization are delineated by the rows labeled GI Clinic, GI Path Lab, and Pathology
Department. The handoffs in the process from one department/area to another are visible by
the lines that cross from one row to another. Again, the starbursts indicate potential problematic areas to be improved in the process.
How to Do a Voice of the Customer (VoC) Analysis
How to Define/Segment the Market
The first thing we need to do is figure out from whom are we going to get our VoC data; that
is, who are the stakeholder segments to be interviewed that can provide input on the CTQs
of interest.
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 167
The simplest method for segmenting a market is to study the SIPOC analysis and focus attention on the outputs and on the customers by asking
What are the outputs of your process, both intended and unintended?
Who are the customers of those outputs?
Are there particular groups of customers whose needs focus on specific outputs?
Today? Tomorrow?
How to Plan the VoC
Next we have to plan the VoC. There are two types of VoC data: reactive data and proactive
data.
Reactive data arrives regardless of whether the organization collects it, for example, customer complaints, product returns or credits, contract cancellations, market share changes,
customer defections and acquisitions, customer referrals, closure rates of sales calls, web page
hits, technical support calls, and sales, to name a few.
Proactive data arrives only if it is collected by personnel in the organization: This is the type of VoC data we need to collect. It is
data obtained through positive action, for example, data gathered through interviews, focus
groups, surveys, comment cards, sales calls, market research, customer observations, and
benchmarking. Some of the steps involved in planning the VoC are the following:
Choosing stakeholders within each stakeholder segment to interview; for example,
customer segments, employee segments, investor segments, regulatory segments,
environmental segments, building and grounds, and so on .
Creating the interview questions .
Setting up the interviews .
Assigning interviews to team members .
How to Collect the Data
This step involves conducting the interviews to collect VoC data from your stakeholder segments. A few things to keep in mind when conducting the interviews:
Record their answers in bullet point form; it makes the data easier to organize later.
Let the interviewee talk; this is time for you to understand her perspective on the process. However, don’t let the interview become a therapy session; gently stay on task.
Write down exactly what they say in their words; you don’t want anything to get lost
in translation. This is important! If profanity is used, record it. It may be an indicator
of the degree of emotion felt by a person being interviewed.
How to Organize and Interpret the Data
The last step involves organizing and interpreting all the VoC data you have collected. Focus
points are the underlying themes for one or more raw VoC statements. Affinity diagrams are
used to create focus points.
From the Library of Pearson HED
168 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Example: Voice of the Customer Analysis for Patient No Shows
at an Outpatient Psychiatric Clinic
Define/segment
the market. Plan the VoC. Collect the data. Organize and interpret the data.
Define/Segment the Market
To define and segment the market for the outpatient psychiatric clinic discussed earlier, the
team went back to the SIPOC analysis they had created and looked at the suppliers and customers. They decided to focus on the following segments during the VoC to help them better
understand the process. The team members decided that each of the following segments form
a stakeholder group; in other words, there are no subsegments of patients or physicians. This
is frequently not the case.
Patients /family
Appointment schedulers
Insurance providers
Patient registration reps in clinics
Nurses
Physicians
Finance office
Plan the VoC
Next up was planning the VoC. The team decided to focus on collecting proactive VoC data
since there was no reactive VoC data available for this process. Next, they did the following:
Choosing Stakeholders within Each Market Segment to Interview
The team decided that they would take a judgment sample (expert opinion) of the following
stakeholder groups to interview. The stakeholders selected in each group were deemed most
appropriate to reflect the views of that group by the team doing the VoC, perhaps with the
help of an expert. The team members decided that they could afford to interview the following number of stakeholders within each segment; it would give them a great understanding
of the process:
Five patients and family members
Two appointment schedulers
Two insurance providers
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 169
Two patient registration reps in clinics
Five nurses
Five physicians
One finance office
As stated previously, the sample size in each segment is a function of the budget available to
perform the VoC. It is an arbitrary number that in the opinion of an expert yields reliable
information about that segment.
Creating Interview Questions
The team then decided on questions to ask each interviewee related to the patient scheduling
process in the Outpatient Psychiatric Clinic at XYZ Hospital. The questions are designed to
understand the process as much as possible:
How do you feel about the patient scheduling process as it relates to patient no shows?
What issues/problems do you see with the patient scheduling process that may lead
to patient no shows?
What solutions/recommendations do you have to decrease the number of patient no
shows?
What feelings or images come to mind when you think about patient no shows?
This last question is important because it captures the circumstance surrounding the
process.
Setting Up Interviews
The process expert assigned one of the team members the task of setting up the interviews
with the various stakeholders making sure to follow the timeline set out by the Black Belt in
the project plan.
Assigning Interviews to Team Members
The process expert then assigned interviews to different team members based on his understanding of each team member, their personality, and their relationships with various
stakeholders.
Collect the Data
The various team members then went out and conducted the interviews to collect VoC data
from their respective stakeholders. The process expert gave the team specific instructions:
Record their answers in bullet point form to make the data easier to organize later.
Let them talk. This is time for you to understand their perspective on the process.
Write down exactly what they say; you don’t want anything to get lost in translation.
From the Library of Pearson HED
170 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Table 6.9 Voice of the Customer Summary Table

Selected
Market
Segment
Raw VoC Data Affinity
Diagram Theme
(Focus Point)
Driving Issue CTQ
Patients “I can’t get an appointment for 3
months. I get here and the wait
ing room is empty. No wonder I
am crazy!” (1)
Note: The number one (1)
indicates the thematic group
the VoC raw data point falls
into; in this case variation
in patients showing up for
their appointments. As you
will see, several raw VoC
data points fall into this
thematic group.
Variation in
patients show
ing up for their
appointments. (1)
Patients not
showing up for
their appoint
ments affects
everyone.
Patient no show
rate, by month.
“They don’t take no shows
seriously, so I don’t feel bad if
something else comes up.” (1)
Variation in sched
uling patients for
appointments. (2)
Time between
date patient
calls to schedule
appointment and
date of appoint
ment is too long.
Turnaround
time to schedule
appointments, by
patient.
“Everyone else seems to miss
appointments and they aren’t
held accountable, so I don’t feel
bad if I do.” (1)
“Why does it take so long to get
an appointment here?” (2)
Appointment
schedulers
“These no shows make our job
so hard.” (1)
“Tough to predict who will no
show and who won’t.” (1)
“The variation in who shows
and who doesn’t makes it tough
to schedule.” (1)
“We hear complaints from
everyone. I am glad they are
finally doing something about
it!” (1)
Insurance
providers
“No shows seem to be a problem
that a lot of hospitals struggle
with.” (1)
“They aren’t good for anybody.”
(1)

From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 171

Selected
Market
Segment
Raw VoC Data Affinity
Diagram Theme
(Focus Point)
Driving Issue CTQ
Patient regis
tration repre
sentatives
“We see a lot of frustration
when patients no show—from
the docs to the nurses to other
patients to staff. We need to
decrease them asap!” (1)
“Sometimes we are busy, some
times we aren’t.” (1)
“It makes it hard for our boss
to schedule us. I feel bad for her
sometimes.” (1)
“Sometimes we get sent home if
there are too many no shows. I
get it but it still sucks.” (1)
Nurses “I have physicians giving me
crap for times they have nothing
to do, like it’s my fault!” (1)
“Sometimes I am super busy,
sometimes it’s slow.” (1)
“I hear patients complaining
that they had to wait a few
months to get in. I have no idea
why.” (2)
Physicians “I am the most expensive
resource and half the time I feel
like I am sitting on my butt.” (1)
Finance office “It definitely has a financial
impact, and with the way things
have been going we need every
penny.” (1)
“They are directly affecting our
bottom line.” (1)
“We need to figure out how to
decrease them and fast!” (1)

Organize and Interpret the Data
Team members analyzed the VoC data in Table 6.9 by the market segments (see column 1 in
Table 6.9 ). Next, they used all the raw VoC data points (see column 2 of Table 6.9 ) to create
thematic groups indicated by the numbers in parentheses in each cell that are summarized in
column 3 of Table 6.9 . Next, team members identified the issue underlying each focus point,
From the Library of Pearson HED
172 A Guide to Six Sigma and Process Improvement for Practitioners and Students
called driving issues (see column 4 in Table 6.9 ). Then team members converted each cognitive issue into one or more quantitative variable, called critical-to-quality (CTQ) variables
(see column 5 in Table 6.9 ).
The team identified two CTQs via the VoC interviews: patient no show rate by month (which
they kind of knew all along would be a CTQ) and turnaround time to schedule appointments
by patient. They decided to focus on the first CTQ, patient no show rate, because they agreed
that turnaround time to schedule appointments could be a function of the no show rate.
How to Do a SIPOC Analysis
To create a SIPOC analysis the team fills in each of the five columns in the SIPOC table in
Table 6.10 by answering the questions in the following sections.
Suppliers
Team members identify relevant suppliers by asking the following questions:
Where does information and material come from?
Who are the suppliers?
Is the Human Resources department a source of inputs into the process (employees)?
Inputs
Team members identify relevant inputs by asking the following questions:
What do your suppliers give to the process?
What effect do the inputs (Xs) have on the process?
What effect do the inputs (Xs) have on the CTQs (outputs)?
What effect do employees (Xs) have on the CTQ(s)?
Process
Team members create a high level flowchart of the process taking particular care to identify
the beginning and ending points of the process.
Outputs:
Team members identify outputs of the process by asking the following questions:
What products or services does this process make, both intended and unintended?
What are the outputs that are critical to the customer’s perception of quality? Or lack
of quality? These outputs are called critical-to-quality (CTQ) characteristics of the
process.
Are there any unintended outputs from the process that may cause problems?
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 173
Customers
Team members identify relevant customers (market segments) by asking the following
questions:
Who are the customers (stakeholders) or market segments of this process?
Have we identified the outputs (CTQs) for each market segment?
See Table 6.10 for the layout of a SIPOC analysis.
Table 6.10 SIPOC Analysis

Process name:
Process owner:
SUPPLIERS INPUTS PROCESS OUTPUTS CUSTOMERS
(Providers
of required
resources)
(Resources
required by the
process)
(Deliverables from the
process)
(Stakeholders who
put requirements on
the outputs)
Here we list all
the suppliers of
the inputs to the
process.
Here we list all
the inputs into
the process.
This is a high level
flowchart of the
process.
Here we list all the
outputs from the pro
cess, both intented and
unintended.
Here we list all the
customers of the
outputs.

How to Create Operational Definitions
The creation of operational definitions requires the following steps. Remember, all users of
the operational definition must agree with the operational definition or it is of no value in
eliminating misunderstandings and arguments.
Step 1: Determine the criterion for the item being operationally defined. The team (customer and supplier or key stakeholders) must establish the criteria on the item being operationally defined. For example “the blanket must be at least 50% wool.” In other words, a
random sample of 15 one inch by one inch pieces of blanket are selected, and if all 15 have at
least 50% wool, the blanket is considered a wool blanket.
Step 2: Test for the item being operationally defined. Next, the team selects the sample of
15 one inch by one inch pieces of the blanket and has them tested by a wool expert to analyze
the proportion of wool in each piece.
Step 3: Decision for the item being operationally defined. Finally the team must decide
whether the test to determine if the item being operationally defined meets the criteria established. For example, if all 15 one inch pieces of wool analyzed by the wool expert are 50%
wool or greater, then the blanket is 50% wool.
From the Library of Pearson HED
174 A Guide to Six Sigma and Process Improvement for Practitioners and Students
An Anecdote about Operational Definitions

At one time I was consulting for a paper mill. The nearest lodgings to the paper mill
was a pretty low class motel. I walked into my room and the maid was cleaning the
toilet with a toilet bowl brush. I was glad to see that. However, then she used the same
brush to clean the toilet seat and used a rag to dry the seat. Finally, she put the strip of
paper around the toilet seat that says sanitized. Clearly, the maid and I had different
operational definitions of sanitized.

How to Do a Failure Modes and Effects Analysis (FMEA)
There are nine steps to conducting an FMEA:
1. Identify the critical parameters (Xs) and their potential failure modes identified in the
cause and effects matrix or diagram through brainstorming or other tools—that is,
ways in which the process step (X) might fail (columns 1 and 2 of Table 6.11 ).
2. Identify the potential effect of each failure (consequences of that failure) and rate its
severity (columns 3 and 4 of Table 6.11 ). The definition of the severity scale is shown
in Table 6.12 .
3. Identify causes of the effects and rate their likelihood of occurrence (columns 5 and 6
of Table 6.11 ). The definition of the likelihood of occurrence scale is shown in Table
6.13 .
4. Identify the current controls for detecting each failure mode and rate the organization’s ability to detect each failure mode (columns 7 and 8 of Table 6.11 ). The definition of the detection scale is shown in Table 6.14 .
5. Calculate the RPN (risk priority number) for each failure mode by multiplying the
values in columns 4, 6, and 8 (column 9 of Table 6.11 ).
6. Identify the action(s) and/or contingency plans for an alternative setting of an X that
will improve the distribution of the CTQ, identify person(s) responsible to implement and maintain the alternative setting of the X, and target completion dates for
reducing or eliminating the RPN for each failure mode (columns 10 and 11 of Table
6.11 ). Actions are the process changes needed to reduce the severity and likelihood
of occurrence, and increase the likelihood of detection, of a potential failure mode;
they are change concepts. Contingency plans are the alternative actions immediately
available to a process owner when a failure mode occurs in spite of process improvement actions. A contingency plan might include a contact name and phone number
in case of a failure mode.
7. Identify the date the action was taken to reduce or eliminate each failure mode
(column 12 of Table 6.11 ).
8. Rank the severity (column 13 of Table 6.11 ), occurrence (column 14 of Table 6.11 ),
and detection (column 15 of Table 6.11 ) of each failure mode after the recommended
action (column 10 of Table 6.11 ) has been put into motion.
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 175
Table 6.11 Failure Modes and Effects Analysis

1
Critical
Parameter
Lack of buy
in by top
manage
ment.
2
Potential
Failure
Mode
Project
doesn’t get
supported.
3
Potential
Failure
Effect
Project
unexecutable.
4
SEV
10
5
Potential
Causes
Lack of
commit
ment.
6
OCC
10
7
Current
Controls
None
8
DET
10
9
RPN
1000
10
Recommended
Action
Top manage
ment must
have skin in the
game.
11
Responsibility
and Target
Date
CEO.
12
Date
Action
Taken
6/2/2019
13
SEV
2
14
OCC
4
15
DET
10
16
RPN
80
Project
exceeds
budget.
Takes too
long to get
data.
Black Belt
gets fired!
Project is
delayed.
Need to hire
new BB.
Heads roll.
3
8
BB lacks
financial
expertise.
BB
lacks IT
expertise.
8
8
BB does
his best.
BB does
his best.
5
5
120
320
Finance rep on
team responsible
for budget.
IT rep on team
responsible for
data.
Finance rep.
IT rep.
6/2/2019
6/2/2019
2
4
5
4
5
5
50
80
Lack of buy
in from
physicians.
Implemen
tation issues.
Unsuccessful
implementa
tion.
7 Entitled
physi
cians.
8 Lots of
bitching
back and
forth.
8 448 Physician
Champion part
of team.
Physician
Champion.
6/2/2019 4 6 8 192
Lack of buy
in from
nurses.
Implemen
tation issues.
Unsuccessful
implementa
tion.
7 Entitled
physi
cians.
8 Lots of
bitching
back and
forth.
8 448 Nursing
Champion part
of team.
Nursing
Champion .
6/2/2019 4 6 8 192
Lack of
awareness.
Lack of
cooperation.
Misaligned
team.
5 Lack of
commu
nication.
5 Rumor
mill in
overdrive.
2 50 Communication
plan created.
Black Belt. 6/2/2019 1 5 2 10
Scope too
broad.
Project
becomes
unmanage
able.
Project
unexecutable.
7 Trying to
boil the
ocean.
7 Planning
by Black
Belt.
3 147 Strong planning
effort by Black
Belt.
Black Belt. 6/2/2019 3 3 3 27

From the Library of Pearson HED
176 A Guide to Six Sigma and Process Improvement for Practitioners and Students
9. Multiply the values in columns 13, 14, and 15 of Table 6.11 to recalculate the RPN for
each failure mode after the recommended action (column 10 of Table 6.11 ) has been
put into motion.
If the after RPN is significantly lower than the before RPN for the RPNs with large before
values, you have likely hit upon a good change concept, or modification to an X (step in
the flowchart). You can see this occurring in “lack of physician buy-in” and “lack of nurse
buy-in.”
Table 6.12 Definition of “Severity” Scale = Likely Impact of Failure

Impact Rating Criteria: A Failure Could…
Bad 10 Injure a customer or employee
9 Be illegal
8 Render the unit unfit for use
7 Cause extreme customer dissatisfaction
6 Result in partial malfunction
5 Cause a loss of performance likely to result in a complaint
4 Cause minor performance loss
3 Cause a minor nuisance; can be overcome with no loss
2 Be unnoticed; minor effect on performance
Good 1 Be unnoticed and not affect the performance

Table 6.13 Definition of “Occurrence” Scale = Frequency of Failure

Impact Rating Time Period Probability of Occurrence
Bad 10 More than once per day > 30%
9 Once every 3-4 days < = 30%
8 Once per week < = 5%
7 Once per month < = 1%
6 Once every 3 months < = .3 per 1,000
5 Once every 6 months < = 1 per 10,000
4 Once per year < = 6 per 100,00
3 Once every 1-3 years < = 6 per million (approx. Six Sigma)
2 Once every 3-6 years < = 3 per ten million
Good 1 Once every 6-100 years < = 2 per billion

From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 177
Table 6.14 Definition of “Detection” Scale = Ability to Detect Failure

Impact Rating Definition
Bad 10 Defect caused by failure is not detectable.
9 Occasional units are checked for defects.
8 Units are systematically sampled and inspected.
7 All units are manually inspected.
6 Manual inspection with mistake proofing modifications.
5 Process is monitored with control charts and manually inspected.
4 Control charts used with an immediate reaction to out of control condition.
3 Control charts used as above with 100% inspection surrounding out of control condition.
2 All units automatically inspected or control charts used to improve the process.
Good 1 Defect is obvious and can be kept from the customer or control charts are used for pro
cess improvement to yield a no inspection system with routine monitoring.

How to Do Check Sheets
Check sheets are used in Six Sigma projects to collect data on CTQs and Xs in a format that
permits efficient and easy data collection and analysis by team members. Three types of check
sheets are discussed: attribute check sheets, measurement check sheets, and defect locations
check sheets.
Attribute Check Sheets
An attribute check sheet is used to gather data about defects in a process. The logical way to
collect data about a defect is to determine the number and percentage of defects generated
by each cause. To create an attribute check sheet simply list the different types of defects and
then tally the number of occurrences for each over a relevant test time period.
Table 6.15 is an example of a defect check sheet for an outpatient scheduling call center for
a health system. This check sheet was created by tallying each type of call defect during four
2-hour time periods for one week. Keeping track of these data provides management with
information on which to base improvement actions, assuming a stable process.
From the Library of Pearson HED
178 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Table 6.15 Outpatient Call Center Defects

Type of Defect Frequency Percentage
Improper use of English language 2 2.9
Grammatical errors in speech 6 8.6
Inappropriate use of words 3 4.3
Rude response 3 4.3
Didn’t know answer to patient’s question 25 35.7
Call took too much time 30 42.8
Not available 1 1.4
Total 70 100

Measurement Check Sheets
Gathering measurement type data about a product, service, or process involves collecting
information such as revenue per month, cost per quarter, cycle time per unit produced, waiting time by customer, temperature, size, length, weight, and diameter by item. This data is
best represented on a frequency distribution, also called a
measurement check sheet.
Table 6.16 is a measurement check sheet showing the frequency distribution of cycle times
to answer 508 patient calls to make appointments (how long the patient is kept on hold with
horrible muzak playing on the phone) in an outpatient scheduling call center for a health
system that came into the call center between 8:00 a.m. and 5:00 p.m. on January 16, 2015.
Table 6.16 Frequency Distribution of Cycle Times

Cycle Time (Minutes) Frequency
05 < 10 minutes 36
10 < 15 minutes 178
15 < 20 minutes 233
20 < 25 minutes 53
25 < 30 minutes 8
Total 508

This type of check sheet is a simple way to examine the distribution of a CTQ or X and its
relationship to specification limits (the boundaries of what is considered an acceptable cycle
time). The number and percentage of items outside the specification limit are easy to identify
so that appropriate action can be taken to reduce the number of defective calls, perhaps using
a process improvement project. For example, suppose the call center has a policy that no
call can last more than 20 minutes. If this policy is in place, 12% (61/508) of call cycle times
are out of specification. This may help employees and management improve the process,
assuming it is stable, while both keeping calls under 20 minutes and giving callers satisfactory answers.
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 179
Defect Location Check Sheet
Another way to gather information about defects in a product is to use a defect location check
sheet. A
defect location check sheet is a picture of a product (or a portion of it) on which a
relevant employee indicates the location and nature of a defect.
Figure 6.8 is an example of a check sheet for collecting data regarding defects on a cube. It
shows a defect in the top left-hand corner of the cube (in this case a dent); the location of
the defect is marked with an “x.” Suppose that an analysis of multiple check sheets reveals
that many x’s are in the upper left corner on the front face of the cube. If this is so, further
analysis might shed light on the type of defect in the upper left corner. In turn, this might lead
employees to identify the root cause of the defects. This, of course, leads to improvements in
the cube production process.
X
Figure 6.8 Defect location check sheet example
Brainstorming
How: Brainstorming is a way to elicit a large number of ideas from a group of people in a
short period of time about an idea, topic, or problem. Members of the group use their collective thinking power to generate ideas and unrestrained thoughts.
Effective brainstorming should take place in a structured session.
The group should be between 3 and 12 people. No animals, babies, or electronic
devices are allowed; they are too distracting.
Composition of the group should include a variety of people who have different
points of view about the idea, topic, or problem.
The group leader should keep the group focused, prevent distractions, keep ideas
flowing, and record the outputs.
The brainstorming session should be a closed-door meeting with no distractions.
Seating should promote the free flow of ideas; a circle is the best seating arrangement.
The leader should record the ideas so everyone can see them, preferably on a flip
chart, blackboard, or illuminated transparency. Or, each participant can write his own
ideas on 3×5 cards; one idea per 3×5 card.
From the Library of Pearson HED
180 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Procedure
The following steps are recommended prior to a brainstorming session:
1. Select the topic or problem.
2. Send out the topic to all brainstorming participants to ensure all agree on the topic
before arriving at the session.
3. Conduct research on the topic in a library or on the Internet.
4. Prepare a list of the identified ideas and provide a copy to each of the brainstorm participants before the brainstorming session. There is no reason to reinvent the wheel.
The brainstorm participants should add ideas to what is already easily accessible
through research.
5. Establish a time and place for the brainstorming session.
6. Invite all attendees to the session.
7. Remind all attendees to study the list of ideas provided to them on the topic.
8. Ask all participants to add additional ideas for discussion in the brainstorming session
and bring them to the session, say one idea per 3×5 card.
The following steps are recommended
at a brainstorming session:
1. Post the topic or problem.
2. Each group member makes an additional list of ideas about the problem.
3. Each person reads one idea at a time from her list of ideas, sequentially, starting at the
top of the list. Group members continue in this circular reading fashion until all the
ideas on everyone’s lists are read.
4. If a member’s next idea is a duplication of a previously stated idea, that member goes
on to the subsequent idea on his list.
5. After each idea is read by a group member, the leader requests all other group members to think of “new ideas.”
6. Members are free to pass on each go-round but should be encouraged to add
something.
7. If the group reaches an impasse, the leader can ask for everyone’s “wildest idea.”
Rules
Certain rules should be observed by the participants to ensure a successful brainstorming
session—otherwise, participation may be inhibited.
1. Don’t criticize, by word or gesture, anyone’s ideas.
2. Don’t discuss any ideas during the session, except for clarification.
3. Don’t hesitate to suggest an idea because it sounds silly. Many times a “silly” idea can
lead to the problem’s solution.
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 181
4. Don’t allow any group member to present more than one idea at a time.
5. Don’t allow the group to be dominated by one or two people.
6. Don’t let brainstorming become a gripe session.
How to Do Affinity Diagrams
Affinity diagrams are used to analyze verbal or pictorial data. In this text, we analyze verbal
data. For example, affinity diagrams take verbal data (on, say, 3×5 cards) and place them
into thematic groups, such that the data (3×5 cards) in a group are similar thematically to
the other data points (3×5 cards) in the group. They can be used to evaluate data from brainstorming or from VoC Customer interviews.
Constructing an affinity diagram begins with identifying a problem. Team composition
usually consists of the same people who participated in the brainstorming session about the
problem under study or conducted the VoC interviews.
A team should take the following steps to construct an affinity diagram:
1. Select a team member to serve as the group’s facilitator.
2. The facilitator transfers all the ideas generated from a brainstorming session to 3×5
cards; one idea per 3×5 card.
3. The facilitator spreads all the 3×5 cards on a large surface (table) in no particular
order, but all cards face the same direction.
4. In silence, all group members simultaneously move the 3×5 cards into groups (clusters) so the 3×5 cards in a group seem to be related; that is, they have an unspoken
underlying theme or affinity for each other.
5. After the group agrees the clusters are complete (usually 3 to 15 clusters emerge), the
group discusses all the 3×5 cards in each cluster and prepares a header card that sums
up the information for each cluster.
6. The facilitator transfers the information from the cards onto a flip chart, or “butcher
paper” and draws a circle around each cluster.
7. The underlying structure of the problem, usually typified by the names of the header
cards, is used to understand the product or process problem.
Refer to Figure 6.3 ; that example was created using the preceding steps to understand why
call center employees are leaving the job.
In Figure 6.3 the team’s view of the problems in call center employees leaving the job are
given by the header cards:
Chairs not comfortable
Not satisfied with job
Poor management
From the Library of Pearson HED
182 A Guide to Six Sigma and Process Improvement for Practitioners and Students
A detailed study of these three categories may help team members understand why call center
employees keep leaving the job.
How to Do Cause and Effect Diagrams (C&E Diagrams)
The following steps are recommended for constructing a cause and effect diagram.
1. State the problem, which is the effect.
2. Select the team; usually the brainstorming team or the VoC interviewer team(s).
3. Affinitize the data (one data point per 3×5 card) into major causes (“fishbones”) or
use the 6 Ms: materials, machines, mother nature, methods, man, and measurement.
4. Put the major causes on a C&E diagram as the major categories.
5. Add subcauses for each major cause and place them on the C&E diagram.
6. Allow time to ponder the subcauses before evaluating them. You may find some
holes in your cause and effect diagram; fill in the holes by brainstorming or just being
thoughtful.
7. Circle subcauses that are most likely contributing to the problematic CTQ (Y).
8. Verify each potential subcause with data or relevant literature and so on.
Refer to Figure 6.4 ; that figure was created to understand reasons that patients are not showing up to their appointments in a hospital clinic.
How to Do Pareto Diagrams
The following steps are recommended for constructing a Pareto diagram. We illustrate a
Pareto diagram with an example concerning sources of defective data entries for a particular
data entry operator.
1. Establish categories for the data being analyzed. Data should be classified according to
defects, products, work groups, size, and other appropriate categories; a check sheet
listing these categories should be developed. In the following example, data on the
type of defects for a data entry operator will be organized and tabulated.
2. Specify the time period during which data will be collected. Three factors important in
setting a time period to study are (1) selection of a convenient time period, such as one
week, one month, one quarter, one day, or four hours, (2) selection of a time period
that is constant for all related diagrams for purposes of comparison, and (3) selection
of a time period that is relevant to the analysis, such as a specific season for a certain
seasonal product. In the data entry example, the time period is four months—January
through April 2014. In the example, the types of defects are recorded as they occur
during the time period and are totaled, as shown in Table 6.17 .
3. Note: Pareto diagrams are most useful in analyzing the data from a control chart only
if the control chart is in a state of statistical control.
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 183
Table 6.17 Record of Defects for Data Entry Operator: Check Sheet to Determine the
Sources of Defects

(Assumption: If the following data is from a control chart, the percentage of defective entries must be stable
over time for proper use of the Pareto diagram.)
Major Causes of Defective Entries Month 1/12 2/12 3/12 4/12 Total
Transposed numbers 7 10 6 5 28
Out of field 1 2 3
Wrong character 6 8 5 9 28
Data printed too lightly 1 1 2
Torn document 1 1 2 4
Creased document 1 1 2
Illegible source document 1 1
Total 15 20 16 17 68

4. Construct a frequency table arranging the categories from the one with the largest
number of observations to the one with the smallest number of observations. The
frequency table should contain a category column; a frequency column indicating
the number of observations per category with a total at the bottom of the column; a
cumulative frequency column indicating the number of observations in a particular
category plus all frequencies in categories above it; a relative frequency column indicating the percentage of observations within each category with a total at the bottom
of the column; and a relative cumulative frequency column indicating the cumulative
percentage of observations in a particular category plus all categories above it in the
frequency table. An “other” category, if there is one, should be placed at the extreme
right of the chart. If the “other” category accounts for as much as 50% of the total, the
breakdown of categories must be reformulated. A rule of thumb is that the “other”
bar should be smaller than the category with the largest number of observations. The
frequency table for the data entry example, in Table 6.18 , shows that two types of
defects (transposed numbers and wrong characters) are causing 82.4% of the total
number of defective entries.
Table 6.18 Frequency Table of Defects for Data Entry Operator: Pareto Analysis to Determine Major
Sources of Defects

Major Cause of
Defective Entries
Frequency Relative
%
Cumulative
Frequency
Cumulative %
Transposed numbers 28 41.2 28 41.2
Wrong character 28 41.2 56 82.4
Torn document 4 5.9 60 88.3

From the Library of Pearson HED
184 A Guide to Six Sigma and Process Improvement for Practitioners and Students

Major Cause of
Defective Entries
Frequency Relative
%
Cumulative
Frequency
Cumulative %
Out of field 3 4.4 63 92.7
Data printed too lightly 2 2.9 65 95.6
Creased document 2 2.9 67 98.5
Illegible source document 1 1.5 68 100.0
Total 68 100.0

5. Draw a Pareto diagram:
Draw horizontal and vertical axes on graph paper and mark the vertical axis with
the appropriate units, from zero up to the total number of observations in the
frequency table.
Under the horizontal axis, write the most frequently occurring category on the far
left, and then the next most frequent to the right, continuing in decreasing order
to the right. In the data entry example, “transposed numbers” and “wrong character” are the most frequently occurring defects and are positioned to the far left;
“illegible source document” accounts for the fewest defective cards and appears at
the far right of the chart. The rightmost category will be “other.” It may very well
be a significant size bar on the Pareto diagram; remember the rules for the “other”
category we discussed previously.
Draw in the bars for each category. For some applications, this may provide
enough information on which to base a decision, but often the percentage of
change between the columns must be determined. Figure 6.9 displays the bars for
the data entry example.
Number of defective entries due to major causes
Transposed
numbers
Wrong
character
Torn
document
Creased
document
Illegible source
document
Out of field
Data too light
0
15
30
45
60
68
Figure 6.9 Pareto diagram of defective data entries for an operator
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 185
Plot a cumulative percentage line on the Pareto diagram. Indicate an approximate
cumulative percentage scale on the right side of the chart and plot a cumulative
percentage line on the Pareto diagram. To plot the cumulative percentage line, or
cum line, start at the lower left (zero) corner and move diagonally to the top-right
corner of the first column. In our example, the top of the line is now at the 28 level,
as in Figure 6.10 (a). Repeat the process, adding the number of observations in the
second column. In our example, the line rests on the 56 level, as in Figure 6.10 (b).
The process is repeated for each column, until the line reaches the total number of
observations level that includes 100% of the observations, as in Figure 6.10 (c).
Title the chart and briefly describe its data sources. Without information on when
and under what conditions the data were gathered, the Pareto diagram will not be
useful.
Typically a Pareto diagram is drawn using Minitab. See Chapter 4 , “Understanding Data:
Tools and Methods,” for instructions on how to create a Pareto diagram using Minitab.
How to Do Gantt Charts
Recall that a Gantt chart is a tool for scheduling projects. Each task or subtask for a project
is listed on the vertical axis, as are the person(s) or area(s) responsible for its completion;
see columns 1 and 2 in Table 6.5 . The horizontal axis is time, see columns 3 and beyond in
Table 6.5 . It shows the anticipated and actual duration of each task by a bar of the appropriate length. The left end of the bar indicates the earliest start time, and the right end of the bar
indicates the latest stop time for the task.
The Gantt chart is especially useful in determining which tasks can be done in parallel or
which must be serially. This can save a lot of time in completing the project.
A great use of a Gantt chart is to provide an executive charged with a big project a tool to
monitor the project so she doesn’t lose sleep over it. For example, the COO of a hospital is
in charge of building a new wing on time and within budget, but she is not a civil engineer
and has no idea what is supposed to happen and when. All she knows is that her butt is on
the line if the project is late and over budget. One way to calm this executive down is to have
a project manager draw a Gantt chart for the project so all the executive has to do is to check
whether every task is on time and under budget. If not, she can call the person responsible
(see column 2 in Table 6.5 ) for that step and find out what is going on. This turns into better
sleep for the executive!
How to Use Change Concepts
The change concepts presented here are not specific enough to be applied directly to making
improvements. Rather, the concept must be considered within the context of a particular
situation and then turned into an idea for a process improvement. The idea needs to be specific enough to describe how the change can be developed, tested, and implemented in the
particular situation. When describing the change concepts, we tried to be consistent in the
degree of specificity or generality of the concepts. Sometimes, a new idea seems at first to be
From the Library of Pearson HED
186 A Guide to Six Sigma and Process Improvement for Practitioners and Students
a new change concept; but often, upon further reflection, it is seen to be an application of one
of the existing change concepts.
Number of defective entries due to major causes
Cumulative percentage of major
causes of defective entries
Transposed
numbers
Wrong
character
Torn
document
Creased
document
Illegible source
document
Out of field
Data too light
0
15
30
45
60 92.7
98.5
100
88.3
82.4
41.2
68
Number of defective entries due to major causes
(b) Bar Two
(a) Bar One
(c) All Bars
Cumulative percentage of major
causes of defective entries
Transposed
numbers
Wrong
character
Torn
document
Creased
document
Illegible source
document
Out of field
Data too light
0
15
30
45
60
56
92.7
98.5
100
88.3
82.4
41.2
68
Number of defective entries due to major causes
Cumulative percentage of major
causes of defective entries
Transposed
numbers
Wrong
character
Torn
document
Creased
document
Illegible source
document
Out of field
Data too light
0
15
30
45
60
56
92.7
98.5
100
88.3
82.4
41.2
68
Figure 6.10 Pareto diagram: Cumulative
percentage line plotting
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 187
The primary purpose of this discussion is to provide help to individuals and teams who
are trying to answer the question: What change can we make (in the Xs) that will result in
improvement of our problematic process (the CTQ)? Change concepts can serve to provoke
a new idea for an individual or team. A team leader can choose one or more of the 70 change
concepts and then the team can explore some ideas for possible application of this concept
to a step in the problematic process (X). The list of ideas should be recorded. After the generation of ideas is complete, the ideas can be discussed and critiqued. Any of the ideas that
show promise can be further explored by the team to obtain a specific idea for a change to a
step in the process, or an X.
Some of the change concepts appear to offer conflicting advice for developing changes. For
example, concept 25 “reduce choice of features” and concept 63 “mass customize” appear
to be aimed in opposite directions. Change concept 4 “reduce controls on the system” and
change concept 51 “standardization (create a formal process)” also suggest conflicting directions. The important consideration is the context in which the change concept is being
applied. The change concepts are listed in Langley J., Nolan K., Nolan T., Norman C., and
Provost L. (1996),
The Improvement Guide, Jossey-Bass, Inc. (San Francisco, CA), or Gitlow,
H., Oppenheim, A., Oppenheim, R., and Levine, D. (2015),
Quality Management, 4th ed.,
Hercher Publishing Company (Naperville, IL).
Eliminate Waste
In a broad sense, any activity or resource in an organization that does not add value to an
external customer can be considered waste. This section provides some concepts for eliminating waste.
1. Eliminate things that are not used. Constant change in organizations results in less
demand for specific resources and activities that were once important to the business. Unnecessary activities and unused resources can be identified through surveys,
audits, data collection, and analysis of records. The next step is to take the obvious
actions to remove the unused elements from the system.
2. Eliminate multiple entry. In some situations, information is recorded in a log or
entered into a database more than one time, creating no added value. This practice is
also called
data redundancy. Changing the process to require only one entry can lead
to improvement in productivity and quality by reducing discrepancies.
3. Reduce or eliminate overkill. Sometimes, a company’s standard or recommended
resources are designed to handle special, severe, or critical situations rather than the
normal situation. Changing the standard to the appropriate level of resources for the
normal situation reduces waste. Additional resources are used only when the situation
warrants it.
4. Reduce controls on the system. Individuals and organizations use various types
of controls to make sure a process or system does not stray too far from standards,
requirements, or accepted practices. While useful for protection of the organization,
these controls can increase costs, reduce productivity, and stifle improvement. Typical forms of controls include a layered management structure, approval signatures,
From the Library of Pearson HED
188 A Guide to Six Sigma and Process Improvement for Practitioners and Students
standardized forms, and reports. A regular review of all the organization’s control
procedures by everyone working in the system can result in identifying opportunities
to reduce controls on the system without putting the organization at risk.
5. Recycle or reuse. Once a product is created and used for its intended purpose, it is
natural to discard it and the by-products created by its use. However, if other uses can
be found for the discarded product or by-products, the cost of producing the product
can be spread out over its use and its reuse.
6. Use substitution. Waste can often be reduced by replacing some aspect of the product
or process with a better alternative. One type of substitution is to include lower-cost
components, materials, or methods that do not affect the performance of the process,
service, or product (sometimes called
value engineering). Another type of substitution
is to switch to another process with fewer steps or less manual effort.
7. Reduce classifications. Classifications are often developed to differentiate elements
of a system or to group items with common characteristics, but these classifications
can lead to system complexity that increases costs or decreases quality. Classification
should be reduced when the complexity caused by the classification is worse than the
benefit gained.
8. Remove intermediaries. Intermediaries such as distributors, handlers, agents, and
carriers may be part of a system. Consider eliminating these activities by linking
production directly with the consumer. Some intermediaries add value to a process
because of their specialized skills and knowledge. Often, however, eliminating these
services can increase productivity without reducing value to the customer.
9. Match the amount to the need. Rather than using traditional standard units or sizes,
organizations can adjust products and services to match the amount required for a
particular situation. This practice reduces waste and carryover inventory. By studying
how customers use the product, more convenient package sizes can be developed.
10. Use sampling. Reviews, checks, and measurements are made for a variety of reasons.
Can these reasons be satisfied without checking or testing everything? Many times,
the standard 100 percent inspection and testing results is a waste of resources and
time. Formal sampling procedures are available that can often provide as good or even
better information than 100 percent checking. This is discussed further in Chapter 13 ,
“DMAIC Model: ‘I’ Is for Improve.”
11. Change targets or set points. Sometimes problems go on for years because some
piece of equipment is not designed or set up properly. Make sure that process settings
are at desirable levels. Investigate places where waste is created, and consider adjustments to targets or set points to reduce the waste.
Improve Work Flow
Products and services are produced by processes. How does work flow in these processes?
What is the plan to get work through a process? Are the various steps in the process arranged
and prioritized to obtain quality outcomes at low costs? How can the work flow be changed
so that the process is less reactive and more planned?
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 189
12. Synchronize. Production of products and services usually involves multiple stages.
These stages operate at different times and at different speeds, resulting in an operation that is not smooth. Much time can be spent waiting for another stage to be
reached. By focusing on the flow of the product (or customer) through the process,
each of the stages can be synchronized.
13. Schedule into multiple processes. A system can be redesigned to include multiple
versions of the same process focused on the specific requirements of the situation.
Rather than a “one-size-fits-all” large process, multiple versions of the process are
available, each tuned to the different types of needs of customers or users. Priorities
can be established to allocate and schedule the inputs to maximize the performance
of the system. The specific processes can then be greatly simplified since they only
address a limited range of input requirements.
14. Minimize handoffs. Many systems require that elements such as a customer, a form,
or a product be transferred to multiple people, offices, or workstations to complete
the processing or service. The handoff from one stage to the next can increase time
and costs and cause quality problems. The work flow can be rearranged to minimize
any handoff in the process. The process can be redesigned so that any worker is only
involved once in an iteration of a process.
15. Move steps in the process close together. The physical location of people and facilities can affect processing time and cause communication problems. If the physical location of adjacent steps in a process are moved close together, work can be
directly passed from one step to the next. This eliminates the need for communication
systems, such as mail, and physical transports, such as vehicles, pipelines, and conveyor belts.
16. Find and remove bottlenecks. A bottleneck or constraint is anything that restricts the
throughput of a system. A constraint within an organization would be any resource
for which the demand is greater than its available capacity. To increase the throughput of a system, the constraints must be identified, exploited if possible, and removed
if necessary. Bottlenecks occur in many parts of daily life; they can usually be identified by looking at where people are waiting or where work is piling up.
17. Use automation. The flow of many processes can be improved by the intelligent
use of automation. Consider automation to improve the work flow for any process
to reduce costs, reduce cycle times, eliminate human slips, reduce repetitive manual
tasks, and provide measurement.
18. Smooth work flow. Yearly, monthly, weekly, and daily changes in demand often
cause work flow to fluctuate widely. Rather than trying to staff to handle the peak
demands, steps can often be taken to better distribute the demand. This distribution
results in a smooth work flow rather than in continual peaks and valleys.
19. Do tasks in parallel. Many systems are designed so that tasks are done in a series or
linear sequence. The second task is not begun until the first task is completed. This is
especially true when different groups in the organization are involved in the different
From the Library of Pearson HED
190 A Guide to Six Sigma and Process Improvement for Practitioners and Students
steps of a process. Sometimes, improvements in time and costs can be gained from
designing the system to do some or all tasks in parallel.
20. Consider people in the same system. People in different systems are usually working
toward different purposes, each trying to optimize their own system. Taking actions
that help people to think of themselves as part of the same system can give them a
common purpose and provide a basis for optimizing the larger system.
21. Use multiple processing units. To gain flexibility in controlling the work flow, try to
include multiple workstations, machines, processing lines, and fillers in a system. This
makes it possible to run smaller lots, serve special customers, minimize the impact of
maintenance and downtime, and add flexibility to staffing. With multiple units, the
primary product or service can be handled on one line to maximize efficiency and
minimize setup time. The less-frequent products and services can be handled by the
other units.
22. Adjust to peak demand. Sometimes it is not possible to balance the demands made
on a system. In these cases, rather than keeping a fixed amount of resources (materials, workers, and so on), historical data can be used to predict peak demands. Then
methods can be implemented to meet the temporarily increased demand.
Optimize Inventory
Inventory of any type can be a source of waste in organizations. Inventory requires capital
investment, storage space, and people to handle and keep track of it. In manufacturing organizations, inventory includes raw material waiting to be processed, in-process inventory, and
finished-goods inventory. For service organizations, the number of skilled workers available
is often the key inventory issue. Extra inventory can result in higher costs with no improvement in performance for an organization. How can the costs associated with the maintenance
of inventory be reduced? An understanding of where inventory is stored in a system is the
first step in finding opportunities for improvement.
23. Match inventory to predicted demand. Excess inventory can result in higher costs
with no improvement in performance for an organization. How can the proper
amount of inventory to be maintained at any given time be determined? One approach
to minimizing the costs associated with inventory is to use historical data to predict
the demand. Using these predictions to optimize lead times and order quantities leads
to replenishing inventory in an economical manner. This is often the best approach
to optimizing inventory when the process involves lengthy production times.
24. Use pull systems. In a pull system of production, work at a particular step in the
process is done only if the next step in the process is demanding the work. Enough
product is ordered or made to replenish what was just used. This is in contrast to
most traditional
push systems, in which work is done as long as inputs are available.
A pull system is designed to match production quantities with a downstream need.
This approach can often result in lower inventories than a schedule-based production system. Pull systems are most beneficial in processes with short cycle times and
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 191
high yields. Some features of effective pull systems are small lot sizes and container
quantities, fast setup times, and minimal rework and scrap.
25. Reduce choice of features. Many features are added to products and services to
accommodate the desires and wants of different customers and different markets.
Each of these features makes sense in the context of a particular customer at a particular time, but taken as a whole, they can have tremendous impact on inventory costs. A
review of current demand for each feature and consideration of grouping the features
can allow a reduction in inventory without loss of customer satisfaction.
26. Reduce multiple brands of same items. If an organization uses more than one brand
of any particular item, inventory costs will usually be higher than necessary since a
backup supply of each brand must be kept. Consider ways to reduce the number of
brands while still providing the required service.
Change the Work Environment
Changes to the environment in which people work, study, and live can often provide leverage
for improvements in performance. As organizations try to improve quality, reduce costs, or
increase the value of their products and services, technical changes are developed, tested, and
implemented. Many of these technical changes do not lead to improvement because the work
environment is not ready to accept or support the changes. Changing the work environment
itself can be a high-leverage opportunity for making other changes more effective.
27. Give people access to information. Traditionally, organizations have carefully controlled the information available to various groups of employees. Making relevant
information available to employees allows them to suggest changes, make good decisions, and take actions that lead to improvements.
28. Use proper measurements. Measurement plays an important role in focusing people
on particular aspects of a business. Developing appropriate measures, making better
use of existing measures, and improving measurement systems can lead to improvement throughout the organization.
29. Take care of basics. Certain fundamentals must be considered to make any organization successful. Concepts like orderliness, cleanliness, discipline, and managing costs
and prices are examples of such fundamentals. It is sometimes useful to take a fresh
look at these basics to see whether the organization is still on track. If there are fundamental problems in the business, changes in other areas may not lead to improvements. Also, when people’s basic needs are not being met, meaningful improvements
cannot be expected in other areas. The Five-S movement, which was the beginning
of quality control in Japanese workshops, got its name from the Japanese words for
straighten up, put things in order, clean up, cleanliness, and discipline.
30. Reduce demotivating aspects of the pay system. Pay is rarely a positive motivator in an organization, but it can cause confusion and become a demotivator. Some
pay systems can encourage competition rather than cooperation among employees.
Another result of some pay systems is the reluctance to take risks or make changes.
From the Library of Pearson HED
192 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Review the organization’s system for pay to ensure that the current system does not
cause problems in the organization.
31. Conduct training. Training is basic to quality performance and the ability to make
changes for improvement. Many changes are not effective if people have not received
the basic training required to do a job. Training should include the “why” as well as
the “what” and the “how.”
32. Implement cross-training. Cross-training means training people in an organization
to do multiple jobs. Such training allows for flexibility and makes change easier. The
investment required for the extra training pays off in productivity, product quality,
and cycle times.
33. Invest more resources in improvement. In some organizations, people spend more
than a full-time job getting their required tasks completed and fighting the fires created in their work. The only changes made are reactions to problems or changes
mandated outside the organization. To break out of this mode, management must
learn how to start investing time in developing, testing, and implementing changes
that lead to improvements.
34. Focus on core processes and purpose. Core processes are the processes directly
related to the purpose of the organization. They can be characterized as those activities that provide value directly to external customers. To reduce costs, consider reducing or eliminating activities that are not part of the core processes.
35. Share risks. Every business is faced with taking risks and reaping their accompanying
potential rewards or losses. Many people become more interested in the performance
of their organization when they can clearly see how their future is tied to the longterm performance of the organization. Developing systems that allow all employees
to share in the risks can lead to an increased interest in performance. Types of plans
for sharing risks and gains include profit sharing, gain sharing, bonuses, and pay for
knowledge.
36. Emphasize natural and logical consequences. An alternative approach to traditional
reward-and-punishment systems in organizations is to focus on natural and logical
consequences. Natural consequences follow from the natural order of the physical
world (for example, not eating leads to hunger), while logical consequences follow
from the reality of the business or social world (for example, if you are late for a
meeting, you will not have a chance to have input on some of the issues discussed).
The idea of emphasizing natural and logical consequences is to get everyone to be
responsible for their own behavior rather than to use power, judge others, and force
submission. Rather than demanding conformance, the use of natural and logical
consequences permits choice.
37. Develop alliances/cooperative relationships. Cooperative alliances optimize the
interactions between the parts of a system and offer a better approach for integration
of organizations.
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 193
Enhance the Producer/Customer Relationship
To benefit from improvements in quality of products and services, the customer must recognize and appreciate the improvements. Many ideas for improvement can come directly from
a supplier or from the producer’s customers. Many problems in organizations occur because
the producer does not understand the customer’s needs, or because customers are not clear
about their expectations of suppliers. The interface between the producer/provider and the
customer provides opportunities to learn and develop changes that lead to improvement.
38. Listen to customers. It is easy for people to get caught up in the internal functioning of the organization and forget why they are in business: to serve their customers.
Time should be invested on a regular basis in processes that “listen” to the customers.
Sometimes it is important to figure out how to communicate with customers farther
down the supply chain, or even with the final consumer of the product or service.
Talk to customers about their experiences in using the organization’s products. Learn
about improvement opportunities.
39. Coach customers to use the product/service. Customers often encounter quality
problems and actually increase their costs because they do not understand all the
intricacies of the product or service. Companies can increase the value of their products and services by developing ways to coach customers on how to use them.
40. Focus on the outcome to a customer. Make the outcome (the product or service)
generated by your organization the focus of all activities. First, clearly understand the
outcomes that customers expect from your organization. Then, to focus improvement
efforts on a particular work activity, consider the question, How does this activity support the outcome to the customer? Make improvements in such areas as the quality,
cost, efficiency, and cycle time of that activity. Organize people, departments, and
processes in a way that best serves the customer, paying particular attention to the
product/customer interfaces. This change concept can also be described as “begin
with the end in mind.”
41. Use a coordinator. A coordinator’s primary job is to manage producer/customer
linkages. For example, an expeditor is someone who focuses on ensuring adequate
supplies of materials and equipment or who coordinates the flow of materials in an
organization. Having someone coordinate the flow of materials, tools, parts, and
processed goods for critical processes can help prevent problems and downtime. A
coordinator can also be used to work with customers to provide extra services. One
example is a case manager, who acts as a buffer between a complex process and the
customer.
42. Reach agreement on expectations. Many times customer dissatisfaction occurs
because the customers feel that they have not received the products or services they
were led to expect as a result of advertising, special promotions, and promises by the
sales group. Marketing processes should be coordinated with production capabilities.
Clear expectations should be established before the product is produced or the service
is delivered to the customer.
From the Library of Pearson HED
194 A Guide to Six Sigma and Process Improvement for Practitioners and Students
43. Outsource for “free.” Sometimes it is possible to get suppliers to perform additional
functions for the customer with little or no increase in the price to the customer. A
task that is a major inconvenience or cost for the customer can be performed inexpensively and efficiently by the supplier. The supplier might be willing to do this task
for “free” to secure ongoing business with the customer.
44. Optimize level of inspection. What level of inspection is appropriate for a process? All products eventually undergo some type of inspection, possibly by the user.
Options for inspection at any given point in the supply chain are no inspection, 100
percent inspection, or reduction or increases to the current level of inspection. A
study of the level of inspection can potentially lead to changes that increase quality of
outcomes to the customers and/or decrease costs. Identifying the appropriate level of
inspection for a process is discussed in Chapter 13 .
45. Work with suppliers. Inputs to a process sometimes control the costs and quality of
performance of a process. Working with suppliers to use their technical knowledge
can often reduce the cost of using their products or services. Suppliers may even have
ideas on how to make changes in a company’s process that will surprise its customers.
Manage Time
Cut cycle time as a strategy for improving any organization. An organization can gain a competitive advantage by reducing the time to develop new products, waiting times for services,
lead times for orders and deliveries, and cycle times for all functions and processes.
46. Reduce setup or startup time. Setup times can often be cut in half just by getting
organized for the setup. Minimizing setup or startup time allows the organization to
maintain lower levels of inventory and get more productivity out of its assets.
47. Set up timing to use discounts. The planning and timing of many activities can be
coordinated to take advantage of savings and discounts that are available, resulting in
a reduction of operating costs. An organization must have a system in place to take
advantage of such opportunities. For example, available discounts on invoices offered
by suppliers for paying bills within ten days of the invoice date require a system that
can process an invoice and cut a check within the discount period. Opportunities to
apply this concept require a flexible process and knowledge of the opportunity to take
advantage of the timing.
48. Optimize maintenance. Time is lost and quality often deteriorates when production
and service equipment break down. A preventive maintenance strategy attempts to
keep people and machines in good condition instead of waiting until there is a breakdown. Through proper design and the study of historical data, an efficient maintenance program can be designed to keep equipment in production with a minimum
of downtime for maintenance. Learning to observe and listen to equipment before it
breaks down is also an important component of any plan to optimize maintenance.
49. Extend specialists’ time. Organizations employ specialists who have specific skills or
knowledge, but not all of their work duties utilize these skills or knowledge. Try to
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 195
remove assignments and job requirements that do not use the specialists’ skills. Find
ways to let specialists have a broader impact on the organization, especially when the
specialist is a constraint to throughput in the organization.
50. Reduce wait time. Reduction in wait time can lead to improvements in many types
of services. Ideas for change that can reduce the time that customers have to wait are
especially useful. This refers not only to the time to perform a service for the customer, but the time it takes the customer to use or maintain a product.
Manage Variation
Many quality and cost problems in a process or product are due to variation. Reduction of
variation improves the predictability of outcomes and may actually exceed customer expectations and help to reduce the frequency of poor results. Many procedures and activities are
designed to deal with variation in systems. Consideration of Shewhart’s concept of common
and special causes opens up opportunities to improve these procedures. By focusing on variation, some ideas for changes can be developed. Three basic approaches can be taken: Reduce
the variation, compensate for the variation, or exploit the variation.
51. Standardization (create a formal process). The use of standards, or standardization,
has a negative and bureaucratic connotation to many people. However, an appropriate amount of standardization can provide a foundation upon which improvement
in quality and costs can be built. Standardization is one of the primary methods for
reducing variation in a system. The use of standardization, or creating a more formal
process, should be considered for the parts of a system that have big effects on the
outcomes, or the leverage points.
52. Stop tampering. Tampering is defined as interfering so as to weaken or change for the
worse. In many situations, changes are made on the basis of the last result observed or
measured. Often these changes actually increase the variation in a process or product,
as illustrated by the Funnel Experiment, discussed in Chapter 1 , “You Don’t Have to
Suffer from the Sunday Night Blues!” The methods of statistical process control can
be used to decide when it is appropriate to make changes based on recent results.
53. Develop operational definitions. Reduction of variation can begin with a common
understanding of concepts commonly used in the transaction of business. The meaning of a concept is ultimately found in how that concept is applied. Simple concepts
such as on time, clean, noisy, and secure, need operational definitions to reduce variation in communications and measurement.
54. Improve predictions. Plans, forecasts, and budgets are based on predictions. In many
situations, predictions are built from the ground up each time a prediction is required,
and historical data is not used. The study of variation from past predictions can lead
to alternative ways to improve the predictions.
55. Develop contingency plans. Variation in everyday life often creates problems. Reducing the variation may eventually eliminate the problems, but how do people cope in
the meantime? One way is to prepare backup plans, or contingencies, to deal with the
From the Library of Pearson HED
196 A Guide to Six Sigma and Process Improvement for Practitioners and Students
unexpected problems. When the variation is due to a special cause that can be identified, contingency plans can be ready when these special causes of variation occur.
56. Sort product into grades. Creative ways can be developed to take advantage of naturally occurring variation in products. Ways of sorting the product or service into different grades can be designed to minimize the variation within a grade and maximize
the variation between grades. The different grades can then be marketed to different
customer needs.
57. Desensitize. It is impossible to control some types of variation: between students
in a class, among the ways customers try to use a product, in the physical condition
of patients who enter the hospital. How can the impact on the outcome (education,
function, and health) be minimized when this variation is present? It can be done by
desensitizing or causing a nonreaction to some stimulus. This change concept focuses
on desensitizing the effect of variation rather than reducing the incidence of variation.
58. Exploit variation. It is sometimes not clear how variation can be reduced or eliminated. Rather than just accepting or “dealing with” the variation, ways can be developed to exploit it. This change concept deals with some ways to turn the negative
variation into a positive method to differentiate products or services.
Design Systems to Avoid Mistakes
Making mistakes is part of being human; they occur because of the interaction of people with
a system. Some systems are more prone to mistakes than others. Mistakes can be reduced
by redesigning the system to make their occurrence less likely. This type of system design
or redesign is called
mistake proofing or error proofing. Mistake proofing can be achieved by
using technology, such as adding equipment to automate repetitive tasks, by using methods
to make it more difficult to do something wrong, or by integrating these methods with technology. Methods for mistake proofing are not directed at changing people’s behavior, but
rather at changing the system to prevent slips. They aim to reduce mistakes from actions that
are done almost subconsciously when performing a process or using a product.
59. Use reminders. Many mistakes are made as a result of forgetting to do something.
Reminders can come in many different forms: a written notice or a phone call, a
checklist of things to accomplish, an alarm on a clock, a standard form, or the documented steps to follow for a process.
60. Use differentiation. Mistakes can occur when people are dealing with things that look
nearly the same. A person may copy a wrong number or grab a wrong part because of
similarity or close proximity to other numbers or parts. Mistakes can also occur when
actions are similar. A person may end up in the wrong place or use a piece of equipment in the wrong way because the directions or procedures are similar to others
they might have used in a different situation. Familiarity that results from experience
can sometimes increase the chance of committing mistakes of association. To reduce
mistakes, steps should be taken to break patterns. This can be done by, for example,
color coding, sizing, using different symbols, or separating similar items.
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 197
61. Use constraints. A constraint restricts the performance of certain actions. A door
that blocks passage into an unsafe area is a constraint. Constraints are an important
method for mistake proofing because they can limit the actions that result in mistakes.
They do not just make information available in the external world, they also make
it available within the product or system itself. To be effective, constraints should be
visible and easy to understand. Constraints can be built into a process so that accidental stopping or an unwanted action that will result in a mistake can be prevented.
Constraints can also be used to make sure that the steps performed in a process or
when using a product are accomplished in the correct sequence.
62. Use affordances. An affordance provides insight, without the need for explanation,
into how something should be used. In contrast to a constraint, which limits the
actions possible, an affordance provides visual (or other sensory) prompting for the
actions that should be performed. Once a person sees the fixtures on a door, he should
be able to determine whether it opens in, opens out, or slides. There should not be a
need to refer to labels or to use a trial-and-error approach. If a process or product can
be designed to lead the user to perform the correct actions, fewer mistakes will occur.
Focus on the Product or Service
Most of the change concepts in the other categories address the way that a process is performed; however, many of the concepts also apply to improvements to a product or service.
This category comprises eight change concepts that are particularly useful for developing
changes to a product or service that does not naturally fit into any of the other groupings.
63. Mass customize. Most consumers of products and services would agree that quality
increases as the product or service is customized to the customer’s unique circumstances. Most consumers would also expect to pay more or wait longer for these
customized offerings than for a mass-produced version. To mass customize means
combining the uniqueness of customized products with the efficiency of mass production.
64. Offer the product or service anytime. Many products and services are available only
at certain times. Such constraints almost always detract from their quality. How can
these constraints be removed? In some cases a technology breakthrough, such as the
ATM, is needed. In other cases, prediction plays an important role—for example,
predicting what type of cars customers will order. However, in many situations the
constraint is created because it is more convenient for the provider of the service than
for the customer. Offering the product or service anytime is different from just reducing wait time. To achieve this goal often takes a totally new conceptualization of the
product or service. For this reason, “anytime” is an important concept for expanding
the expectations of customers.
65. Offer the product or service anyplace. An important dimension of quality for most
products and services is convenience. To make a product or service more convenient,
free it from the constraints of space. Make it available anyplace. For products, the constraint of space is often related to the size of the product. Making a product smaller or
From the Library of Pearson HED
198 A Guide to Six Sigma and Process Improvement for Practitioners and Students
lighter without adversely affecting any of its other attributes almost always improves
the quality of the product. One of the most striking examples is the miniaturization of
the computer to the point that it can now be carried in a briefcase and used virtually
anywhere.
66. Emphasize intangibles. Opportunities for improvement can be found by embellishing a product or service with intangible features. Three ways to accomplish this are
by miniaturizing, providing information (electronically or otherwise), and developing
producer-customer relationships.
67. Influence or take advantage of fashion trends. The features and uniformity of a
product or service define its quality. Uniformity is often assumed to exist in a product
or service, while features can affect customer expectations. Features are frequently
subject to fashion trends.
68. Reduce the number of components. Reducing handoffs was one of the change concepts for simplifying a process. Similarly, reducing the number of component parts is
a way to simplify a product. Components in this context can mean component parts,
ingredients, or multiple types of the same component. Reduction in the number of
components can be achieved through design of the product so that one component
performs the functions previously performed by more than one; or by standardizing
the size, shape, or brand of similar components; or by packaging components into
modules.
69. Disguise defects and problems. In some instances, especially in the short term, it may
be more effective to hide the defect in a product or service than to remove it. However,
the longer-term strategy is to remove the defect. Included in this category are actions
taken to make the defect more palatable to the customer. This change concept does
not include false advertising, in which claims about the product are made that are not
true, nor does it include defects that are hidden at the time of sale only to emerge in
later use of the product.
70. Differentiate product using quality dimensions. Customer satisfaction is improved
as the match between process output and customer needs/wants is increased. The
degree of matching is determined using customer research. Customer research can
provide an understanding of customer’s needs and wants.
How to Do Communication Plans
Communication plans are created by thinking about what human factors need to be addressed
in your organization as relates to a process improvement project or initiative. Some possible
questions to consider are
Does everyone in the organization understand process improvement in general?
Who do we need to communicate to?
Who is committed (not just supportive) to the project from top management?
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 199
What does the project mean to top management?
What does the project mean for individual employees?
What does the project do for the organization?
What is expected from various stakeholders with respect to the project?
What are the potential benefits and risks of the project; do we need to communicate
them?
What is the timeline for the project?
What resources will be required for the project; especially human resources?
How will people be affected by the process improvement project, and how can we
alleviate their potential concerns?
Once you have answered these questions you need to plan different events/communications
using channels most appropriate for the selected audience by filling in the columns on the
communication plan presented in Table 6.19 .
Event/communication—What do you need to communicate?
Participants—Who does the event/communication need to reach? Are there opinion
leaders (those who have an active voice in a community and whose advice is sought
by others) that we want to be involved to help facilitate project buy-in?
Medium—How are you going to reach them? Possible ways are meetings, website,
email, posters, information sessions, conference calls, private discussions, and so on.
Frequency—How often will this event/communication take place?
When—What is the specific date(s) and time it will happen?
Lead—Who is taking the lead on the communication?
Scheduled—Who is doing the scheduling and are they doing it on time?
Status—TBD (to be done), Done (completed), WIP (work in progress)
Notes—Are there any miscellaneous notes?
Table 6.19 Communication Plan Outline

#
1
Event/
Communication
Participants/
Audience
Medium Frequency When Lead Scheduled? Status Notes
2
3
4

From the Library of Pearson HED
200 A Guide to Six Sigma and Process Improvement for Practitioners and Students

#
5
Event/
Communication
Participants/
Audience
Medium Frequency When Lead Scheduled? Status Notes
6

See Table 6.7 for an example of a communication plan for reducing no shows in a clinic
project.
Takeaways from This Chapter
A flowchart is a tool used to define and document a process.
Voice of the Customer analysis involves surveying stakeholders of a process to understand their requirements and needs.
A SIPOC analysis is a simple tool for identifying the Suppliers and their Inputs into a
process, the high level steps of a Process, the Outputs of the process, and the Customer
segments interested in the outputs.
An operational definition promotes understanding between people by putting communicable meaning into words.
Measurement systems analysis studies are used to calculate the capability of a measurement system to determine whether the data can be used for meaningful analysis
of a problematic process.
Failure Modes and Effects Analysis (FMEA) is a tool used to identify, estimate, prioritize, and reduce risk among the Xs for a project.
Check sheets are used for collecting or gathering data in a logical format, called rational subgrouping. The data collected can be used to construct a control chart, a Pareto
diagram, or a histogram.
Brainstorming is a process used to elicit a large number of ideas from a group of
people in a brief amount of time. Team members use their collaborative thinking
power to generate unlimited ideas and thoughts.
An affinity diagram is a tool used by teams to organize and consolidate a substantial
and unorganized amount of verbal, pictorial, and/or audio data relating to a problem.
A cause and effect (C&E) diagram, also known as a fishbone diagram as its structure
resembles the spine and head of a fish, is a tool that is used to organize the potential
causes of a problem, select the most probable cause, and verify the cause and effect
relationship between the problem and the most probable cause.
Pareto diagrams are used to identify and prioritize issues that contribute to a problem
we want to solve. Pareto analysis focuses on distinguishing the “vital few versus the
trivial many.”
From the Library of Pearson HED
Chapter 6 Non-Quantitative Techniques: Tools and Methods 201
A Gantt chart is a simple scheduling tool. It is a bar chart that plots tasks and subtasks
against time.
Change concepts are approaches to change that have been found to be useful in developing solutions that lead to improvement in processes.
A communication plan is created for a project to identify and appease human concerns regarding the project as many people will typically be affected by it.
References
Gitlow, H., A. Oppenheim, R. Oppenheim, and D. Levine (2015), Quality Management,
4th ed. (Naperville, IL: Hercher Publishing Company). This book is free online at
hercherpublishing.com.
Gitlow, H. and D. Levine (2004),
Six Sigma for Green Belts and Champions: Foundations, DMAIC, Tools and Methods, Cases and Certification (Upper Saddle River, NJ:
Prentice-Hall).
Langley J., K. Nolan, T. Nolan, C. Norman, and L. Provost (1996),
The Improvement Guide
(San Francisco: Jossey-Bass).
Additional Readings
Pyzdek, Thomas (2003), Quality Engineering Handbook, 2nd ed. (New York: Marcel Dekker).
Rath and Strong (2000),
Six Sigma Pocket Guide (Lexington, MA: Rath and Strong).SeverityO
ccurrenceRPNSeverityOccurrence
From the Library of Pearson HED
This page intentionally left blank
From the Library of Pearson HED
203
7
Overview of Process
Improvement Methodologies
What Is the Objective of This Chapter?
The objective of this chapter is to introduce the different methods available to solve process
improvement problems. We give you a high level overview of the methods and some simple
examples. Later in the book we go into more detail about the methods, as well as when to use
them and how to use them.
SDSA Cycle
The SDSA cycle is a series of four steps used to standardize a process and reduce variation
in process outcomes (Gitlow et al., 2015; Gitlow and Levine, 2004). After forming a project
team, determining a process to standardize, and agreeing on a way to measure success, the
following steps are executed:
Standardize—Each employee flowcharts the process under study, usually their collective job. They are then brought together and agree upon a best practice flowchart
that takes the best aspect of their individual flowcharts and eliminates the worst
aspects of their individual flowcharts. Basically they move from their individual current state flowcharts to a new and hopefully improved standardized future state flowchart that all employees using the flowchart agree upon and follow. Standardization
can dramatically reduce variability in the outcomes of a process. It is also a great test
of a workforce’s ability to exhibit discipline in the way they do their jobs. Personal
discipline to follow the standardized flowchart, or an improved flowchart, is
critical
to a successful Six Sigma organizational transformation.
Do—Each employee uses the new, standardized flowchart, and they collectively
gather data of their efforts so the results can be analyzed in the Study step.
Study—Employees analyze the results of the experiment or test on the new standardized flowchart to see whether it has the intended outcome.
From the Library of Pearson HED
204 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Act—Employees decide whether to adopt the new standardized flowchart, and if the
answer is yes, they lock it in with documentation and training, or modify the plan and
repeat the SDSA cycle. If the answer is no, the team goes back to the beginning and
tries again.
SDSA Example
A medical records department in a hospital must file 80% of all patient records within 30 days
of the patient leaving the hospital; it is the law. Currently, the percentage of medical records
filed within 30 days has an average of 22% and a standard deviation of 2%.
Standardize—First, the three employees who work in the medical records department flowchart the medical records process as they see it. Second, they compare
their flowcharts; see the flowcharts for employees A, B, and C in Figure 7.1 . Third,
they develop one best practice flowchart that includes all their individual flowcharts’
strengths and avoids all of their individual flowcharts’ weaknesses; see the flowchart
called Standardized Flowchart in Figure 7.1 .
Employee A Employee B Employee C Standardized
Flowchart
Figure 7.1 Employee flowcharts and standardized flowchart for medical records department
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 205
Do—All employees involved in the process perform a pilot test using the new best
practice flowchart, called the Standardized Flowchart in Figure 7.1 .
Study—The employees involved in the process compare the percentage of medical
records filed within 30 days after the pilot test with the data from before the pilot test.
We can see from the Current panel and the Standardize panel in the chart in Figure
7.2 that it has improved from 22% to about 45% through standardization!
Figure 7.2 Current and standardized panel showing improvement
Act—Finally, employees make the combined flowchart the best practice method that
they all follow through documentation and training.
Anecdote about the SDSA Cycle

One of the authors always forgot which teeth he had flossed and consequently had to
start over. It was very frustrating. So, he developed a standardized best practice method.
He has four bridges in in his mouth all in the same place: the last two back teeth on the
top and bottom on both sides of his mouth. So, he standardized how he flossed.
Standardize: First the upper left bridge, next the lower left bridge, then the lower right
bridge, then the upper right bridge, then the upper teeth from left to right, and finally,
the lower teeth from left to right.
Do: He followed the standardized method for one month.
Study: He studied the results and realized he had never gotten mixed up!
Act: He locked in the standardized method and now uses it every day.

From the Library of Pearson HED
206 A Guide to Six Sigma and Process Improvement for Practitioners and Students
PDSA Cycle
The PDSA cycle, also called the Deming cycle or the Shewhart cycle, consists of four steps
used to continually improve a process by reducing variation, centering the mean on the
desired nominal level, or eliminating waste from the process (Gitlow et al., 2015; Gitlow and
Levine, 2004; Deming, 1994). After forming a project team, determining an aim for the team,
and agreeing on a way to measure success (a metric), the following steps are executed:
Plan—Team members figure out what changes can be made to the current best
practice process (see standardized flowchart on the left of the Figure 7.3 ) to achieve
improved results from the revised best practice method for the metric of interest
(see revised practice flowchart on the right of Figure 7.3 )—for example, deleting an
unnecessary step in the process as shown in Figure 7.3 . Basically, team members are
moving from the current state flowchart to a new and hopefully improved future state
flowchart by changing the flow of the process. In this book we discuss many methods
for creating the revised best practice method. These methods are ideas on how to
improve a process that are noted in the revised best practice flowchart.
Standardized
flowchart
Revised best
practice method
Improved best
practice flowchart
Figure 7.3 Standardized and revised best practice flowcharts
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 207
Do—Next, team members conduct an experiment or test of the revised flowchart
(Plan) on a small scale and collect data so the results can be analyzed in the Study step.
Study—Team members, perhaps with the assistance of an expert, analyze the results
of the experiment or determine whether it has the intended outcome. Note: Originally
this step was known as “Check,” but Dr. Deming changed it to “Study” as he felt check
meant to stop someone in ice hockey.
Act—Team members decide whether to (1) adopt the revised best practice flowchart
and lock it in with documentation and training, (2) abandon the revised best practice
method because it failed to yield the desired results, or (3) modify the revised best
practice method and repeat the PDSA cycle.
These four steps are repeated over and over again in the quest for never ending improvement.
PDSA Example
Using the SDSA cycle we were able to improve the percentage of medical records filed within
30 days from 22% to about 45% by standardizing the process among employees, which is
good, but not enough to get to the 80% required by law. The next course of action is to turn
the PDSA cycle:
Plan—After various conversations with process stakeholders, team members find
that records require three signatures from the lab before being released from the lab
for filing by the medical records department. Upon further study, it is determined that
two of those signatures are unnecessary and just increase the cycle time it takes to file
the medical records.
Do—The lab agrees to eliminate the two signatures. Then team members run a pilot
test on the new process without the two signatures from the lab.
Study—Next, team members compare the percentage of medical records filed within
30 days after the new pilot test (panel 3 in Figure 7.4 , called After Lab) with before the
pilot test (panel 2 in Figure 7.4 , called Standardize). We can see from Figure 7.4 that it
has improved from 45% to about 65% through the elimination of the two signatures
in the lab!
Act—Finally, team members lock in and sustain the new process through documentation and training. The PDSA cycle continues forever, although perhaps with less data
collection, to constantly improve the process absent capital investment. Once capital
investment is required, additional financial considerations come into play. These
financial considerations are not discussed in this book. However, many problems that
are attempted to be fixed through capital investment can frequently be fixed using the
PDSA cycle without any capital investment.
From the Library of Pearson HED
208 A Guide to Six Sigma and Process Improvement for Practitioners and Students
NOTE
The team was about to run the PDSA cycle again to move from 65% to 80% of medical records filed within 30 days when the CEO who was also the Project Champion
was fired, c’est la vie!
Anecdote about the PDSA Cycle

I was out to a Tai dinner with a friend who was 70 years old. He always ordered Spring
Break Duck as his entrée. We have been going to this restaurant at least twice a week
for years. One day I noticed that he always leaves the onions over on his plate. I said
to myself, “Is it possible he doesn’t know he can order the duck without onions?” It
seemed to stretch the boundaries of credibility that he could be 70 and be so unaware
of something as common as this. So I asked him, why don’t you order the duck without
onions. And my 70-year-old friend said to me: “Can I do that?” I was amazed that in 70
years he hadn’t realized that he could change items on a restaurant menu. So this was
an opportunity for a turn of the PDSA cycle.
Plan—The next time he eats Spring Break Duck, he will order it without onions
and with extra pineapple.
Do—The next time he orders Spring Break Duck without onions and with extra
pineapple.
Study—The waiter and the restaurant owner didn’t kick him out of the restaurant
for being obnoxious and overly demanding.

Figure 7.4 After Lab
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 209
Act—He realized that he could order his food the ways he likes it, and that the
restaurant personnel are happy to oblige him, because then he will be more likely
to return to the restaurant.
This story is amazing and true. PDSA can really improve our lives!
Kaizen/Rapid Improvement Events
Kaizen is Japanese for “improvement” or “change for the best” (Wikipedia: Kaizen); it refers
to a philosophy or practices that focus upon continuous improvement of processes. Kaizen
refers to activities that continually improve all functions and involves all employees. Kaizen
aims to eliminate waste by improving standardized processes.
Kaizen is a daily process, the purpose of which goes beyond simple productivity improvement. It is also a process that, when done correctly, humanizes the workplace, eliminates
overly hard work (called “
muri” in Japanese), and teaches people how to perform experiments on their work using the PDSA cycle and how to learn to spot and eliminate waste in
processes. In all, kaizen provides a humanized approach for workers to increase productivity;
in other words, kaizen nurtures employees and promotes participation in process improvement activities.
Kaizen works as follows: The PDSA cycle is turned, and at the Do stage employees make small
and rapid modifications to the Plan stage and determine their success in the Study phase of
the PDSA cycle. The small and rapid changes to the Plan find any problems with the Plan,
display the problems, clear the problems that stand in the way of the Plan, and acknowledge
the correctness of the modification to the Plan. Next, employees Study the results of the effort
by viewing the critical metric (CTQ or Y), and finally, Act. Kaizen or Rapid Improvement
Events are typically conducted on processes where the root cause is known but the solution
to eliminate the root cause is not.
Kaizen or Rapid Improvement Events generally consist of three stages that correspond to
the PDSA cycle. An example of a typical Kaizen or Rapid Improvement Event (RIE) looks
like this:
Pre-event (Plan)
Identify the area to be kaizened.
Identify team members.
Plan the event (secure meeting space, reserve equipment, obtain supplies, order meals,
prepare training, create certificates, and so on).
Conduct Voice of the Customer (VoC) analysis and perform process observations.
Collect baseline data.
Create Kaizen/Rapid Improvement Event (RIE) charter.
From the Library of Pearson HED
210 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Obtain management commitment.
Select the Project Champion.
Prepare a schedule of meetings with the Project Champion in which he can hear the
results of the kaizen blitz.
Event (Plan, Do, and Study)
Three to five day events consisting of
Event kickoff.
Conduct basic RIE training.
Review the charter with all relevant personnel.
Review the data.
Map out current state process.
Observe the current state process.
Do root cause analysis on current state process.
Brainstorm/prioritize/select improvements to the current state process.
Develop an implementation plan for the future state process.
Train staff on new future state process.
Create new tools for the future state process.
Implement future state process.
Prepare and deliver final report to the Process Owner and the Project Champion.
Post-event (Act)
Follow up on outstanding action items.
Celebrate success.
Track progress over time, perhaps on the organization’s dashboard.
Update the Project Champion.
Review data periodically to make sure the process has not fallen back into its old bad
habits.
Kaizen/Rapid Improvement Events Example
First case starts (FCS) in an operating room (OR) are defined as the percentage of first
cases of the day in each OR that start on time. Start time is measured by when the patient
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 211
is wheeled into the OR (wheels in). So any case that has “a wheels in time” greater than the
scheduled time is defined as late.
XYZ hospital has a problem with first case starts in its 16 ORs; only 55% of cases start on
time. Due to the high cost of OR time (reported to be in the neighborhood of $1,800/hour)
and the opportunity cost of cases not scheduled (if cases are delayed long enough it means
that other margin generating cases cannot be done), the CEO asks the process improvement
team to get involved to improve FCS.
Since many of the root causes are thought to be known (physician late, patients not cleared),
but no solution has been put in place, the team decides to conduct a Kaizen/Rapid Improvement Event.
Pre-Event
The Process Improvement Executive names a Project Leader, who forms the team, which
consists of employees from related areas including operating room staff, nursing, surgeons,
surgical services, anesthesiology, and central sterile staff. Over the next month the team
completes all the pre-event work, such as completing the project charter, planning the event,
conducing Voice of the Customer interviews, and collecting baseline data. At this point the
team is ready for the event.
Event
To limit disruption to the OR, the team schedules the event for a Friday, Saturday, and
Sunday. As an aside, each participant is given two vacation days to be used at a later time
determined by management.
Over the next three days, the team maps out the current state process, creates a fishbone diagram to identify (confirm) potential root causes, studies data and identifies the main reasons
for the delays, brainstorms potential solutions, maps out the new future state process, and
creates an implementation plan.
They find that FCS are delayed for the following three reasons:
Surgeons are late.
Patients are not cleared properly.
Missing items.
The solutions they come up with to address the preceding issues are the following:
Surgeons are late—The surgeon on the team worked with his colleagues to come up
with a plan. Since first case starts are coveted by surgeons, any surgeon who is late for
two or more first cases may lose his first case start privileges for the next month. Also,
surgeons are required to be in the hospital a minimum of 30 minutes prior to their
scheduled start.
From the Library of Pearson HED
212 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Patients not being cleared properly—Each patient is required to be scheduled for
pre-admission testing and clearance, and anesthesia is to be notified when the patient
is cleared.
Missing items—A standard checklist of items to be kept in each OR was created, and
a member of the OR staff makes sure each item is in each room the night before.
Post-Event
Since the ORs were closed on the final day of the event, the team had to wait until the next
week when the ORs opened back up to implement changes.
All staff is trained on the new process and despite the usual resistance from some employees,
the changes are implemented. Minor changes are made along the way, but the new plan
seemed to be working well.
The team collects a month’s worth of data and presents it to the CEO, who is anxious to see
how the team did. To her amazement, the team had gone from 55% of FCS on time to 89%
of FCS on time.
She thanks the team and throws a party for everyone involved to celebrate!
All new processes are locked into place with documentation, and the team creates a plan to
track the data to ensure that the improvements are sustained.
Kaizen and Rapid Improvement Events are not discussed further in this book. If you are
interested in this topic, many excellent resources are available on the Web.
DMAIC Model: Overview
Based on the scientific method, the DMAIC model is like the PDSA cycle in that it is used to
improve a current process, product, or service (Gitlow et al., 2015; Gitlow and Levine, 2004).
You can think of the DMAIC model like the PDSA cycle on steroids in that it is much more
detailed in the number of substeps within each of its five phases.
The easiest way to explain the DMAIC model is to think of it in terms of the CTQ (Y) is a
function of one or more Xs. For example, suppose that it is believed that patient waiting time
is affected by insurance type, physician, and availability of an examination room. This could
be stated as follows:
Patient waiting time is a function of insurance type, physician, and availability of an
examination room.
If we call Patient waiting time Y, insurance type (X
1), physician (X2), and availability
of an examination room (X
3), then we can write the relationship between the Y and
the Xs as follows:
Y = f(X
1, X2, X3) .
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 213
where:
Y is a measure of patient waiting time.
f is the symbol of relationship.
X
1 is insurance type.
X
2 is physician.
X
3 is availability of an examination room.
The five phases of the DMAIC model are
Define, Measure, Analyze, Improve, and Control.
They are aimed at finding revised methods that modify how the Xs are performed with the
intention of moving the CTQ (Y) in a specified direction, reducing variation in the output
(Y), or both.
Define Phase
The Define phase involves four steps; they are
1. Prepare a project charter. This is the rationale for doing the project.
2. Understand the relationships between Suppliers-Inputs-System-OutputsCustomers (called SIPOC analysis). SIPOC analysis allows team members to get an
initial understanding of all the stakeholders of the project.
3. Analyze Voice of the Customer (VoC) data to identify the critical to quality (CTQs
or Ys) important to customers. A VoC analysis provides data from each stakeholder
group directly to team members about the needs and wants of each stakeholder group
in the language of the stakeholders.
4. Develop an initial project objective. A project objective states the following: the problematic process to be studied, a metric for measuring the CTQ of the process, a direction for the CTQ (increase or decrease), a target for the metric that is rational not
arbitrary, and a deadline for the completion of the project.
Measure Phase
The Measure phase involves five steps; they are
1. Create an operational definition for each CTQ. Operational definitions give meaning to terms so that all concerned parties agree on the definition of the metric being
studied in the DMAIC model.
2. Develop a data collection plan for the CTQ.
3. Perform a measurement system that determines if there is accurate data about each
CTQ.
4. Collect and establish a baseline for each CTQ.
5. Estimate the process capability for each CTQ.
From the Library of Pearson HED
214 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Once all of this is done, you can collect data to determine the performance of the process.
In other words, you can determine the current state performance of your CTQ or Y. This
allows you to compare the current process performance, as measured by the CTQ, with the
performance of the improved process.
Analyze Phase
The Analyze phase involves the following steps:
1. Create a detailed flowchart of the current state process. This allows all concerned parties to understand exactly how the process works in its current form.
2. Identify the Xs, which are factors that cause each of your CTQs to be problematic. Xs
can be identified in many ways, such as analyzing your current state flowchart, talking
to process experts, surveying the literature, analyzing data, using the list of 70 change
concepts, and benchmarking, to name a few sources of Xs.
3. Reduce the number of potential Xs that can impact the CTQ by using a Failure Modes
and Effect Analysis (FMEA). FMEA is a tool used to identify the Xs that are critical
to reducing the variation of the CTQ, moving the CTQ toward the desired nominal
level, or eliminating waste to get better results for the CTQ.
4. Develop an operational definition for each X (see Measure phase).
5. Develop a data collection plan for the Xs (see Measure phase).
6. Create a measurement system for each of the Xs (if necessary; see Measure phase).
7. Collect data and testing theories to determine which of the Xs cause your CTQs to be
problematic.
8. Develop hypotheses/takeaways about the relationships between the critical Xs and the
CTQ(s).
Testing of theories between CTQs and Xs involves determining how and why each X causes
the CTQ to be problematic. For example, you can develop hypotheses like the following (if
you have 5 Xs):
Y = f(X
1), or Y = f(X2, X3,X5)
A test would look like this:
If cycle time for a nurse to answer a call bell in a hospital is the CTQ (or Y), and X
1 is the
nurse, then we can see from Figure 7.5 that nurse A takes longer to answer a call bell than
nurse B. In this hospital, X1 is a potentially important X for improving the CTQ.
Suppose that the cycle time to answer a call bell in a second hospital is recorded for Nurse X
and Nurse Y for 100 consecutive patients. From the distributions of call bell response times in
Figure 7.6 , we can easily see that the nurses have the same call bell response time distribution.
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 215
Consequently, we develop a hypothesis that “XNurse” is NOT a potentially important X to
explain the behavior of call bell response time (CTQ or Y) in the second hospital, unless it is
correlated to another X that is significant and interacts with X
Nurse.
Figure 7.5 Dot plot of Nurse A versus Nurse B
Figure 7.6 Dot plot of Nurse X versus Nurse Y
From the Library of Pearson HED
216 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Improve Phase
The Improve phase involves the following steps:
1. Generate potential solutions to optimize the mean, variation, and shape of the distribution of the CTQ. This can be done many ways; for example, experimental designs,
FMEA, brainstorming, using the 70 change concepts (discussed previously in this
book), simulation, or benchmarking. (
Note: We explain each of these methods when
we get into the Improve phase in more detail later in the book.)
2. Select solutions.
3. Create a flowchart for the new optimized process.
4. Identify and mitigate risk for the new process.
5. Run a pilot test of the new improved process.
6. Collect and analyze data from the pilot test.
7. Implement the process full scale across the organization.
Control Phase
The Control phase involves the following steps:
1. Reduce the effects of collateral damage to related processes.
2. Standardize improvements.
3. Develop a control plan for the Process Owner.
4. Identify and document the benefits and costs of the project.
5. Input the project into the Six Sigma database.
6. Diffuse the improvements throughout the organization.
7. Champion, Process Owner, and Black Belt sign off on the project.
DMAIC Model Example
ABC Health System has a problem with patients not showing up for their scheduled appointments in one of its behavioral health clinics. This causes idle time for the physicians and staff
and represents lost revenue for the system. The process improvement team undertakes a Six
Sigma DMAIC project to solve the problem.
Define
In the Define phase the team completes their business case including financial impact to the
department, speaks to different process stakeholders by doing a VoC analysis, and identifies their process Y (CTQ); that is, “percentage of no shows by month.” Their final project
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 217
objective is to reduce patient no shows in the behavioral health clinic from 30% per month
to as low as possible per month by January 1, 2015.
Measure
In the Measure phase the team collects data on the CTQ (Y) to get a baseline of current performance and finds that 30% of patients do not show up to their appointments on average in
the largest behavioral health clinic.
Analyze
In the Analyze phase the team uses a variety of methods to identify the Xs or the causes for
patients not showing up to their appointments and finds that there are three main factors
responsible for the patient no shows:
Age of the patient and time of appointment
No reminder call
Lack of penalty for no shows
Specifically the team finds that younger patients no show more often if their appointments
are early in the morning. They also find that patients who receive an appointment reminder
call have a higher likelihood of showing up for their appointment. Finally, during the VoC
interviews the team finds that because there is no penalty for not showing up, the patients
had a laissez faire attitude about keeping their appointments.
Improve
The team modifies the process as follows to address the critical Xs found in the Analyze
phase:
Schedule older patients earlier in the day and younger patients later in the day.
Have reminder calls made to the patients two days prior to their appointment.
Institute a $50 penalty for patients who do not cancel at least 48 hours in advance.
After conducting a pilot test with these three process changes the team saw no shows drop to
an average of 5% over the next three months.
Control
To sustain their improvements the team locks in the process changes by documenting the
newly improved process and trains all employees in the new process. They then celebrate
their success and hand the process over to the Process Owner, which includes reviewing no
show data monthly to ensure the new process is still being followed. Additionally, the Process
Owner continues to turn the PDSA cycle to drive the percentage of no shows even lower than
the 5% mark.
From the Library of Pearson HED
218 A Guide to Six Sigma and Process Improvement for Practitioners and Students
DMADV Model: Overview
Sometimes you hit a wall with a current process in terms of making further improvements.
If this happens you need to redesign the process, or sometimes you have no process at all
so you need to create one. In both of these cases, you use what is called the DMADV model
(Gitlow, et al., 2006). By using the DMADV model you are designing the process, product, or
service with customer requirements in mind at every step to ensure a high quality outcome.
The DMAIC model is about improving a process, product, or service, while the DMADV
model is about inventing and/or innovating a process, product, or service.
The
DMADV model is the Design for Six Sigma (DFSS) model used to create major new features of existing products, services, or processes, or to create entirely new products, services,
or processes. It has five phases: Define, Measure, Analyze, Design, and Verify/Validate.
Define Phase
The Define phase of the DMADV model has five components; they are
Establishing the background and business case for the project.
Assessing the risks and benefits of the project.
Forming the Six Sigma DMADV team.
Developing the project plan.
Writing the project objective.
Measure Phase
The Measure phase of a Design for Six Sigma project has three steps; they are
1. Segmenting the market, designing and conducting Voice of the Stakeholder surveys.
2. Using the survey results as inputs to find Critical to Quality characteristics (CTQs
or Ys).
3. Creating a matrix to understand the relationships between the needs and wants of
stakeholders and the features of the product, service, or process design.
Analyze Phase
The Analyze phase contains four steps; they are
1. Generate several possible designs at a high level; no details are developed in this step.
2. Compare all the possible designs and select the best one based on relevant criteria.
3. Perform a risk analysis on the best high level design.
4. Create a high level model of the best design.
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 219
The aim of these four steps in the Analyze phase is to develop preliminary designs from
which a best design can be identified based on relevant criteria.
Design Phase
The Design phase of a Design for Six Sigma project has four steps; they are
1. Construct a detailed design of the “best” design from the Analyze phase.
2. Develop and estimate the capabilities of the critical features of the best design. These
features are called Critical to Process (CTP) elements. CTPs are the components that
make up the detailed design. They include detailed drawings of each component, as
well as nominal values and specification limits.
3. Do a trial run of the new design.
4. Prepare a verification plan to enable a smooth transition among all affected departments. A verification plan ensures that the final design works well for all relevant
stakeholders.
Verify/Validate Phase
The intent of the Verify/Validate phase is to
Facilitate buy-in of Process Owners.
Design a control and transition plan so that the design can hit the ground running.
Conclude the DMADV project and celebrate the success.
Turn the new product, service, or process over to the Process Owner.
The final part of the Verify phase is to maintain communication between the Champion
and the Process Owner. These lines of communication alleviate any confusion or other
unforeseen problems that will inevitably develop. It ensures that the conceptual design is not
compromised by outside forces and neglect. All phases of the DMADV model end with a
tollgate review by the Process Owner and the Project Champion.
DMADV Model Example
ABC Hospital has seen a substantial increase in the number of transfer admissions coming
into the hospital. Transfer admissions are admissions of patients coming from other hospitals
that cannot provide the specialized services that ABC Hospital can. However, due to delays
in the process and a lack of standardization, referring hospitals have started to send their
patients elsewhere. Currently transfers are handled by nursing administration, but after looking at industry best practices, the C-suite of the hospital decides to create a Transfer Center
within the hospital. The Transfer Center would have employees whose sole job is to ensure a
standardized process to screen and admit transfers from other hospitals. To design the new
From the Library of Pearson HED
220 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Transfer Center to meet customer expectations in every way, the CEO commissioned the
process improvement team to use the DMADV model to make it happen.
Define
In the Define phase the team completes their business case including financial impact to
the department, conducts a risk assessment on the project, and forms the team. Their final
project objective is to create a Transfer Center that increases transfers to the hospital from
400 per year to as high as possible per month by January 1, 2015. At a contribution margin
of $3,000 per transfer, the financial impact is significant.
Measure
In the Measure phase the team segments the market and conducts VoC interviews with various stakeholders, including physicians, nurses, hospital staff, insurance providers, and referring hospitals. The CTQs for the team are number of transfers and number of complaints
from referring hospitals, patients, and ABC Hospital staff. The team also creates a matrix to
understand the relationships between the needs and wants of stakeholders and the features
of the product, service, or process design.
Analyze
In the Analyze phase the team uses a variety of methods to identify several low level design
options; specifically the team looks at
Various IT and data collection systems (four systems were looked at)
Number of staff (using three, four, or five employees was discussed)
Skillset of staff (having a registered nurse versus not having a registered nurse)
Location of Transfer Center within the hospital (near the Emergency Department or
near the Nursing Administration office)
Pricing packages for international patients (three options were discussed)
Hours (24/7 operation versus usual business hours)
The team then uses a matrix to select the best design, which includes
Using an add-on from the vendor who is providing the new electronic medical record
system.
Three staff is deemed the proper amount to handle the projected increased volume.
A registered nurse is critical to properly screen patients.
Being close to nursing administration is important as that is where bed assignment
occurs.
Pricing their services similarly to their closest competitors.
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 221
Providing service 24/7 as their competitors do not offer 24/7 service, and management
decided that the incremental cost is worth the differentiation from the competition
that it provides.
Design
In the Design phase, the team creates a detailed design for the best design from the Analyze
phase. An implementation plan is then created to execute on the detailed design.
Verify
The Verify phase includes meeting with the Process Owners and Project Champion to review
the control and transition plan and then celebrating the completion of the project!
Part of the Verify phase is creating and tracking the CTQs for the project and calculating the
financial impact.
After a year, the number of transfers to the hospital increased from 400 per year to 1,150, an
increase of 287%! At a contribution margin of $3,000 per patient, the financial impact was
$2.25 million! The number of complaints was reduced to next to none.
We do not cover the DMADV model in this book as it requires a book of its own. We simply
want to expose you to the methodology at a high level. For details of the DMADV model, see
Gitlow, Levine, and Popovich (2006).
Lean Thinking: Overview
Lean thinking is a process orientation that focuses on reducing waste by eliminating nonvalue added steps in a process. Lean thinking has several tools and methods that we briefly
discuss in this chapter; they are the 5S methods, Total Productive Maintenance (TPM), quick
changeovers, poka-yoke, and value stream mapping. If you are interested in learning about
lean thinking in more depth, see the references at the end of this chapter. It is important to
realize that all the lean tools and methods briefly discussed here can be used as potential
change concepts to manipulate the Xs to optimize the CTQs in the Six Sigma DMAIC model.
The 5S Methods
The first step in process improvement is frequently accomplished using the 5S methods (Gitlow, 2009). The 5S methods are simple techniques for highlighting and eliminating waste,
inconsistency, and unreasonableness from the workplace. The 5S methods can be used to
accomplish these tasks by creating a neat and tidy workplace; everything has a place and
everything is in its place.
Additionally, the 5S methods and standardization (SDSA) can be used to determine whether
the workforce in an organization has the personal discipline to maintain over time a neat and
tidy standardized workplace. Think of your junk closet at home. One day you clean it and
throw out a ton of garbage. Two months later it is back to where it was before your 5Sed the
From the Library of Pearson HED
222 A Guide to Six Sigma and Process Improvement for Practitioners and Students
closet. No discipline! Discipline is difficult to develop in a workforce; it takes years to really
hone a personal discipline culture.
Table 7.1 briefly describes each of the 5S methods, followed by some explanations.
Table 7.1 Description of the 5S Methods

5Ss Description
Sort Eliminate unnecessary things (or put them away) and make necessary things visible.
Systematize Order essential things so that they can be quickly and easily accessed and put away.
Spic and Span Clean machines, equipment, and the work environment.
Standardize Develop “best practices” to make the preceding 3Ss habits.
Self-discipline Spread and maintain the 5S culture throughout the organization. Get everyone to use
the preceding 4Ss in work every day.

1. Sort means organizing things so they are easy to find. Sort divides things that are
needed from things that are not needed, thus, creating a workplace that has only what
is needed and not a lot of junk that makes finding what is needed a time consuming
task.
2. Systematize means tidily placing “things” that are needed in their proper places so
anyone can access or put them away.
3. Spic and Span is an attitude that considers a dirty and untidy workplace intolerable; it
is proactive cleaning and maintenance. Make cleaning a cultural cornerstone of your
organization. Spic and Span is analogous to personal hygiene for people.
4. Standardize is the development of an integrated system of “best practice” methods
for sorting, systematizing, and spic and span. Standardize is used to make the first 3Ss
a habit for each individual using it.
5. Self-discipline is about spreading and maintaining the 5S culture throughout an
organization. Everyone should be involved in the 5S culture.
5S Methods Example
We demonstrate the 5S methods with an example that resonates with many of us, a messy
closet! Many people’s clothes closets at home are a bit of a disaster in that there is so much
clutter in them that they end up spending more time finding something than they spend getting dressed in the morning. The 5Ss are a perfect way to Lean out that problematic closet.
Start with
Sort. Go through your closet and pull out anything that you haven’t worn in the
past year and put it in a pile. Unless it is seasonal clothing, odds are if you haven’t worn it in
the past year it is simply taking up valuable closet space and making it hard for you to find
what you need. Donate it to someone in need or give it to a friend and let it take up space in
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 223
their closet! You might even want to remove seasonal items and put them somewhere else
until that season approaches.
Your next task is then to
Systematize and organize your closet so things are in logical places
and are easy to find. Perhaps put your shirts on one side, your dress clothes on another, your
jeans on another, and have a specific spot to hang your belts. If you really want to get sassy
you can organize them by color as well.
Then you are going to
Spic and Span or clean your closet. This can be accomplished by
throwing away old or broken hangers, dusting and cleaning the area, vacuuming the floor,
and throwing away anything that has accumulated since the last time (probably never) that
you cleaned your closet.
Next is
Standardize; this makes the preceding 3Ss a habit.
Finally determine how you will
Self-discipline your 5S efforts. This involves spreading the
5S throughout the entire organization (family).
Total Productive Maintenance (TPM)
Total Productive Maintenance (TPM) is useful for maintaining plant and equipment with
total involvement from all employees (Venkatesh, 2005; Japan Institute of Plant Maintenance, 1997). Its objectives are to dramatically increase production and employee morale by
1. Decreasing waste
2. Reducing costs
3. Decreasing batch sizes
4. Increasing production velocity
5. Increasing quality
TPM is built on a base of the 5S methods. Problems cannot be clearly seen when the workplace is unorganized. Making problems visible is the first step to eliminating them. TPM has
seven component parts. They are
jishu hozen (autonomous maintenance); kaizen; planned
maintenance; quality maintenance; training; office TPM; and safety, health, and environment
TPM.
TPM Example
An example on preventative maintenance that is near and dear to all of us is our bodies. This
is something that not only helps each of us individually but also helps us as a nation reduce
our skyrocketing healthcare costs.
Autonomous maintenance—These are things that the operator (or human being)
does without the need of any specific skill. In terms of the human body this consists
of personal hygiene, eating, getting enough sleep, reducing stress, and so on.
Planned maintenance—This refers to actions that can be taken on a routine or recurring basis. Some examples related to the human body would be annual physicals,
From the Library of Pearson HED
224 A Guide to Six Sigma and Process Improvement for Practitioners and Students
yearly skin cancer screenings, colonoscopies after the age of 50, annual mammography, and so on.
Training—These are skills that the operator (human) does not have currently that
would help prevent equipment breakdowns in the future. In terms of the human body
some examples would be stress management classes, nutrition education, learning
how to meditate, learning how to monitor blood pressure or glucose levels, and so on.
Health maintenance—Finally, equipment design changes are fundamental design
changes to the equipment (body) itself to prevent breakdowns in the future. Some
examples related to improving the human body would be to begin a weight training
program to improve strength of the body, walking daily to improve your cardiovascular system to prevent heart issues, stretching daily to improve flexibility and prevent
muscle pulls, removing tonsils to prevent recurring strep throat infections.
Quick Changeover (Single Minute Exchange of Dies—SMED)
Quick changeover (Productivity Development Team, 1996), or Single Minute Exchange of
Dies (SMED), is a technique that team members can use to analyze and then reduce
Setup time for equipment (including tools and dies) and people (for example, shift to
shift setup for cashiers in a supermarket)
Resources required for a changeover
Materials necessary for a changeover
Quick changeovers are accomplished through the SMED system. Single minute states that
the time it takes to change a process from producing product A to product B, and so on, is
less than 10 minutes. Quick changeovers create the opportunity to institute small batch sizes,
or even one-piece flows, due to the short switch over times from unit A to unit B. Ideally, one
unit batch sizes occur when switchover time from product A to product B is zero, or very
small. This allows organizations to produce to customer demand, not to produce a batch
of A, then due to a long switchover time, produce a large batch of B. This creates large and
expensive inventories.
SMED Example
Auto races are competitive in that after 500 miles cars may be separated by only 50 feet. As
all cars need to make “pit stops” to change tires and refuel, performing this type of quick
changeover can be a real competitive advantage to a team and mean the difference between
first place and tenth place.
It takes most people 15 minutes to change one tire, while a NASCAR pit crew can change a
tire in less than 15 seconds! Why? They are using SMED principles such as
Performing as many steps as possible before the pit stop begins (having all the tools
and parts ready for when the car stops, redesigning the wheel so that all nuts are
attached to the tire except one, necessitating only tightening of the nuts)
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 225
Using a coordinated team to perform multiple steps in parallel (having a team working on changing the tires while a separate team works on refueling)
Creating a highly optimized and standardized process with roles and responsibilities
to execute while the car is stopped as well as before and after it is stopped and practicing to make sure everyone can execute their responsibility
Poka-Yoke
Poka-yoke (pronounced POH-kah YOH-kay) is Japanese for “mistake-proofing device”
(Shingo, 1986). A mistake-proofing device is used to prevent the causes of defects and/or
defective output (called errors), or to inexpensively inspect each item produced to determine
whether it is conforming or defective. Poka-yoke devices promote Zero Quality Control
(ZQC). ZQC means the production or delivery of zero errors, scrap, downtime, or rework.
Poka-Yoke Examples
Sponge bags to let surgical teams count sponges to prevent sponges from being left
inside patients after surgery.
Bar coding on patient bracelets to prevent medication errors.
Rumble strips—those little carved depressions in the road that provide the driver with
tactile and audible warning that the car is not in its lane
Beeps in cars if key is left in ignition so you don’t lock your keys in your car
Ceiling paint that goes on pink but dries white to let know if you have missed a spot
Gas cap tethers, which keep you from leaving your gas cap behind after refueling your
car
An Anecdote for Poka-Yoke

There are many humorous examples of poka-yoke. One such example concerns keep
ing urinals clean in public places. This largely amounts to keeping the floors clean and
dry, which is often a problem because of the poor aim of the male users of the facilities.
One poka-yoke device is to fill the urinal with ice. We assume that most of the male
readers of this text have encountered this situation. One purpose of the ice is to bring
out the hunter gene that lies deep in most males from prehistoric times. Frequently
what happens is that the male going to the bathroom takes aim at one particular ice
cube with great concentration and attempts to completely melt it, thereby satisfying his
need for killing the cube and having a successful hunting experience. By the way, this
has been shown to reduce spillage around the urinal by over 80%.

From the Library of Pearson HED
226 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Value Streams
Value stream (Rother and Shook, 2003) is a term that describes all the value added and nonvalue added process steps and decisions necessary to move a product or service from supplier
to customer. These steps include design and redesign, raw material flows, subcomponent
flows, information flows, production and service flows, and people flows, to name a few steps.
Value stream maps can be created for different points in time. A current state value stream
map follows a product or service from supplier to customer to represent current conditions.
A future state value stream map incorporates the opportunities for improvement identified
in the current state value stream map to improve the process at a future point in time using
the previous lean tools and methods (see Figures 7.7a and 7.7b ).

COIL Corp. 6‐week
forecast
Production control
(MRP)
Daily
order
90/60/30‐day
forecasts
COOL MACHINE
Corp.
Rolls of coils 20,000 units/mo.
20‐day month
Weekly schedule
Weekly
fax

 

 

C/T = 2 sec.
C/O = 60 min.
Uptime = 85%
27,000 sec.

 

available
27,000 sec.
available
27,000 sec.
available
27,000 sec.
available
5 days 10 days*
1.5 days 1.8 days 1.65 days 3.3 days
Production
lead time =
18.25 days
Value‐added
time = 198 sec.
2 sec. 42 sec. 50 sec. 64 sec. 40 sec.

1
Package size = 20
2 shifts
(450 minutes each)
Daily ship
schedule
Monday +
Wednesday
1x day
I
Coils
5 days
C/T = 42 sec.
Weld 1

C/O = 10 min.
Uptime = 100%
2 shifts

7,500L
2,500S
C/T = 50 sec.
Weld 2

C/O = 10 min.
Uptime = 80%
2 shifts

I
1,000L
500S
C/T = 64 sec.
Assembly 1
1 1 1 1

C/O = 0
Uptime = 100%
2 shifts

I
1,200L
600S
C/T = 40 sec.
Assembly 2

C/O = 0
Uptime = 100%
27,000 sec.
available
2 shifts

I
1,100L
550S
Shipping Dept.
I
2,200L
1,100S
* 7,500 + 2,500 = 10,000
10,000 units/1,000 demand per day = 10 days
Figure 7.7a Example of current and future state value stream maps
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 227

COIL Corp. 6‐week
forecast
Production control Daily
order
90/60/30‐day
forecast
COOL MACHINE
Corp.
Daily
order

 

 

20,000 units/mo.
20 days/mo.
1,000 units/day
Package size = 20
2 shifts
(450 minutes each)
1x da y

Daily order
OXOX
20

20
Bin
Coil
Daily delivery COILS

L S Cutting and bending Welding and assembly
Change
over
3
Batch

L S

Production lead
time = 5 days
1.5 days 1.5 days 2 days
Processing
time = 164 sec.
2 sec. 162 sec.

Coil
Weld
change‐over

Takt = 54 sec.
C/T = 162 sec.
C/O = 0 sec.
2 shifts
Uptime = 100%

 

EPE = 1 shift
C/O < 10 min.

Future‐State Value Stream Map: View 4
Total C/T
1.5 days
Weld
uptime
20
20
Figure 7.7b Example of current and future state value stream maps
We do not discuss lean tools and methods any further in this book. If you are interested in
lean tools and methods, a wealth of literature is available on the Web. Remember, each lean
tool can be used to find Xs for improving a CTQ!
Takeaways from This Chapter
The SDSA cycle is a series of four steps used to standardize a process and reduce variation and waste in process outcomes.
The PDSA cycle consists of four steps used to continually improve a process by reducing variation, moving the process toward nominal, and eliminating waste.
From the Library of Pearson HED
228 A Guide to Six Sigma and Process Improvement for Practitioners and Students
A kaizen or Rapid Improvement Event (RIE) is a methodology that consists of a three
to five day event that empowers front-line staff and utilizes their process knowledge
to analyze and make immediate improvements to problematic processes.
Based on the scientific method, the DMAIC model is a five-phase method used to
improve a current process, product, or service.
The DMADV model is a five-phase method used to design a new or redesign a current
process, product, or service.
Lean thinking is focused on promoting the reduction of waste by eliminating nonvalue added steps in a process.
The 5Ss are a lean methodology focused on highlighting and eliminating waste, inconsistency, and unreasonableness from the workplace. It is used to create a neat and tidy
workplace.
TPM is a theory that is useful for maintaining plants and equipment with total
involvement from all employees.
Quick changeover (SMED) is a technique that team members can use to analyze and
then reduce setup time for equipment and people, resources required for a changeover, and materials necessary for a changeover.
Poka-yoke is Japanese for mistake-proofing device. A mistake-proofing device is used
to prevent the causes of defects and/or defective output (called errors), or to inexpensively inspect each item produced to determine whether it is conforming or defective.
Value stream is a term that describes all the value added and non-value added process
steps and decisions necessary to move a product or service from supplier to customer.
References
Deming, W. E. (1994), The New Economics for Industry, Government, Education, 2nd ed.
(Cambridge, MA: Massachusetts Institute of Technology).
Gitlow, H. (2009),
A Guide to Lean Six Sigma (Boca Raton, FL: CRC Press – A Division of
Taylor and Francis).
Gitlow, H., D. Levine, and E. Popovich (2006),
Design for Six Sigma for Green Belts and
Champions: Foundations, DMADV, Tools and Methods, Cases and Certification
(Upper
Saddle River, NJ: Prentice-Hall).
Gitlow, H., A. Oppenheim, R. Oppenheim, and D. Levine (2015),
Quality Management: Tools
and Methods for Improvement,
4th ed. (Naperville, IL: Hercher Publishing Company).
This book is free online at hercherpublishing.com.
Gitlow, H. and D. Levine, D. (2004),
Six Sigma for Green Belts and Champions: Foundations, DMAIC, Tools and Methods, Cases and Certification (Upper Saddle River, NJ:
Prentice-Hall).
From the Library of Pearson HED
Chapter 7 Overview of Process Improvement Methodologies 229
Japan Institute of Plant Maintenance (1997 English Edition), Focused Equipment Improvement for TPM Teams (New York: Productivity Press).
Productivity Development Team (1996),
Quick Changeovers for Operators: The SMED System (New York: Productivity Press).
Rother, M. and J. Shook (2003),
Learning How to See (Cambridge, MA: The Lean Enterprise).
Shingo, S. (1986),
Zero Quality Control: Source Inspection and the Poka-Yoke System (New
York: Productivity Press).
Venkatesh, J. (2005),
An Introduction to Total Productive Maintenance (TPM), The Plant
Maintenance Resource Center, 1996-2005. Revised: Sunday, 27-Feb-2005.
From the Library of Pearson HED
This page intentionally left blank
From the Library of Pearson HED
231
8
Project Identification and Prioritization:
Building a Project Pipeline
What Is the Objective of This Chapter?
The objective of this chapter is to discuss the process used to create the project pipeline at
your organization. We walk you through project identification, project screening and scoping, project prioritization and selecting projects, and finally managing the project pipeline.
The process of creating and managing the pipeline is typically led by a Process Improvement
Executive and senior-level Executive Champions.
Project Identification
The first step in the process is project identification. We have found that potential process
improvement projects can come to the Process Improvement Executive, senior level executives, employees, and customers, to name a few sources, in a variety of ways.
There are four basic categories of ways to identify process improvement projects; they are
summarized in Table 8.1 . Each of the methods in the cells of Table 8.1 is explained in this
chapter.
Table 8.1 Project Identification Matrix

Proactive Reactive
Internal Strategic/tactical plans
Voice of the Employee (VoE) interviews
Employee focus groups or forums
Employee surveys
Employee feedback (suggestions and
complaints)
External Voice of the Customer (VoC) interviews
Customer focus groups
Customer surveys
Regulatory issues dealt with proactively
Customer feedback (suggestions and
complaints)
Regulatory issues dealt with reactively

From the Library of Pearson HED
232 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Internal Proactive
The first category for project identification we discuss is internal proactive sources. Internal
refers to the fact that the data for project identification comes from within the organization;
for example, Voice of the Employee interviews.
Proactive refers to the fact that employees in
the organization actively collect the data for project identification; it does not come passively
to the organization.
Strategic/Tactical Plans
Strategic planning is a process frequently used within organizations to identify the objectives
necessary to pursue the mission; this includes setting the mission and objectives, determining
actions to achieve the objectives, and allocating resources to achieve the objectives. A strategy
describes how the objectives will be achieved given the budget. Many times those actions
involve potential process improvement projects.
Most projects evolve out of an organization’s strategic plan. Each section of the organization
submits, usually on a yearly basis, its strategic plans for the next year. Part of the plan is a gap
analysis that compares the current state with the desired state objectives. If the gap between
the current state and desired state is larger than the organization would like, a tactical thrust
is initiated to close that gap. The tactical thrust in this case is a potential process improvement
project that depending on the gap could be a 5S project, an SDSA project, a PDSA project, a
Six Sigma DMAIC project, a Six Sigma DMADV project, or a Lean project.
Potential strategic/tactical projects may also be identified through internal metrics found on
Organizational dashboardsA dashboard is a set of interlocking and cascading
objectives with appropriate metrics (called
key performance indicators [KPIs] or
CTQs) that go from the top to the bottom of the organization; the objectives and
metrics define employees’ jobs. Many dashboards have adopted the four types of
objectives: employee, process, customer, and financial.
Balanced scorecardsA balanced scorecard is similar to a dashboard, except it
always focuses on four types of objectives: financial, customer, process, or employee
learning and growth. The basic idea is happy employees improve the processes, which
creates customer satisfaction, and ultimately, results in good financial performance.
Voice of the Employee (VoE) Interviews—Finding the Problems
Another way of identifying projects is by having a process improvement professional meet
individually with senior executives or managers to identify potential projects by asking any
of the following questions:
What concerns do you have within your organization/area/department/ division?
What is your actual performance versus your desired performance? (to identify performance gaps)
How are you not meeting your customers’ expectations? What comments or feedback
have you received from your customers?
From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 233
What are your competitive threats?
How is the market changing?
What are your performance results from past years? Are they stable, predictable, and
acceptable?
What are the areas in which you have less dollars (revenue, profit, and so on) than
expected in your strategic plan?
Are you over budget? Is your budget based on rational expectations? If you are exceeding a rational budget where are you exceeding it? What are you going to do about it?
Remember the effects of treating common variation as special variation.
After all these meetings you should have a list of problems and issues. If there are many vague
and similar issues you may want to consolidate them using an affinity diagram. You may
have to meet with the senior leaders again to dig deeper into the issues to get at their root
causes so you understand the project that is needed to resolve the issue(s) caused by some
problematic process.
Employee Focus Groups
An employee focus group is a type of qualitative research. They typically consist of 6 to 12
employees and are facilitated by a moderator. The members of the focus group share an
interest in the well-being of the organization. The purpose of a focus group is to dive deep
into a few high priority issues to learn about employees’ perceptions, feelings, opinions,
beliefs, needs, and wants, which may lead to ideas for process improvement projects.
Focus groups that are well run help uncover feelings and emotions that are not visible in
surveys due to the depth of the questioning on a selected few high priority topics. Employees
that are on the line often see issues not visible to management. Hence, focus groups can be a
valuable tool to identify issues and ideas for potential process improvement projects.
However, if management does not react quickly to the employees’ suggestions, the focus
groups become a joke! At a minimum, management has to listen to the suggestions and get
back to the employees on what they plan to do or not do, but true effectiveness comes when
management acts on the suggestions.
Employee Forums
Many organizations conduct quarterly, semi-annual, or annual employee forums to give their
employees a chance to
Present feedback to management on their areas of the organization.
Communicate issues and concerns on their areas of the organization in particular, or
the entire organization in general.
Ask questions about the organization.
Give their views and opinions about the organization.
From the Library of Pearson HED
234 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Employee forums are another great way to identify potential process improvement projects.
Front-line employees who attend the forums may have their finger on the pulse of the company to a greater degree than management with their areas, and may be exposed to issues that
management wouldn’t ordinarily be aware of. Again, if management does not react quickly
to the employees’ suggestions, the forums become a joke! Again, at a minimum, management
has to listen to the suggestions and get back to the employees with their plans for acting on,
or not acting on, their suggestions, but true effectiveness comes when management acts on
the valuable suggestions.
Employee Surveys
Many organizations distribute employee engagement or satisfaction surveys on an annual
basis. Often these surveys give employees an opportunity to give suggestions on how the
organization may improve. This can be a great way to identify potential process improvement projects.
The problem with employee surveys is that the information is qualitative in nature and is
difficult to analyze. A potential solution is to take the qualitative employee feedback and use
techniques such as the affinity diagram and the cause and effect diagram to organize the data
into meaningful opportunities for improvement of problematic processes. Once again, if
management does not react quickly to the employees’ suggestions, the surveys become a joke!
Internal Reactive
The second category for project identification we discuss is internal reactive sources. Internal
is defined as before—for example, employee feedback. Reactive refers to the fact that the data
comes to the organization as a direct result of doing business; again, like employee feedback.
Employee Feedback (Suggestions or Complaints)
Another way to identify potential process improvement projects is through unsolicited
employee feedback in the form of complaints or suggestions, which may come in various
forms:
Through email
Via suggestion boxes
Complaints filed through human resources
From coworkers, employees, or managers
Valuable information can be obtained from employee feedback so it must be encouraged and
responded to for it to continue. One way to deal with employee feedback is through a closed
feedback loop as shown in Figure 8.1 . First, collect feedback from the preceding sources and
analyze it on a regular basis. Second, investigate whether the feedback (data) indicates that
a problem warrants further action. Make sure that you communicate to employees that you
listened to their feedback. Finally, if you have made changes based on the employee feedback,
refine the changes through continuous improvement; see Figure 8.1 .
From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 235
It is important to ensure that employees are satisfied with the outcome(s) of their input.
So following up with employees who have given feedback may be an investment you want
to make. Often all that employees want is to be listened to and thanked for their feedback.
Remember, if management does not react quickly to the employees’ suggestions, the feedback becomes a joke.
Refine Changes Take Action
Communicate
Feedback
Collect Data
The Closed-Loop Feedback Process
Figure 8.1 The closed-loop feedback process (source: www.mindtools.com)
An Anecdote about Collecting Customer Information

Years ago when you went to buy a new or used car, you would make a deal with the
salesperson and then he would have to go back to the sales manager for approval.
What you didn’t know was that some car retailers had a hidden microphone under
the desk where you were sitting. The sales manager and the salesman were listening
to your conversation about the lowest price you would pay for the car, and of course
used this information against you. The moral to this anecdote is that data collection
is not always done ethically. You need to be careful.

External Proactive
The third category for project identification we discuss is external proactive sources. External
refers to the fact that the data for project identification comes from outside of the organization; for example, Voice of the Customer interviews. Proactive refers to the fact that employees in the organization actively collect the data for project identification; it does not come
passively to the organization.
Voice of the Customer (VoC) Interviews
Another way of identifying projects is by having a process improvement professional meet
individually with your customers (stakeholders) to identify potential projects by asking the
following questions:
From the Library of Pearson HED
236 A Guide to Six Sigma and Process Improvement for Practitioners and Students
How do you feel about our product/process/service?
What do you like about our product/process/service? How can we do better?
What areas can we improve upon with respect to our product/process/service?
How are we not meeting your expectations?
How do you feel about our products/services/processes?
What images and emotions come to mind when you think about our products/
processes/services?
Customer Focus Groups
Like employee focus groups discussed previously, a customer focus group is a type of qualitative research typically consisting of 6 to 12 customers. The group is facilitated by a moderator,
and its members are customers who share an interest in the well-being of the organization.
The purpose of a focus group is to dive deeply into customers’ (stakeholders’) perceptions,
feelings, opinions, beliefs, needs, and wants, which may lead to ideas for process improvement projects.
Focus groups that are well run help uncover feelings and emotions that are not visible in
surveys. Hence, focus groups can be a valuable tool to identify issues and ideas for potential
process improvement projects.
Customer Surveys
Customer surveys can either be proactive or reactive. Both proactive and reactive surveys
may be done by organizations or third-party vendors (for example, Hospital Consumer
Assessment of Health Care Providers and Systems [HCAHPS] inpatient surveys in healthcare) in various different ways:
Phone surveys
Email surveys
Mail surveys
Personal interviews
Surveys must be done properly or they yield garbage information; for more information on
how to conduct a proper survey refer to any basic marketing research book.
External Reactive
The fourth category for project identification we discuss is external reactive sources. External is defined as before; for example, customer feedback. Reactive is defined as before; for
example, reacting to regulatory compliance warnings.
From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 237
Customer Feedback (Suggestions or Complaints)
Another way to identify potential process improvement projects is through unsolicited customer feedback in the form of complaints or suggestions, which may come in various forms:
Through email
Via suggestion boxes
From complaints filed through the customer service department
From customers directly at the point of service
Valuable information can be obtained from customer feedback, so it must be encouraged
and responded to for it to continue. One way to deal with customer feedback is through a
closed feedback loop as shown previously in the closed-loop feedback process in Figure 8.1 .
It is important to manage feedback by ensuring customers are satisfied with the outcome of
their feedback. So following up with customers who have given feedback may be an investment you want to make. Often all that customers want is to be listened to and thanked for
their feedback. Again, nothing gets the customers’ attention like acting on their feedback.
Regulatory Compliance Issues
Regulatory compliance describes an effort organizations make to ensure that they are aware
of and take the necessary steps to ensure they comply with various laws and regulations. For
example, healthcare organizations are typically faced with many compliance requirements
aimed at patient safety and information security, service delivery, operational practices, and
electronic medical record management. Such requirements include various regulatory bodies
and industry standards such as
HIPAA (Health Insurance Portability and Accountability Act)
JCAHO (Joint Commission on Accreditation of Healthcare Organizations)
SOX (Sarbanes Oxley Act)
CMS (Centers for Medicare and Medicaid Services)
The preceding list is merely some common regulatory bodies; there are many more federal
and state regulations for patient safety that healthcare organizations need to comply with, as
well as other organizations, such food and drug companies and utilities. Depending on the
violation, the repercussions for noncompliance vary among the regulatory bodies—from
warnings to fines to exclusion from participating in some programs to being shut down.
Civil and criminal penalties also can be imposed for violations of certain regulations. Many
industries have regulatory bodies that
must receive proper attention—or else !
Process improvement projects related to regulatory compliance can therefore be identified
both proactively and reactively.
From the Library of Pearson HED
238 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Proactive
Organizations with a strong regulatory compliance department can stay ahead of the curve
by staying up to date on the latest rules and regulations, and building objectives and metrics
to track them into their organizational dashboards or balanced scorecards. By establishing
and monitoring key metrics they can identify potential problems immediately and execute
process improvement projects to eliminate them.
Reactive
Periodically, healthcare organizations are subject to either scheduled or unscheduled surveys
and/or visits from various regulatory bodies, for example, the Joint Commission in healthcare organizations. Often they find areas where the organization is not in compliance. To
be compliant, the organization must execute a task, or if the issue is more complicated, the
organization must undertake a process improvement project.
NOTE
Suggestions on issues and potential problems from interviews, focus groups, surveys, employee forums, and the like may result in a plethora of vague qualitative
data. One way to make sense of all these great suggestions is to consolidate them
using an affinity diagram and a cause and effect diagram.
Using a Dashboard for Finding Projects
A managerial dashboard is a tool used by management to clarify and assign accountability for
the “critical few” key objectives, key indicators, and projects/tasks needed to steer an organization toward its mission statement (Gitlow et al., 2015; Gitlow and Levine, 2004). They
do this by creating an interlocking and cascading set of objectives with metrics throughout
the organization, from top to bottom. The objectives clarify employees’ jobs. Managerial
dashboards have both strategic and tactical benefits.
Structure of a Managerial Dashboard
The president’s objectives and indicators emanate from the mission statement (see row 1 and
columns 1 and 2 of Table 8.2 ). Direct reports identify their area objectives and area indicators by studying the president’s key indicators (column 2 of Table 8.2 ) that relate to their
area of responsibility. The outcome of these studies is to identify the key area objectives and
area indicators (see columns 3 and 4 of Table 8.2 ) required to improve the president’s key
Indicator(s) (see column 2) to achieve a desirable state for presidential key objective(s) (see
column 1). This process is cascaded throughout the entire organization until processes are
identified that must be improved or innovated with potential process improvement projects
or tasks (see column 5 of Table 8.2 ).
From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 239
Table 8.2 Generic Managerial Dashboard

Mission Statement: A mission statement is a declaration of the reason for the existence of an
organization. It should be short and memorable, as well as noble and motivational.
President Direct Reports Potential Process
Improvement Projects
or Tasks
Presidential
Objectives
Presidential
Indicators
Area Objectives Area Indicators
Presidential
objectives
that must be
achieved to
attain the mis
sion statement.
A key indicator
is a measure
ment that
monitors the
status of a
key objective.
One or more
presidential
indicators show
progress toward
each presiden
tial objective.
Area objectives
are established
to move each
presidential
indicator in the
proper direction.
One or more
area indicators
show progress
toward each
area objective.
Process improvement
projects or tasks are used to
improve or innovate pro
cesses to move indicators
in the proper direction.

Example of a Managerial Dashboard
Table 8.3 shows an example of a managerial dashboard for the ABC Hospital. The mission of
ABC Hospital is to be the best university teaching hospital in the universe! It has a classic type
of organizational structure led by a Chief Executive Officer, Chief Operating Officer, Chief
Financial Officer, Chief Medical Officer, and Chief Nursing Officer. One of the areas that the
Chief Operating Officer is responsible for is Surgical Services. Table 8.3 shows an example
of a dashboard between the Chief Operating Officer and the Director of Surgical Services.
Table 8.3 Partial Managerial Dashboard for Hospital ABC

Mission Statement: To be the best university teaching hospital in the universe!
Chief Operating Officer (COO) Director of Surgical Services (DSS) Potential Process
Improvement
Projects
COO
Objectives
COO Indicators DSS Objectives DSS Indicators
To have
operating
rooms that
are efficient
and cost
effective
% of first case
surgeries that
start on time by
day
To improve
first case sur
gery start times
% of first start sur
geries that are on
time by surgeon
by day
First case start project
% of surgeries
cancelled day of
surgery by day
To decrease
cancellations
% of cancella
tions by insurance
provider by age of
patient by day
Note: Data indicates
that cancellations are
very low; therefore
this is not a candidate
for a project.

From the Library of Pearson HED
240 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Managing with a Dashboard
Top management uses a dashboard at monthly operations review meetings for several purposes. First, managers use dashboards to clarify key objectives and accountability among
all personnel and areas. Second, managers use dashboards to promote statistical thinking
by monitoring key indicators using control charts. For example, is the sales volume for last
month due to a special or common cause of variation in the selling process? Third, a manager
uses dashboards to develop and test hypotheses concerning potential changes to processes.
A hypothesis test analyzes the effect of a change concept on a key indicator, and hence, on its
objective. Fourth, a manager uses dashboards to ensure the routine and regular updating of
key indicators and to prevent processes from sliding back into their old bad habits.
Managers can use (all or some of) the following questions when conducting a monthly review
meeting to get the most out of their dashboard:
Are the key objectives and key indicators on the dashboard the “best” set of objectives
and indicators to attain the mission statement?
Is the dashboard balanced in respect to employee, process, customer, and financial objectives? Do any areas have too much (or too little) representation on the
dashboard?
What products and/or services are most critical to your organization achieving its
mission statement? List the top five or ten products and/or services.
Are objectives being met in a timely fashion?
What methods are used to manage, perform, and improve the processes that underlie
key objectives?
Which key indicators on the dashboard are used to measure customer satisfaction
and dissatisfaction? Are these measures operationally defined? Are these measures
adequate?
What process is used to motivate employees to work on improvement projects? Hint:
Managers redefine work to include doing work and improving work using the PDSA
cycle. As things improve, employees experience the rush of intrinsic motivation from
an improved and well done job.
Does your organization have the ability to identify the return on investment from its
dashboard? How is return on investment measured?
Project Screening and Scoping
The Process Improvement Executive has a list of potential projects; now what? The next steps
in the process are to screen and scope potential projects as follows:
Determine quickly which ones are worth investigating further. There are questions
you can ask that eliminate projects right away so as not to waste your time and effort.
From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 241
Estimate the potential benefits of each project. If there is no benefit, why do it?
Estimate the potential costs of each project. If the cost are too high in respect to the
benefits, why do it?
Determine which methodology would most likely be used for each potential project,
for example, SDSA, PDSA, DMAIC, and so on.
Create a high-level project charter and problem statement that can be used to prioritize the remaining potential projects.
Questions to Ask to Ensure Project Is Viable
A few questions can be asked right away to disqualify a potential process improvement
project before it wastes valuable time and resources. Some of these questions are as follows:
Is the root cause and solution already known? If the root cause and the solution are
already known, there is no point starting a process improvement project because you
already know what you have to do. So like Nike says, “Just Do It!”
Does the project have firm Champion commitment? If you do not have firm Champion commitment, stop right now and run in the other direction! The number one
key success factor is the commitment of top management (your Champion); if you
don’t have it, find another project or another Champion.
Is data readily available or collectable? The process improvement methodologies we
are teaching you in this book all necessitate data both to determine a baseline for the
current state of the process, as well as to help identify root causes. If data is not readily
available, or at least collectible, you should think long and hard about taking on that
project.
Is the scope narrow? Can it be broken into smaller projects? The last thing you want
is a “solve world hunger” type project; they are almost impossible to execute and you
will lose momentum quickly as their scope is way too big. An alternative is to break
it down into smaller projects; if this is not possible, you probably want to stay away!
Are there organizational constraints that make this project risky? Will key people
in the organization be committed to keep the project on time and moving forward?
Are there political issues that would side track the project? Are key people always
in reactive mode and constantly putting out fires? If there is a potential for team
members to frequently be pulled away to work on “more pressing issues,” or they are
overwhelmed with the “crisis du jour” (fires), you may want to reconsider this project.
If either of these situations exist, the project Champion must really be committed
to eliminate these barriers to a successful project! Additionally, schedule slippage is
something you need to avoid as any loss in momentum can derail your project for
good.
Are there environmental issues that you need to consider? Is there something going
on outside the scope of your organization that will change the landscape in your
industry and eliminate the need for this project?
From the Library of Pearson HED
242 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Estimating Project Benefits
Process improvement projects have both soft benefits and hard benefits (Gitlow et al., 2015;
Gitlow and Levine, 2004). Examples of soft benefits include improving quality and morale,
and decreasing cycle time. Examples of hard (financial) benefits include increasing revenues
or decreasing costs. The Process Improvement Executive “guesstimates” the dollar impact of
the process improvement project so management has an idea of its financial impact during
the project prioritization process. This “guesstimate” will be refined through iterative learning as the project proceeds.
There are two taxonomies for classifying the potential cost related benefits that may be realized from a process improvement project.
Taxonomy 1: Cost Reduction Versus Cost Avoidance
Cost reduction includes costs that fall to the bottom of the profit and loss statement. A cost
reduction can be used to offset price, increase profit, or can be reinvested elsewhere by management. Cost reductions are calculated by comparing the most recent accounting period’s
actual costs with the previous accounting period’s actual costs. Cost avoidance includes
those costs that can be reduced, if management chooses to do so, but until action is taken no
real costs are saved. Examples include reducing labor hours needed to produce some fixed
volume of work. Unless an increased volume of work is completed with the same headcount,
no real savings are realized. The impact of cost avoidance is not visible on the profit and loss
statement and is difficult to define, but is still important in meeting organizational goals.
Taxonomy 2: Tangible Costs Versus Intangible Costs
Tangible costs are easily identified—for example, the costs of rejects, warranty, inspection,
scrap, and rework.
Intangible costs are costs that are difficult to measure—for example, the costs of long cycle
times, many and long setups, expediting costs, low productivity, engineering change orders,
low employee morale, turnover, low customer loyalty, lengthy installations, excess inventory, late delivery, overtime, lost sales, and customer dissatisfaction. It is important to realize
the some of the most important benefits are unknown and unknowable. Hence, the “guesstimate” of benefits in the Define phase of a Six Sigma project often identifies a minimum
estimate of intangible benefits.
The Process Improvement Executive develops a formula to “guesstimate” the potential benefits that the organization may realize due to the process improvement project. For example,
here is a possible formula:

Cost reductions
PLUS Cost Avoidance
PLUS Additional Revenue
LESS Implementation Costs
EQUALS Financial Benefits
_______________________________
_______________________________
_______________________________
_______________________________
_______________________________

From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 243
Project Methodology Selection—Which Methodology Should I Use?
Before prioritizing a project it is usually a good idea to determine the type of methodology you will be using; this allows you to project the time the project will take as well as the
resources needed. Many times it is tough to know what methodology you will use until you
have done further investigation, and sometimes the methodology you thought you would use
changes once you get started.
You always want to choose the methodology that helps you execute your project the quickest
to realize your objectives, while utilizing the least amount of resources. This is a fairly subjective decision that is best made by your Process Improvement Executive who can draw upon
years of experience and leverage his understanding of different organizational factors, such
as resources available, time needed, project complexity, political environment, and so on.
The authors created the flowchart in Figure 8.2 to help you figure out which methodology
to use on a given project.

START
CELEBRATE
SUCCESS
AND PDSA
FOREVER!
STOP
SDSA DMADV/DFSS
Do we need to
create/redesign
or simply
standardize
processes?
Is there
currently a
standardized product/
process/
service?
Can it be
broken into
smaller
projects?
Is scope
narrow?
Is root cause
known?
Is solution
known?
No No
No
Variation
Waste
Yes
Yes
Yes
Standardize New Yes
Lean thinking
Complex
PDSA DMAIC
Simple or
complex
problem?
Decrease waste
or reduce variation?
Which Methodology to Use?

NO PROJECT
(CULTURE OR
MANAGEMENT
ISSUE)
No
No
Yes
Kaizen/Rapid
Improvement
Event
Just do it!
(Project mgmt)
Simple Figure 8.2 Which methodology to use?
From the Library of Pearson HED
244 A Guide to Six Sigma and Process Improvement for Practitioners and Students
The first question we ask is: Is there currently a standardized product/process/service?
If there is not a current standardized product/process/service, then the project is one
of the following:
An SDSA project— If everyone is doing things differently (they each have their
own flowchart/process), team members need to get everyone together and take the
best aspects from their individual flowcharts—that eliminates the weaknesses from
their individual flowcharts—and create one best practice flowchart that everyone
follows. This requires personal discipline. Using the 5Ss helps to develop personal
discipline in the workforce.
DMADV/DFSS project— If there is no process or it is too screwed up to continue,
we have to design one from scratch using the Design for Six Sigma (DMADV)
methodology.
If there is a current standardized product/process/service then:
If the scope is not narrow and it cannot be broken down into smaller projects, it
is likely an organizational culture or management issue and we advise you to stay
away!
If the scope is narrow, the root cause is known, and the solution is known, we call
this a Just Do It project. All you need to do is create a list of action items and project manage it to success.
If the scope is narrow, the root cause is known, and the solution is not known, we
conduct a Kaizen/Rapid Improvement Event over three to five days to come up
with and implement a solution.
If the scope is narrow and the root cause is not known, the question is do we want
to decrease variation/change a problematic CTQ (Y), or do we want to remove
waste in a problematic CTQ?
If we want to decrease variation and the problem is not too complex, we suggest
using the PDSA cycle. If the problem is more complex, we suggest using the Six
Sigma DMAIC model.
If we want to remove waste, we suggest using one of the Lean thinking
methodologies.
NOTE
Complex? A problem may be considered complex if it has some or all of the
following characteristics:
More than one CTQ
Conflicting CTQs (one gets better and the other gets worse)
From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 245
Involves organizational politics
Requires an Information Technology solution
Requires capital investment
Estimating Time to Complete Project
At this point we like to make a rough estimate on how long the project will take so that we
can prioritize and select projects. Many different methodologies can be used to estimate
the time needed to complete a project. Many great project management books cover these
methodologies in detail, so we are merely going to touch on a few of the ones you can use:
Top down (ambiguous) estimating—Develop an estimated timeline based on past
experience and past projects. At this point when we are only looking for a rough estimate to prioritize the project, this is the method typically used. The other methods
listed here can be used after the project actually starts.
Bottom up estimating—Breaking down big tasks into smaller ones and then estimating the time it will take to complete each one.
Three-point estimating—With three-point estimating you come up with an expected
estimate, a pessimistic estimate, and an optimistic estimate for each activity.
Project end date estimating—Sometimes you will be given a hard deadline by when
a project must be completed. In this case you consider the needed deliverables and
resources available to create a timeline working backward.
Expert judgment estimating—Many times you will have an expert who has worked
on a similar project previously. Based on their experience they will have an idea of
how long this type of project will take.
The Process Improvement Executive will estimate the time it takes to complete the project
likely using top down estimating, and then once the project begins, you may want to use one
of the preceding methods to narrow it down even further. At this stage, keep in mind some
things to minimize the time it takes to complete a project, such as:
Choose projects that have reliable and easily accessible data.
Choose projects in areas where teams have total management support (Champion)
and are allowed to meet as often as necessary.
Potential project stakeholders are available for interviews (Interviewing stakeholders
as part of the VoC analysis is time consuming. If your stakeholders have limited availability to participate it will slow you down.)
Commitment by the Champion to change. At some point you will want to test change.
You want to make sure your Champion will cut through any delays; so make sure you
pick projects where this won’t be an issue.
From the Library of Pearson HED
246 A Guide to Six Sigma and Process Improvement for Practitioners and Students
There are other actions you can take to speed up projects that we discuss later when covering
different methodologies. These are just the ones you want to be cognizant of when prioritizing and selecting projects:
The project does not involve politics.
The project requires minimal resources.
The project is under the Process Owner’s and Champion’s control, and they are “all
in” on the project.
The project directly affects the end user.
Creating a High Level Project Charter
The last step of project scoping is creating a high level project charter and problem statement
for each project that you want to add to the pipeline. This will be carried forward to the next
section where it is used to prioritize potential process improvement projects.
The project charter will be refined and expanded in detail during the initial part of the project. This is just to give the steering committee a starting point when prioritizing projects.
Table 8.4 shows a format you can use for a high level project charter.
Table 8.4 Format for High Level Project Charter

Project Name: Identify the issue that is the potential process improve
ment (PI) opportunity.
Department/Area Name: What department or area will be impacted by the project?
High Level Problem Description: Provide a concise description of the issue that needs to be
addressed by the project team.
Project Type: Methodology expected to be used (PDSA, DMAIC, etc.).
Time to complete: Estimated time to complete project using a Gantt chart.
Problematic objective: Problematic objective from the dashboard or other
source.
Problematic indicator: Metric that measures the status of the problematic objec
tive.
Project Scope: Process steps (flowchart) or functional boundaries of the
proposed project. Where does the project start and where
does it stop?
Potential Financial Impact: Projected result of the project and anticipated savings or
revenues to be realized.
Current Baseline Performance: Actual current performance of problematic indicator.

From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 247

Project Champion: Senior executive who reviews projects, removes impedi
ments for the team, and can secure adequate resources
and support for the project team.
Process Owner: The manager of the process under study who has the
authority to change the process.
Finance Champion: Member of the Finance department who can help the
team estimate financial impact of the project.
Project Leader: Process improvement expert who will be the project
leader.
Potential Team Members: Employees who will work on the project part-time.

Problem Statement
The problem statement is a paragraph that states the issue at hand in succinct terms so that
you can understand the problem and its impact for the organization as a whole. A typical
problem statement should be organized as follows:
Our organization is experiencing an issue with ____________________________. The
department/area where this issue is most problematic is _______________________. This
issue has been in existence for _________________. The impact of the issue on the organization is _________________________________________________. It is affecting the
organization by _________________________________________________. The financial
impact of the issue can be estimated at ______________________. Customers affected by
the issue are ________________________________.
Prioritizing and Selecting Projects
Now that you have identified projects, screened out projects that are not worth doing, and
scoped the ones that you will consider doing, you have essentially created a project pipeline.
However, just because a project makes it to the project pipeline does not mean it is a project you will take on. Due to limited time and resources, often you have to further prioritize
projects so that you can maximize both your time and resources. One such methodology is
to prioritize projects using a project prioritization matrix that essentially compares different
projects in your pipeline and ranks them against each other based on certain criteria.
Depending on the organization, prioritization is done by senior executives who have process
improvement training, as well as Process Improvement Executives. They then present the list
of potential projects to a steering committee who makes the final decision on which projects
to execute.
From the Library of Pearson HED
248 A Guide to Six Sigma and Process Improvement for Practitioners and Students
The first step in the process is to select the criteria for which you compare and rank the projects in your pipeline. There are many possible criteria you may use to compare projects. You
can come up with your own, but some criteria we have seen used are the following:
Financial impact (return on investment [ROI])
Data availability
Data quality
Customer satisfaction
Employee satisfaction
Other stakeholder satisfaction (for example, investors, regulators, building, and
grounds)
Ease of implementation
Alignment with mission
Time to complete project (consider methodology and available resources)
Ability to implement solution
Probability of quick results
Low investment cost
Minimum collateral damage on other processes
Scalable and replicable across organization
Resources available
Competitive advantage the project may create
Prioritizing Projects Using a Project Prioritization Matrix
The project prioritization matrix seen in Table 8.5 is a matrix we use to prioritize projects
in our pipeline. The first column contains the evaluation criteria that we use to compare
potential projects. The second column contains the weights that are usually assigned by the
Finance department. The columns to the right contain the different potential projects in our
pipeline.
From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 249
Table 8.5 Project Prioritization Matrix

Evaluation Criteria Weight Project A Project B Project C Project D
Criteria 1 Weight 1
Criteria 2 Weight 2
Criteria 3 Weight 3
Criteria 4 Weight 4
Criteria 5 Weight 5
Criteria 6 Weight 6
Weighted average of
potential projects

The cell values are assigned by top management, and they evaluate the strength of the relationship between each criteria and each potential project.
They are defined as follows:
0 = no relationship between criteria and project
1 = weak relationship between criteria and project
3 = moderate relationship between criteria and project
9 = strong relationship between criteria and project
Each cell value is then multiplied by its respective weight and summed so that each potential
project has a weighted average. The projects are then prioritized by weight, which ranks them
in terms of importance to the organization based on the evaluation criteria. Obviously this
is not an exact science and is subjective in nature, so it is still possible that management will
make an executive decision to go with a project that is lower on the list due to reasons they
deem important.
In our example, we compare the four potential projects in our pipeline in Table 8.6 based on
the criteria we selected and on the weights that came out of our evaluation criteria matrix.
Table 8.6 Project Prioritization Matrix Example

Evaluation Criteria Weight Project A Project B Project C Project D
Financial impact 0.39 9 3 9 3
Customer satisfaction 0.28 9 3 3 9
Employee satisfaction 0.16 3 9 1 9
Probability of quick results 0.12 9 1 3 9
Data availability and quality 0.03 9 3 3 1
Ability to implement solution 0.02 3 1 1 3
Weighted average of potential 7.92 3.68 4.98 6.3

From the Library of Pearson HED
250 A Guide to Six Sigma and Process Improvement for Practitioners and Students
According to our project prioritization matrix the projects are prioritized as follows:
1. Project A (7.92)
2. Project D (6.3)
3. Project C (4.98)
4. Project B (3.6)
So we select project A (7.92) as the highest priority project. We keep going down the list of
projects until resources are depleted.
Final Project Selection
The organization typically creates a process improvement steering committee that balances
and manages the project pipeline. This steering committee is made up of senior executives
who have been through process improvement training, process improvement personnel,
and various key Champions and stakeholders throughout the organization. The committee
reviews the pipeline, including high-level project charters and problem statements, as part
of their monthly meeting and allocates resources they deem most important to the mission
of the organization.
Executing and Tracking Projects
This section of the chapter discusses executing and tracking process improvement projects.
The first part of this section explains the infrastructure needed to allocate resources to projects, while the second section discusses the role and significance of presidential reviews and
monthly steering committee meetings to successful projects.
Allocating Resources to Execute the Projects
Once the projects are selected, the Process Improvement Executive identifies a project leader
(process improvement professional), as well as other change agents with the needed skill
sets, including process improvement professionals, process experts, and process stakeholders. Champions and process improvement experts rely on their understanding of the organization and their employees to assign the proper people to the right projects to leverage
strengths, weaknesses, and time and resource availability to execute projects.
Once the project leader is assigned she rallies the troops with a kickoff meeting that includes
Background on the project
Roles and responsibilities for the project
Methodology to be used by the project
From the Library of Pearson HED
Chapter 8 Project Identification and Prioritization: Building a Project Pipeline 251
Project plan
Communication plan
Risk abatement plan
Project charter
Monthly Steering Committee (Presidential) Reviews
Many organizations have a monthly steering committee review to hold people accountable,
to track progress and benefits, as well as to review the pipeline and any new projects. The
Project Champions of the various projects present updates on their projects to the steering
committee. This is
crucial as it ensures that top management stays committed to the projects
in their areas. The project teams create the presentation and materials, and can help answer
detailed questions, but it must be the Champion who does the talking. The steering committee also reviews and prioritizes the pipeline if resources are available to execute new projects.
Takeaways from This Chapter
Creating a project pipeline consists of four main steps: project identification, project
screening and scoping, prioritizing projects, and managing the pipeline.
There are four basic categories of ways to identify process improvement projects:
internal proactive, internal reactive, external proactive, and external reactive.
Another way to identify projects presented by the authors is using managerial dashboards or balanced scorecards.
Project scoping consists of asking questions to ensure a project is viable, calculating
potential financial benefits, figuring out which methodology would potentially be
used to execute the project, estimating the time the project will take, and finally creating a high level project charter.
Next, the team prioritizes projects by selecting and ranking criteria to compare potential projects using a project prioritization matrix.
The steering committee then selects the projects the team will be working on.
The Champion and Process Improvement Executive then allocate resources to execute the projects.
Finally a monthly steering committee meeting is held to hold people accountable, to
track progress and benefits, as well as to review the pipeline and any new projects.
From the Library of Pearson HED
252 A Guide to Six Sigma and Process Improvement for Practitioners and Students
References
Gitlow, H., A. Oppenheim, R. Oppenheim, and D. Levine (2015), Quality Management: Tools
and Methods for Improvement
, 4th ed. (Naperville, IL: Hercher Publishing Company).
This book is free online at hercherpublishing.com.
Gitlow, H. and D. Levine (2004),
Six Sigma for Green Belts and Champions: Foundations,
DMAIC, Tools and Methods, Cases and Certification,
(Upper Saddle River, NJ:
Prentice-Hall).
From the Library of Pearson HED
253
9
Overview of Six Sigma Management
What Is the Objective of This Chapter?
The objective of this chapter is for you to learn what you need to know about the fundamentals of Six Sigma including its non-technical and technical definitions, where it came from, its
benefits, the key ingredient for success with Six Sigma, roles and responsibilities, and finally
some terminology you need to know to implement Six Sigma at your organization. Six Sigma
can be defined with both non-technical and technical definitions.
Non-Technical Definition of Six Sigma Management
You can think of the non-technical definition of Six Sigma management as your elevator
pitch (60 second response). Six Sigma management can be defined as the relentless and rigorous pursuit of the reduction of variation in all critical processes to achieve continuous and
breakthrough improvements that impact the bottom-line and/or top-line of the organization
and increase customer satisfaction (Gitlow et al., 2015).
Technical Definition of Six Sigma
The technical definition is a bit more complicated to understand; hence we are placing it in
Appendix 9.1 at the end of this chapter. For those readers with some background in statistics,
read the appendix. For those readers without an understanding of basic statistics, skip it for
now. Essentially, the technical definition states that a process should be improved so that it
generates no more than 3.4 defects per million opportunities (Gitlow et al., 2015).
Where Did Six Sigma Come From?
Six Sigma was originally developed by Bill Smith at Motorola in 1985. He created the Six
Sigma methodology to increase profitability while reducing defects by introducing the concept of
latent defect, which revolved around reducing variation in processes that would then
reduce defects to improve customer satisfaction and save money. Jack Welch at General
Electric and Larry Bossidy at Allied Signal popularized the Six Sigma approach in the early
From the Library of Pearson HED
254 A Guide to Six Sigma and Process Improvement for Practitioners and Students
and mid-1990s by attributing their increase in market capitalization to the results attained
by their drive for Six Sigma quality (Gitlow and Levine, 2004).
Benefits of Six Sigma Management
There are two types of benefits from Six Sigma management: benefits to the organization
and benefits to stakeholders. Benefits to an organization are gained through the continuous
reduction of variation and centering of processes on their desired (nominal) levels (Gitlow
et al., 2015). The benefits to the organization are the following:
Improved process flows .
Reduced total defects .
Improved communication (provides a common language to everyone involved with
a process) .
Reduced cycle times .
Enhanced knowledge (and enhanced ability to manage that knowledge) .
Higher levels of customer and employee satisfaction .
Increased productivity .
Decreased work-in progress (WIP) .
Decreased inventory .
Improved capacity and output .
Increased quality and reliability .
Decreased unit costs .
Increased price flexibility .
Decreased time to market .
Faster delivery time .
Conversion of improvements into hard currency .
Benefits to stakeholders are a by-product of the organizational benefits. The benefits to
stakeholders include
Stockholders receive more profit due to decreased costs and increased revenues.
Customers are delighted with products and services.
Employees experience higher morale and more satisfaction from joy in work.
Suppliers enjoy a secure source of business.
From the Library of Pearson HED
Chapter 9 Overview of Six Sigma Management 255
Key Ingredient for Success with Six Sigma Management
The key ingredient for a successful Six Sigma management process is the commitment (not
only support) of top management. Executives must have a burning desire to transform their
organizations into Six Sigma enterprises. This means total commitment from the top to the
bottom of the organization. Additionally, another ingredient for a successful Six Sigma management process is a labor force capable of following, and constantly improving, best practice
methods for the processes they are part of. This requires personal discipline. My favorite
example of employees desiring to exhibit personal discipline in following the best practice
method is in training infantry on how to fight and protect themselves in a war zone. Why
would someone not follow the best practice method that has been honed over thousands of
years to keep themselves alive?
Six Sigma Roles and Responsibilities
Several jobs in an organization are critical to the Six Sigma management process. They are
Senior Executive (CEO or President), Executive Committee (Senior Vice Presidents), Champion, Master Black Belt, Black Belt, Green Belt, and Process Owner. The roles and responsibilities of each of these jobs are described as follows (Gitlow et al., 2015; Gitlow and Levine,
2004).
Senior Executive
The Senior Executive provides the impetus, direction, and alignment necessary for Six Sigma’s ultimate success. The Senior Executive should do the following:
Study Six Sigma management.
Lead the executive committee in linking objectives and metrics to Six Sigma projects
(dashboards).
Participate on appropriate Six Sigma project teams.
Maintain an overview of the system to avoid suboptimization.
Maintain a long-term view.
Act as a liaison to Wall Street, explaining the long-term advantages of Six Sigma management, if appropriate.
Constantly and consistently, publicly and privately, champion Six Sigma management.
Improve the Six Sigma process.
Conduct project reviews at the monthly operations review meeting.
The most successful, highly publicized Six Sigma efforts have one thing in common—unwavering, clear, and committed leadership from top management. There is no doubt in anyone’s
mind that Six Sigma is “the way we do business.” Although it may be possible to initiate Six
Sigma concepts and processes at lower levels, dramatic success is not possible until the Senior
Executive becomes engaged and takes a leadership role.
From the Library of Pearson HED
256 A Guide to Six Sigma and Process Improvement for Practitioners and Students
Executive Steering Committee
The members of the executive committee are the top management of an organization; they
report directly to the CEO. They should operate at the same level of commitment for Six
Sigma management as the Senior Executive. They should
not be allowed to follow their own
management style in their areas even if they deliver good results. The reason for this is t