You are on page 1of 312

Business Process

Improvement Toolbox
Second Edition
Also available from ASQ Quality Press:

Root Cause Analysis: Simplified Tools and Techniques, Second Edition

Bjørn Andersen and Tom Fagerhaug
Bringing Business Ethics to Life: Achieving Corporate Social Responsibility
Bjørn Andersen
Performance Measurement Explained: Designing and Implementing Your
State‑of-the-Art System
Bjørn Andersen and Tom Fagerhaug
Managing Service Delivery Processes: Linking Strategy to Operations
Jean Harvey
Integrating Inspection Management into Your Quality Improvement System
William D. Mawby
Enterprise Process Mapping: Integrating Systems for Compliance and
Business Excellence
Charles G. Cobb
The Process-Focused Organization: A Transition Strategy for Success
Robert A. Gardner
Decision Process Quality Management
William D. Mawby
The Process Auditing Techniques Guide
J.P. Russell
Defining and Analyzing a Business Process: A Six Sigma Pocket Guide
Jeffrey N. Lowenthal
The Certified Manager of Quality/Organizational Excellence Handbook,
Third Edition
Russell T. Westcott, editor
Making Change Work: Practical Tools for Overcoming Human Resistance
to Change
Brien Palmer

To request a complimentary catalog of ASQ Quality Press publications, call

800‑248-1946, or visit our Web site at
Business Process
Improvement Toolbox
Second Edition

Bjørn Andersen

ASQ Quality Press

Milwaukee, Wisconsin
American Society for Quality, Quality Press, Milwaukee 53203
© 2007 by American Society for Quality
All rights reserved. Published 2007
Printed in the United States of America
12  11  10  09  08  07   5  4  3  2  1

Library of Congress Cataloging-in-Publication Data

Andersen, Bjorn.
  Business process improvement toolbox / Bjorn Andersen.—2nd ed.
    p. cm.
  ISBN 978-0-87389-719-8 (alk. paper)
  1.  Reengineering (Management)  2.  Benchmarking (Management) 
  3.  Reengineering (Management)—Charts, diagrams, etc.  I.  Title.

  HD58.87.A53 2007

No part of this book may be reproduced in any form or by any means, electronic,
mechanical, photocopying, recording, or otherwise, without the prior written permission
of the publisher.

Publisher: William A. Tony

Acquisitions Editor: Matt Meinholz
Project Editor: Paul O’Mara
Production Administrator: Randall Benson

ASQ Mission: The American Society for Quality advances individual, organizational, and
community excellence worldwide through learning, quality improvement, and knowledge

Attention Bookstores, Wholesalers, Schools, and Corporations: ASQ Quality Press

books, videotapes, audiotapes, and software are available at quantity discounts with bulk
purchases for business, educational, or instructional use. For information, please contact
ASQ Quality Press at 800-248-1946, or write to ASQ Quality Press, P.O. Box 3005,
Milwaukee, WI 53201-3005.

To place orders or to request a free copy of the ASQ Quality Press Publications Catalog,
including ASQ membership information, call 800-248-1946. Visit our Web site at or

  Printed on acid-free paper


List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv

Chapter 1­  Business Process Improvement . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Why Is Improvement Necessary? . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 A Business Process Improvement Framework . . . . . . . . . . . . . . . . . 4

Chapter 2 Understanding the Organization’s Stakeholders

and Strategic Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Stakeholder Expectations of Improvement . . . . . . . . . . . . . . . . . . . . 9
2.2 Reviewing the Organization’s Strategy . . . . . . . . . . . . . . . . . . . . . . . 17
2.3 SWOT Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 Competitive Forces Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5 Strategy Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Chapter 3  Understanding Your Current Business Processes . . . . . . . . . 27

3.1 From Processes to Departments to Business Processes . . . . . . . . . . 27
3.2 Definition of a Business Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3 Classification of Business Processes . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4 Identifying the Organization’s Business Processes . . . . . . . . . . . . . . 37
3.5 Business Process Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.6 Relationship Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.7 Traditional Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.8 ­ Cross-­Functional Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.9 Flowchart Divided into Process Areas . . . . . . . . . . . . . . . . . . . . . . . . 54
3.10 ­ Several-­Leveled Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.11 Flowchart with Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.12 Paper and Pencil or PC? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

vi  Contents

Chapter 4 Using Performance Measurement in Business Process

Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.1 The Role of Performance Measurement in Business Process
Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2 Implementing a Performance Measurement System . . . . . . . . . . . . . 67
4.3 Performance Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.3.1 Hard versus Soft Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.3.2 Financial versus Nonfinancial Indicators . . . . . . . . . . . . . . . . 71
4.3.3 Leading versus Lagging Indicators . . . . . . . . . . . . . . . . . . . . . 72

Chapter 5  Creating a Business Process Improvement Road Map . . . . . 75

5.1 Improvement Project Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2 Trend Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.3 Spider Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.4 Performance Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.5 Criteria Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.6 Quality Function Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

Chapter 6  Organizing for Business Process Improvement . . . . . . . . . . 93

6.1 Business Process Improvement Skills . . . . . . . . . . . . . . . . . . . . . . . . 93
6.2 Organizational Modes That Support Business Process Improvement 95
6.3 Quality Circles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
6.4 Cross-­Functional Teams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
6.5 Stimulating an Improvement Culture . . . . . . . . . . . . . . . . . . . . . . . . 103

Chapter 7  The Business Process Improvement Toolbox . . . . . . . . . . . . . 107

7.1 The Need for a Toolbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
7.2 A Business Process Improvement Model . . . . . . . . . . . . . . . . . . . . . 108
7.3 The Tools in the Toolbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

Chapter 8 Tools for Collecting Data about the

Performance Shortcoming . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
8.1 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
8.2 Surveying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
8.3 Check Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
8.4 Problem Concentration Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

Chapter 9  Tools for Analyzing the Performance Shortcoming . . . . . . . 123

9.1 Critical Incident . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
9.2 Pareto Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
9.3 Cause-­and-Effect Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
9.4 Five Whys Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Contents  vii

9.5 Scatter Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

9.6 Histogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
9.7 Relations Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
9.8 Matrix Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
9.9 Is–Is Not Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
9.10 Bottleneck Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

Chapter 10 Tools for Generating Ideas and Choosing

among Them . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
10.1 Brainstorming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
10.2 Brainwriting/Crawford Slip Method . . . . . . . . . . . . . . . . . . . . . . . . . 159
10.3 Six Thinking Hats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
10.4 Nominal Group Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
10.5 Paired Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

Chapter 11  Tools for Creating Improvements . . . . . . . . . . . . . . . . . . . . . 167

11.1 Streamlining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
11.1.1 Bureaucracy Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
11.1.2 Redundancy Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
11.1.3 ­ Value-­Added Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
11.1.4 Process Cycle Time Reduction . . . . . . . . . . . . . . . . . . . . . . . . 172
11.2 Idealizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
11.3 QFD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
11.4 Statistical Process Control/Control Chart . . . . . . . . . . . . . . . . . . . . . 182
11.4.1 Definitions of Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
11.4.2 Basic Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
11.4.3 Types of Control Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
11.4.4 Constructing Control Charts . . . . . . . . . . . . . . . . . . . . . . . . . . 189
11.4.5 Interpreting the Control Charts . . . . . . . . . . . . . . . . . . . . . . . . 196
11.4.6 Process Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
11.5 Six Sigma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
11.5.1 The Six Sigma Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
11.5.2 Six Sigma in the Organization . . . . . . . . . . . . . . . . . . . . . . . . 206
11.6 Business Process Reengineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
11.6.1 A Process for Conducting BPR . . . . . . . . . . . . . . . . . . . . . . . . 209
11.6.2 Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
11.6.3 Reengineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
11.6.4 Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
11.6.5 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
11.7 Benchmarking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
11.7.1 Definitions of Benchmarking . . . . . . . . . . . . . . . . . . . . . . . . . 221
11.7.2 Conducting a Benchmarking Study . . . . . . . . . . . . . . . . . . . . 227
viii  Contents

Chapter 12  Tools for Implementing Improvements . . . . . . . . . . . . . . . . 237

12.1 A∆T Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
12.2 Tree Diagram and Process Decision Program Chart . . . . . . . . . . . . . 241
12.3 Force Field Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Chapter 13  Sample University—Improving Student Satisfaction . . . . . 251
13.1 Description of the Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
13.2 Development of Performance Priorities . . . . . . . . . . . . . . . . . . . . . . 252
13.3 Understanding the Current Processes and
Performance Shortcoming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
13.4 Analysis of the Performance Shortcoming . . . . . . . . . . . . . . . . . . . . 257
13.5 Generating Ideas and Improvement Proposals . . . . . . . . . . . . . . . . . 258
13.6 Implementation of Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
Chapter 14  Green Carpet Seed—Revamping the Business Model . . . . 263
14.1 Description of the Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
14.2 Development of Performance Priorities . . . . . . . . . . . . . . . . . . . . . . 264
14.3 Understanding the Current Processes and
Performance Shortcoming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
14.4 Generating Improvement Ideas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
14.5 Developing an Improved Process . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
14.6 Implementation of the New Service . . . . . . . . . . . . . . . . . . . . . . . . . 271
Appendix  Template Package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
SWOT Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
Trend Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
Spider Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Performance Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Criteria Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
Quality Function Deployment (QFD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Relationship Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
Cross-Functional Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
Check Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Pareto Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Cause-and-Effect Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
Five Whys Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
Scatter Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Histogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Relations Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
Is–Is Not Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
Paired Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
Control Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Force Field Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
List of Figures

Figure 1.1 Without maintenance and improvement, the

performance level decreases. . . . . . . . . . . . . . . . . . . . . . . . . . 3
Figure 1.2 Business process improvement framework. . . . . . . . . . . . . . . 5
Figure 1.3 The Deming wheel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Figure 2.1 The stakeholder model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Figure 2.2 Stakeholder classification matrix. . . . . . . . . . . . . . . . . . . . . . . 12
Figure 2.3 The Kano model and the three types of stakeholder
requirements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Figure 2.4 Stakeholder strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Figure 2.5 SWOT analysis results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Figure 2.6 Five competitive forces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Figure 2.7 Generic strategy map. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Figure 2.8 Strategy map for the shipping company. . . . . . . . . . . . . . . . . . 24
Figure 3.1 The contradiction between vertical departments and
horizontal processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Figure 3.2 A simple business process. . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Figure 3.3 The same simple business process, now with
department boundaries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Figure 3.4 Primary and support processes. . . . . . . . . . . . . . . . . . . . . . . . . 35
Figure 3.5 Business processes in ENAPS. . . . . . . . . . . . . . . . . . . . . . . . . 36
Figure 3.6 Business processes derived from strategy and stake-
holders. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Figure 3.7 The hierarchy of process modeling.. . . . . . . . . . . . . . . . . . . . . 41
Figure 3.8 Example of a relationship map. . . . . . . . . . . . . . . . . . . . . . . . 44
Figure 3.9 The resulting relationship map for the supply process. . . . . . 46
Figure 3.10 Flowchart for the supply process. . . . . . . . . . . . . . . . . . . . . . . 47
Figure 3.11 Flowchart for the filing process. . . . . . . . . . . . . . . . . . . . . . . . 49
Figure 3.12 Standard IDEF0 diagram design. . . . . . . . . . . . . . . . . . . . . . . 50

  List of Figures

Figure 3.13  ample IDEF0 diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

S 50
Figure 3.14 Example of a ­cross-­functional flowchart. . . . . . . . . . . . . . . . . 52
Figure 3.15 Cross-functional flowchart for the reporting process. . . . . . . 53
Figure 3.16 Example of a flowchart divided into process areas. . . . . . . . . 55
Figure 3.17 Level 0 chart for the process. . . . . . . . . . . . . . . . . . . . . . . . . . 56
Figure 3.18 Level 1 chart for procurement. . . . . . . . . . . . . . . . . . . . . . . . . 57
Figure 3.19 Level 1 chart for reporting. . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Figure 3.20 Flowchart with bank customer movement statistics. . . . . . . . 61
Figure 4.1 The performance measurement system design process. . . . . . 69
Figure 4.2 Performance measurement as an early warning system. . . . . 73
Figure 5.1 Trend analysis for three example measures. . . . . . . . . . . . . . . 77
Figure 5.2 Results of the trend analysis. . . . . . . . . . . . . . . . . . . . . . . . . . 78
Figure 5.3 Example of a spider chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Figure 5.4 Spider chart for the central performance measures. . . . . . . . . 80
Figure 5.5 A generic performance matrix. . . . . . . . . . . . . . . . . . . . . . . . . 82
Figure 5.6 Performance matrix for the six performance measures. . . . . . 83
Figure 5.7 Matrix for criteria testing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Figure 5.8 Criteria testing for a hair salon chain. . . . . . . . . . . . . . . . . . . . 85
Figure 5.9 The basic structure of QFD. . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Figure 5.10 A chain of QFD charts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Figure 5.11 The house of quality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Figure 5.12 Benchmarking of others’ offers. . . . . . . . . . . . . . . . . . . . . . . . 88
Figure 5.13 Symbols for the relational matrix. . . . . . . . . . . . . . . . . . . . . . 88
Figure 5.14 Symbols for the roof matrix. . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Figure 5.15 QFD chart for an internal IT department. . . . . . . . . . . . . . . . . 91
Figure 6.1 Traditional organization with vertical silos. . . . . . . . . . . . . . . 96
Figure 6.2 Organizing along business processes with process
owners. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Figure 6.3 Organization of quality circles. . . . . . . . . . . . . . . . . . . . . . . . . 102
Figure 7.1 Stages in business process improvement work. . . . . . . . . . . . 109
Figure 8.1 Example of a check sheet. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Figure 8.2 Check sheet for recording the causes of lost bids. . . . . . . . . . 119
Figure 8.3 Cutout of city map with streetlight problem
concentration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Figure 9.1 Critical incidents for the switchboard operators. . . . . . . . . . . 125
Figure 9.2 A general Pareto chart with a line for cumulative
importance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Figure 9.3 Pareto chart for value of lost jobs. . . . . . . . . . . . . . . . . . . . . . 127
Figure 9.4 The structure for a fishbone chart. . . . . . . . . . . . . . . . . . . . . . 128
Figure 9.5 Process chart with fishbone charts for each process
step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Figure 9.6 Fishbone chart for shaft manufacturing. . . . . . . . . . . . . . . . . . 131
Figure 9.7 List representation of five whys. . . . . . . . . . . . . . . . . . . . . . . . 132
Figure 9.8 Five whys analysis chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
List of Figures  xi

Figure 9.9  catter charts showing different degrees of correlation. . . . . .

S 134
Figure 9.10 Scatter chart to analyze weather and jobs. . . . . . . . . . . . . . . . 136
Figure 9.11 Check sheet for the example. . . . . . . . . . . . . . . . . . . . . . . . . . 138
Figure 9.12 Histogram for the example. . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Figure 9.13 Centered process with a small variation width. . . . . . . . . . . . 140
Figure 9.14 Centered process with a large variation width. . . . . . . . . . . . . 140
Figure 9.15 Process with a small variation width, but not centered. . . . . . 141
Figure 9.16 Histogram with two peaks. . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Figure 9.17 Histogram that is cut off. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Figure 9.18 The principles of a qualitative relations diagram. . . . . . . . . . . 142
Figure 9.19 Qualitative relations diagram for a poorly functioning
measurement system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Figure 9.20 A general quantitative relations diagram. . . . . . . . . . . . . . . . . 144
Figure 9.21 Quantitative relations diagram for the example. . . . . . . . . . . . 145
Figure 9.22 Types of matrix diagrams. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
Figure 9.23 Relations symbols. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Figure 9.24 Matrix diagram for the video rental chain. . . . . . . . . . . . . . . . 148
Figure 9.25 Generic table for is–is not analysis. . . . . . . . . . . . . . . . . . . . . 149
Figure 9.26 Ski resort is–is not analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . 150
Figure 9.27 A bottleneck resource limits the flow from the
entire process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
Figure 9.28 Emergency ward bottleneck analysis. . . . . . . . . . . . . . . . . . . . 154
Figure 10.1 Individual ranking card. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Figure 10.2 Resulting flip chart list. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Figure 10.3 Paired comparisons matrix for internal communication. . . . . 166
Figure 11.1 ­Value-­added analysis questions. . . . . . . . . . . . . . . . . . . . . . . . 171
Figure 11.2 ­Cost-­cycle time chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
Figure 11.3 Corresponding chart after completed ­value-­added
analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Figure 11.4 A flowchart with steps marked according to
streamlining results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
Figure 11.5 ­Cost-­cycle time chart for the original process. . . . . . . . . . . . . 176
Figure 11.6 ­Cost-­cycle time chart for the streamlined process. . . . . . . . . . 176
Figure 11.7 Analysis of the current and ideal process in idealizing . . . . . . 177
Figure 11.8 Idealizing in a printing company. . . . . . . . . . . . . . . . . . . . . . . 178
Figure 11.9 The general house of quality, the chart used by QFD. . . . . . . 180
Figure 11.10 House of quality for process improvement. . . . . . . . . . . . . . . 181
Figure 11.11 Normal distribution curve. . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Figure 11.12 Types of control charts and when to use them. . . . . . . . . . . . . 188

Figure 11.13 Example of an X chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Figure 11.14 Example of an R chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Figure 11.15 Control chart with one point outside the control limits. . . . . 197
Figure 11.16 Chart with several points outside the control limits. . . . . . . . . 198
Figure 11.17 Process with change in level. . . . . . . . . . . . . . . . . . . . . . . . . . 198
xii  List of Figures

Figure 11.18 A
 series of 12 or 14 points on the same side of
the center line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Figure 11.19 Process with trend. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Figure 11.20 Process probably involving tampering. . . . . . . . . . . . . . . . . . . 200
Figure 11.21 Chart displaying cycles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Figure 11.22 p chart for one day’s production. . . . . . . . . . . . . . . . . . . . . . . 203
Figure 11.23 Control chart divided into operators. . . . . . . . . . . . . . . . . . . . 203
Figure 11.24 The process for conducting a BPR project. . . . . . . . . . . . . . . 210
Figure 11.25 Process design using the clean sheet approach. . . . . . . . . . . . 215
Figure 11.26 The current process for developing a tailor-­made system. . . 219
Figure 11.27 The ideal process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Figure 11.28 Operational definition of benchmarking. . . . . . . . . . . . . . . . . 222
Figure 11.29 Benchmarking out of the box. . . . . . . . . . . . . . . . . . . . . . . . . . 224
Figure 11.30 Three types of benchmarking based on what is
being compared. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Figure 11.31 Recommended combinations of types of
benchmarking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Figure 11.32 Progress of a typical benchmarking study. . . . . . . . . . . . . . . . 226
Figure 11.33 Models for organizing a benchmarking study. . . . . . . . . . . . . 226
Figure 11.34 The benchmarking process based on the
benchmarking wheel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
Figure 11.35 Ambition pyramid for benchmarking partners. . . . . . . . . . . . 230
Figure 11.36 Three levels of benchmarking information. . . . . . . . . . . . . . . 231
Figure 11.37 Methods and tools for the observation phase. . . . . . . . . . . . . . 232
Figure 12.1 Matrix for assessing implementation order. . . . . . . . . . . . . . . 238
Figure 12.2 Flowchart for the A∆T analysis. . . . . . . . . . . . . . . . . . . . . . . . 240
Figure 12.3 A principal tree diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Figure 12.4 Tree diagram for introducing a computer system. . . . . . . . . . 243
Figure 12.5 A general PDPC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
Figure 12.6 PDPC for the library. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Figure 12.7 Force field diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Figure 12.8 Force field analysis for a new library computer system . . . . . 248
Figure 13.1 Spider chart for external comparison. . . . . . . . . . . . . . . . . . . . 253
Figure 13.2 Performance matrix for establishing priorities. . . . . . . . . . . . 254
Figure 13.3 Flowchart for the course process. . . . . . . . . . . . . . . . . . . . . . . 255
Figure 13.4 Fishbone chart for lacking communication ability. . . . . . . . . . 257
Figure 13.5 Matrix diagram for Sample University. . . . . . . . . . . . . . . . . . 258
Figure 13.6 Force field analysis for the implementation. . . . . . . . . . . . . . . 261
Figure 14.1 Criteria testing matrix for Green Carpet Seed. . . . . . . . . . . . . 266
Figure 14.2 ­Cross-­functional flowchart for the
dealer management and support process. . . . . . . . . . . . . . . . . 267
Figure 14.3 PDPC for Green Carpet Seed. . . . . . . . . . . . . . . . . . . . . . . . . . 273
List of Tables

Table 4.1 Differences between hard and soft indicators. . . . . . . . . . . . . 70

Table 4.2 Examples of financial indicators. . . . . . . . . . . . . . . . . . . . . . . 72
Table 7.1 Tools for different phases in the improvement work. . . . . . . . 110
Table 9.1 Measures of the diameter of a hole for 125 work pieces. . . . . 137
Table 9.2 Determining the number of classes, C. . . . . . . . . . . . . . . . . . . 137
Table 9.3 Formats with the number of variables and relations for matrix
diagrams. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Table 11.1 Calculation factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
Table 11.2 Factors for estimating the standard deviation. . . . . . . . . . . . . 201
Table 11.3 Focus areas for systematic reengineering. . . . . . . . . . . . . . . . 212
Table 11.4 Breakthroughs through benchmarking. . . . . . . . . . . . . . . . . . . 224


n the years since the first edition of this book was published, in 1999, I have
received a great deal of feedback from readers. The book has been used as
curriculum in many different courses, including some that I teach. The stu‑
dents have responded positively, and those without much work experience find the
numerous examples useful. The larger bulk of readers are, I believe, people who
work with process improvements in their jobs. They seem to come from all types
of sectors and industries, and again the feedback has been predominantly positive.
I never intended this book to be an academic text that would further the state of the
art in this field, and this is probably one reason for the book’s reception. There are
between 12,000 and 15,000 ­first-­edition copies in circulation.
From time to time, I flip through the book, either to review some topic before
a lecture or to explain something to students in more detail. Whenever I do, I
cannot help noticing many shortcomings in the first edition. I have learned a lot
about improvement work during these years, and the general knowledge about
these issues has increased significantly. In the end, it became clear that a revision
was in order if the book is to remain useful.
The second edition contains quite a few extensive changes. First, the entire
structure of the first half of the book has been reworked. I was never too happy
with the overall improvement model that formed the basis for the first edition, and
I believe this is much clearer now. Material on how to organize improvement work
and create a culture that supports a relentless quest for improvement has been added,
as have additional details on strategic and performance planning. Some new tools
and techniques have been added to the toolbox, and a new, extensive case study,
Green Carpet Seed, should help facilitate the more coherent use of the toolbox.
As for the first edition, the objective is still to give readers practical insight into
how they can create a coherent business process improvement system. I genuinely
believe that consistently working on improving various aspects of how things are
done, large and small, is the key to success for any organization. If you can com‑

xvi  Preface

bine a culture that never thinks the organization is where it needs to be and a ­well-
­stocked toolbox, you will have a good chance of a successful future.
The first half of the book presents an overall business process improvement
model. The ensuing chapters deal with understanding and modeling your current
business processes, using performance measurement in improvement work, creat‑
ing a business process improvement road map, and organizing for improvement
work. The second half of the book presents the overall toolbox, and each phase of
the overall improvement model is discussed in its own chapter. For each of these
phases, a selection of suitable tools is presented, which includes information on
each tool’s background, steps for using the tool, and an example of its use. The
last two chapters contain two more extensive case studies that illustrate use of the
full methodology. And finally, a number of templates that support most of the tools
presented can be found in the appendix.
I should point out that the book is suited for employees and managers at any
organizational level in any type of industry, including service, manufacturing, and
the public sector. It is also useful as a textbook for students in courses relating to
quality management and continuous improvement.
My sincere thanks go to all those who have inspired the writing of this book,
including colleagues, classroom participants in courses, and companies with which
I have worked. Thanks also to my wife, Hilde Mentzoni Andersen, for reviewing
the draft. And finally, I must express my gratitude to all the great people at ASQ
Quality Press, who have performed excellently at every stage of the process.

Bjørn Andersen
Trondheim, Norway, March 1, 2007
Business Process Improvement

o some, the term improvement is well understood and almost second nature.
I guess I fall into that category, but I also understand that continually think‑
ing about improvement is more alien to others. It therefore seems logical to
begin with an introduction to improvement in general.

1.1  Why Is Improvement Necessary?

Take almost any area of life or any industrial sector. If you trace its history, you’ll
realize that dramatic changes have occurred, most likely in a relatively short time
span. These changes apply to the products or services offered in the marketplace,
the technologies used, the way activities are organized, and so on.
For example, consider travel. Decades ago, only a very small portion of the
population, regardless of which part of the world you look at, had access to or the
means to travel. Air travel was prohibitively expensive for most people, and rail and
sea travel were too slow to allow people to go great distances. The only mode of
transportation available to large parts of the population in some countries was the
automobile. Although there is room for improvement in the world of travel, there
is no doubt that there have been significant changes. In most parts of the world, air
travel is now quite affordable. People are traveling everywhere, and more people
are seeing distant shores than at any other time in history. From an environmental
point of view, travel has probably become too accessible.
Or take telecommunication. I remember when my parents decided it was too
cumbersome to walk two blocks to the nearest phone booth to make a call, so they
had a phone line from the ­state-­owned monopoly provider installed in the house.
This was in the 1970s. The waiting time was about four months, and the phone
you got came in ­standard-­issue gray, with no other options available. Today, no
one would accept this type of customer service, and fortunately we don’t have to.

  Chapter One

In most countries, telecommunication, along with many other basic services that
used to be run by the state, has been deregulated. As customers, we can choose
our supplier, and the range of services has grown immensely. I have traveled a bit
over the last few years, and I am still amazed by how easy it is to turn on my cell
phone or computer with a wireless Internet card and have both work. My last trip
took me to Venezuela, and I had been told that in order to use my cell phone, I’d
have to buy a separate SIM card from a local provider. But when I arrived, both the
phone and the ­high-­speed Internet worked right away. Turns out the information I
had was two months old.
There have also been considerable changes in the field of electronics. Do you
own a DVD player, or maybe a DVD recorder, with a hard drive of x gigabytes?
Perhaps you own an MP3 player? Many of us own one or more of these units, and
we take it for granted that new technology will emerge before we wear out our cur‑
rent equipment. When I bought my first CD burner, I knew it was only a matter of
time before the DVD burner was available. And when I got my first DVD burner,
the consumer electronics industry had already made it clear that ­dual-­layer burners
were imminent. Now, the first ­Blu-­ray Disc players are available, and I will bet you
anything they are not the last evolutionary step in optical storage technology. What
is remarkable is the price development for these types of products. The first DVD
players cost from $2000 to $3000 (luckily, having understood how these things
work, I avoided buying at that time). Today, you can pick up a DVD player right
next to light bulbs in any chain store for $49.99.
These are only a few examples, but I daresay you’ll find the same trend no mat‑
ter where you look. In fact, I’d say the adage of “the good old days” is simply not
true. Granted, I’m old enough to claim this myself now and then, usually referring
to ­hip-­hop music or the lack of exercise by kids these days, but for the most part,
things are actually much better today. We say that people were much more polite
“in the old days,” but I generally find customer service to have improved in almost
any industry I can think of. We also like to say that people used to lead much better
lives, with less stress and much more time to enjoy themselves, but this is not true.
Most people spent much more time earning enough money to just get by, or they
spent their time doing household chores that appliances do for us today. At no point
in history have people had wider choices of what to do with their lives and better
chances of fulfilling their dreams and needs than they do today.
Why? Because there is an ingrained human quest for what’s better, what’s
around the next bend. It is this drive that has led individuals to explore new con‑
tinents, to travel into space, and to go below the surface of the oceans. This same
drive has led to countless inventions, including four generations of DVD burners.
If you agree with this argumentation, there simply is no way of curbing improve‑
ment efforts in organizations; people will keep going forward no matter what.
In addition, many forces, both internal and external, have caused improve‑
ments to become a necessity in today’s marketplace:
Business Process Improvement  

• The performance level of most processes shows a tendency to decrease over

time unless forces are exerted to maintain it. This means that to maintain the
current standards, it is necessary to perform some degree of maintenance.
If we want to create improvement and renewal, this requires efforts beyond
pure maintenance, as shown in Figure 1.1.
• If an organization does not improve, you can be quite certain its competi‑
tors will. Should the unlikely scenario occur that neither an organization nor
its competitors improve, there are always other actors willing to enter the
market segment.
• Today’s customers are more demanding and, frankly, are spoiled. Sup‑
ply and the quality of the supply are ever increasing, which in turn causes
expectations to rise dramatically. If it is not possible to exceed expectations,
which is the ideal situation, they at least have to be met. If they are not met,
you are guaranteed to lose the customer.

Generally, this means that what was satisfactory a few years ago barely passes
today, and quite certainly will soon be below expectations. It is therefore irrelevant
to discuss whether we have to improve; the question, rather, is how much do we
improve and how quickly do we do it.

Performance level




No improvement effort


Figure 1.1  Without maintenance and improvement, the performance level decreases.
  Chapter One

There is, of course, no definitive answer to this, but a rather general one
is that continuous improvement, combined with occasional breakthroughs, is
needed. In this respect, it is worth pointing out that experience indicates that
an organization that emphasizes continuous improvement usually also pos‑
sesses the creativity and attitude toward improvement necessary to create an
occasional breakthrough (Hayes and Wheelwright, 1984). The opposite, trying
exclusively to achieve a breakthrough and neglecting the continuous improve‑
ment activity, might prove difficult in creating a breakthrough when one is
After observing a large number of improvement projects in very different
types of organizations, I have found that people differ greatly when it comes to
innate improvement skills. Some people, albeit only a few, are born pioneers or
inventors or architects of ­large-­scale change, and some organizations seem to be
natural champions at continuous improvement. Sadly, although all of us share
humankind’s drive toward progress, many of us have discovered that improvement
is a skill that must be learned and practiced. I have also realized how much easier
it is to exert this skill when it’s linked to something specific, such as a ­well-­defined
improvement project or the use of an improvement tool, where certain steps that
need to be executed are outlined. Regardless of how you organize the improvement
efforts and which tools you use, business processes are a good starting point for
improvement work. Together, these observations converge into the main idea of
this book, giving readers a ­well-­stocked toolbox for use in improving their organi‑
zations’ business processes.

1.2  A Business Process Improvement Framework

Although I argue that a drive toward improvement is ingrained in humankind, I

am not saying that improvement work that is goal oriented, includes people, and
performs well in an organization is easy. There are a number of prerequisites
that need to be in place to succeed in this game. While the diagram in Figure 1.2
might not be 100 percent complete, I have tried to portray what I believe are
the most important elements an organization needs to understand or master to
perform well with its improvement efforts. In the spirit of the topic of the book,
this is an updated, hopefully improved, version of an improvement model pre‑
sented in the first edition. The old model outlined a sequence of activities that
naturally belonged to a comprehensive improvement process. Some of these
elements also prevail in the new version, but in the years since the first edition,
I’ve come to realize that there are a number of other elements that warrant men‑
tion in such a framework.
The elements of the business process improvement framework are the
Business Process Improvement  

• The organization’s stakeholders and its strategy, which are included because
they form the overall priorities of the organization. They define the critical
areas where the organization needs to excel, and, even if this is not always
the case, they should be the backdrop against which any improvement proj‑
ect ideas are assessed.
• A thorough understanding of the current business processes. I have already
argued that business processes are an ideal unit of analysis for improvement
work, and understanding how a process is currently performed is an invalu‑
able platform for improving it.
• Performance measurement, which plays an integral part in improvement
work. It helps assess which business processes should be improved, it gives
insight into the current execution of the process, and it allows tracking of
the improvement effects created.
• A business process improvement road map, which is a type of game plan for
the improvement efforts. Such a road map will typically outline which busi‑
ness processes need to be improved over the long term, and it will contain
details about more immediate projects.
• An organizational structure with skills, incentives, and attitudes that foment
improvement. This is in itself a vast topic and an area of critical importance
for an organization that has high improvement ambitions. If the ­employees


Strategy measurement

Improvement Organization and skills

road map for improvement

Business process

Figure 1.2  Business process improvement framework.

  Chapter One

don’t have basic values that support continuous improvement, or there

are no incentives to invest extra effort in improvements, or no one has the
responsibility of initiating improvements, the effects will be dismal.
• The business process improvement toolbox, which this book is ultimately
about. Time and time again, I have seen that effective, ­easy-­to-use improve‑
ment tools help set a direction, guide improvement projects on their way,
and overcome a sense of not knowing how to get started with improvement

The framework combines elements that determine the direction of the improve‑
ment efforts, provide a platform of current business understanding, and give tools
and techniques for running improvement projects. The next chapters of the book
deal with the elements in the framework, and the toolbox itself is presented over
several chapters.
What this version of the improvement framework fails to illustrate is the cycli‑
cal nature of improvement work. Improvement projects often progress through
phases of assessing the current situation, creating and planning improvements,
implementing the changes, and measuring whether the improvement targets were
reached. If the targets were not reached, the process is often repeated until they
have been achieved. This cycle applies to both single improvement projects and the
organization as a whole, and the unique strength that comes with such a virtuous
circle is this stubborn attitude of never being content. When this culture settles and
prevails throughout the organization, there are no limits to how far it can go.
Those familiar with the work of W. Edwards Deming will see that such a cyclic
idea is based on the principles of the Deming wheel (Deming, 1986). The Deming
wheel—or the cycle of plan, do, check, act—is shown in Figure 1.3. This control
loop illustrates a general approach for conducting continuous improvement. The
activities of the four phases are as follows:

• Plan—In this phase, the problem is analyzed and activities to remedy it are
planned, that is, improvement planning.
• Do—In this phase, the activities planned in the previous phase are carried
out. The purpose is primarily to experiment with the solution.
• Check—In this phase, measurements are made to evaluate whether the
activities had any effect on the problem, that is, a performance evaluation.
• Act—Finally, in this phase, the process is modified according to the activities
that were confirmed to give results. After performing the four phases, you
will have either an improved process or a starting point for a new cycle.
Business Process Improvement  

Act Plan

Check Do

Figure 1.3  The Deming wheel.

The purpose of the Deming wheel, besides describing a systematic approach to

improvement, is that the wheel should continuously be in motion by performing
this process. This is one of the main messages of this book and one that will be
repeated several times. The next chapter deals with the organization’s stakeholders
and its strategy, which is one area where this relentless pursuit of improvement
must be emphasized.
Understanding the Organization’s
Stakeholders and Strategic

ou have not misread the title of this book; it is about business process
improvement. Although the book is not about strategic planning, the strat‑
egy of the organization strongly impacts its improvement efforts. This
chapter explains this influence and presents a couple of useful techniques at this
preimprovement stage.

2.1  Stakeholder Expectations of Improvement

There are many ways of presenting improvement ideas and proposals for improve‑
ment projects in an organization. In the 1970s and perhaps partway into the 1980s,
quality circles were a popular approach for bringing people together to discuss and
develop improvement ideas. In many organizations, mailboxes (physical or elec‑
tronic) have been used to collect ideas and suggestions from employees. Likewise,
feedback from customers can be collected more or less systematically through
forms in stores or offices, through an e-mail service, or through a discussion Web
page. Also, managers for each area or level of the organization can be responsible
for recommending a number of improvement ideas per year. In most cases, several
mechanisms will be in play.
The message I want to convey is this: While a richness of improvement ideas
and enthusiasm in the organization for creating ideas is extremely important, a bal‑
ance must be struck where ideas sanctioned for further investigation or implemen‑
tation are coordinated. An organization, no matter its size, will have only so many
free resources—that is, time, people, and money—that can be invested in carrying
out improvement projects. Keeping the operations going will normally take prior‑
ity, leaving improvement work as an extra effort that is often sacrificed when other
tasks require attention. To ensure that the return on the improvement investments

10  Chapter Two

is maximized, projects should address areas of the business that will make a differ‑
ence to the stakeholders.
Following this logic, an important platform for creating the improvement road
map, mentioned in the previous chapter, is a solid understanding of who the orga‑
nization’s stakeholders are, what expectations they hold, what they are satisfied
with, and in which areas they would like to see the organization improve. In other
words, a stakeholder analysis, a process that now seems to be fairly well known and
widespread in most industrial sectors, should be carried out. Stakeholder analysis
can easily be defined as a tool in strategic planning, but for improvement purposes,
it also serves other useful functions:

• When conducted at a higher level, that is, for the entire organization or busi‑
ness unit, it creates an overall understanding of the organization’s position
on the competitive map of the sector it operates in. This allows top man‑
agement to better understand the direction it should follow to maintain or
strengthen this position, thereby providing a basis for prioritizing improve‑
ment projects.
• Performing a stakeholder analysis once or twice per year can in itself gen‑
erate improvement ideas. Depending on the level of detail of the analysis,
discussions about, or even with, key stakeholders can take place regarding
their satisfaction with the organization and what they would like to see done
• Even in individual improvement projects, a ­lower-­level stakeholder analy‑
sis can be useful, especially in the beginning of a project. Consider a small
group of health club employees who have been given the task of reversing
the decreasing turnover at the ­in-­house juice bar over the last few months. A
limited stakeholder analysis for the juice bar would quickly give the group
useful insight into members’ refreshment preferences, how members spend
their time in the club, and the incentives that the employees have for offer‑
ing good service. It would most likely have to be followed up by other
analysis tools, but it would still be a good starting point.

You may have noticed that I use the word stakeholder with the expectation that
readers already know what it means. I hope this is justified, and I know that many
fellow authors take the same position, as the term has become highly entrenched in
management literature. Over the years, I have, however, tried to figure out where
the term really comes from, and I am not sure I have succeeded yet. Wikipedia
(2006) is usually reliable, and the explanation found there links the word to a tradi‑
tional usage from law and gambling: “a third party who temporarily holds money
or property while its owner is still being determined.” This might very well be true,
and it might precede a theory I heard recently, related to the American pioneers
Understanding the Organization’s Stakeholders and Strategic Direction  11

exploring the western U.S. territories during the 1800s. Having enjoyed movies
and TV series like Far and Away, Centennial, and How the West Was Won over the
years, and lately Into the West, I find this theory fascinating, perhaps even credible:
The pioneers went west to find a new life, most of them seeking land where they
could establish homes and farms. Land was claimed by planting a flag to mark a
plot until a formal deed was issued. The flag was attached to a stake. Thus, accord‑
ing to this view, a stakeholder was a person setting out to claim land.
I should also mention that stakeholder analysis is not one unified, ­generally
­agreed-upon tool with certain steps to be performed. It is more of a commonsense
exercise aimed at generating a higher awareness of the various stakeholders who
have a vested interest in the organization and, at a minimum, the needs, require‑
ments, and expectations these stakeholders have. Over the years, I have extensively
studied and worked with stakeholder analysis in many enterprises, and the fol‑
lowing seems to me like a very useful and workable approach (more details about
these steps follow after the list):

1. Identify the organization’s stakeholders, normally through a brainstorming

2. Group or classify the stakeholders according to some criteria to determine
which are the most important.
3. Identify the needs and expectations held by the stakeholders, an exercise
for which the Kano model has proved suitable.
4. Try to anticipate the likely behavior of the stakeholders toward the
5. With the insight generated through these steps, develop strategies for how
to handle the stakeholders.

The only feasible way I have encountered for identifying stakeholders is to conduct
a good ­old-­fashioned brainstorming (see section 10.1 for more details about brain‑
storming). If a small group of people from different areas of the organization gets
together, it normally comes up with the main stakeholders in an hour or two. Who‑
ever undertakes this task of identifying the stakeholders should remember that there
are some obvious and not so obvious stakeholders out there. The obvious ones are
the customers, suppliers, competitors, employees, and perhaps the owners. Some
that are easy to overlook include the partners, the media, government/authorities at
various levels and of different types, the local community, and a number of pressure
groups. Figure 2.1 illustrates how some of these stakeholders surround the organi‑
zation and constitute important parts of the organization’s environment.
You can find several models for classifying stakeholders; see, for example,
D’Herbemont and Cesar (1998). One that I find useful assesses each stakeholder’s
12  Chapter Two

Authorities Owners Management

Suppliers Customers

institutions Competition

Employees Environment partners

Figure 2.1  The stakeholder model.

potential for impacting and cooperating with the organization, as shown in Fig‑
ure 2.2 (Savage, Nix, Whitehead, and Blair, 1991).
The total set of stakeholders will be scattered across this matrix. In general, the
farther to the left and top of the matrix, that is, toward “high” on both dimensions,
the more important the stakeholders typically will be for the organization.
All these stakeholders hold certain expectations toward the organization. Iden‑
tifying these expectations and translating them into more tangible quantities can
probably be done by following different avenues. The organization most likely
already possesses a fairly high level of knowledge about some stakeholders’
requirements, for example, owners or customers. For others, it might be difficult to
tell what they do or do not want. For pressure groups advocating equal opportuni‑
ties, the media in general, or public authorities, the expectations are probably much
less clearly understood, and hidden agendas may even exist. If you map all of the
probable and improbable expectations your stakeholders harbor, you will undoubt‑
edly discover that these are plentiful, not always coherent, and very wide ranging.
To bring some order to these expectations and be able to differentiate between the
important and the not-so-important ones, a very useful diagram called the Kano
model is at your fingertips (Kano, Seraku, Takahashi, and Tsuji, 1984). As you can

Potential to impact the organization

High Low
Potential for cooperation
with the organization


Figure 2.2  Stakeholder classification matrix.

Understanding the Organization’s Stakeholders and Strategic Direction  13

see in Figure 2.3, the model is nothing more than an ­awareness-­creating diagram
showing that there are different types and levels of stakeholder requirements.
The straight diagonal line in the figure portrays the clearly expressed require‑
ments of the stakeholder. Generally, these are the only demands the stakeholder
will describe if asked about her or his wishes. If the stakeholder is a major share‑
holder of the organization, he or she could express the requirements that the return
on investment be a minimum of 7.2 percent, that he or she be granted a seat on the
board, and so on.
In addition, there also exists a set of requirements that are so basic that they
are not even expressed, as indicated by the lower curve. For the shareholder, these
could be that the organization does not go broke and lose its capital, that it does
not get involved in criminal or otherwise unethical activities that could harm the
shareholders, that business is conducted according to general rules and customs,
and so on.
Together, these two sets of requirements constitute a complete set of demands
imposed by the stakeholder toward the organization. The satisfaction depends on
how well both sets of requirements are met. It will be of no help if the investment
returns 10 percent but the shareholder is arrested for criminal activities in which

Loves the
organization’s output

Performance quality:

Excitement quality:
Not known
Not expressed
Trend sensitive
Stakeholder specific
Degree of achievement

Basic quality:
Often forgotten

Hates the
organization’s output

Figure 2.3  The Kano model and the three types of stakeholder requirements.
14  Chapter Two

he has involved the company. In other words, satisfying expressed requirements

cannot rectify shortcomings in the basic demands. On the other hand, satisfying
every single one of the basic requirements will not lead to complete satisfaction
unless the expressed requirements have also been fulfilled. At best, this will elimi‑
nate dissatisfaction. The danger is that the stakeholder takes it for granted that the
organization is aware of the basic requirements, but this might not be the case.
Such silent assumptions are one of the main focuses when clarifying requirements
in the stakeholder analysis.
If these two sets of requirements are defined and satisfied, the foundation
for satisfaction should be firmly established. To further enhance satisfaction,
and even delight the stakeholder, we can look to a third set of requirements.
Requirements is not really the correct word, as these are conditions not
expressed by the stakeholder, and often the stakeholder itself is not even aware
of these needs. For the shareholder, this could include having the organization
make all arrangements for transportation to general assemblies, free access to
the organization’s products or services, a special Web site for the shareholders
that presents updated information of interest to them, and so on. If both the
basic and expressed requirements have been satisfied, the fulfillment of such
extra “requirements” can create true delight. These are often the little extras
required to ensure loyalty and access to the best stakeholders of all kinds. Note,
though, that if such extra requirements are delivered once or a few times, they
are often added to the expressed or even basic requirements that must be ful‑
filled to avoid dissatisfaction.


The financial department of a large company produced weekly reports

for the sales department that contained the credit status of the custom-­
ers. As he always did, the responsible accountant put his heart into the
accuracy of the report, even if this often took quite some time and could
sometimes delay the report. The sales department was not depending on
minute accuracy, but was depending on having the report on time to be
able to issue order confirmations at the agreed time. It assumed that the
financial department knew about this and did nothing to attempt to com-­
municate it to the accountants. Instead, the sales department continued to
be annoyed by the delayed reports, without appreciating the accuracy, and
nothing improved.

When trying to anticipate how the stakeholders will act toward the orga‑
nization, it can be useful to take the stakeholder classification matrix one step
further. Figure 2.4 shows the same matrix as in Figure 2.2 but now with labels
for each field.
Understanding the Organization’s Stakeholders and Strategic Direction  15

Potential to impact the organization

High Low

Potential for cooperation

with the organization


Nonsupportive Marginal

Figure 2.4  Stakeholder strategies.

These labels do not represent exact predictions of behavior for each group, but
they indicate from experience how each group might act. For each group, there are
also some recommendations about suitable strategies to follow:

• Supportive: should be involved in relevant discussions and decisions

• Marginal: should simply be monitored
• Nonsupportive: use a defensive strategy and minimize the dependency on
the stakeholder
• Mixed blessing: best handled through cooperation

To recap, the outcome of the stakeholder analysis should constitute a very good
start for the ensuing improvement efforts.


A large bank that was a leader in developing Internet banking services

had pushed this approach to a point where customers had virtually no
reason to see a bank employee; everything could be done electronically
or, in the worst case, by phone. Two years after the project was deemed
complete, figures showed that the bank was losing certain customers,
especially elderly customers and customers with ­ above-­average income
or investment funds. Since these were also the most profitable customers,
this posed a serious challenge to the bank, and it decided to undertake a
stakeholder analysis.
A large number of stakeholders were identified, including authorities,
investment funds, financial consultants, employees, trade unions, and so on.
These stakeholders were subsequently analyzed according to ­importance
16  Chapter Two

for the bank, and since the problem addressed was loss of customers, only
those found relevant in this matter were included in the process: customers
in different segments (25 years and below, 25–50 years, 50–65 years, 65
years and above, and investors), financial consultants (who often advised
customers about which bank to choose), and trade unions (who brokered
banking deals on behalf of their members).
For each of these, the Kano principles were applied to identify needs
and expectations. To simplify, only selected expectations identified in the
analysis of the customers of age 65 and above are included here.

Performance Expectations
• Competitive interest rates
• Competitive fees and expenses
• Easy-to-understand information from the bank

Basic Expectations
• Easy access to a customer representative when required
• Active counseling regarding major financial decisions
• The bank has my interests in mind

Excitement Expectations
• Branch office open 24/7
• Invitations to annual events sponsored/hosted by the bank
• Birthday attention
• Internet banking available

From this analysis, the team was surprised to find that this customer seg-­
ment took it for granted that there would be a physical branch office they
could go to and where they would find people who could give them advice.
Internet banking was simply a possible extra service for some of these
customers. For the customers with more funds to invest, the picture was
almost the same: they relied on active and close contact with a representa-­
tive from the bank to aid their financial decisions.
After seriously considering continuing the policy to shift customers to
electronic channels and accepting the loss of those who felt uncomfortable
with this, the bank decided to change its strategy. A new brand was estab-­
lished to handle customers with greater need for personal contact. This
group started with five pilot offices in which the branch office was rebuilt
and tailored to the needs of these customers. The approach was a huge
success, with these offices winning back many more customers than had
been lost the last few years, obviously due to filling a gap that few banks
Understanding the Organization’s Stakeholders and Strategic Direction  17

concentrated on. Today, this bank has by far the highest market share in this
customer segment and operates close to 50 such branch offices.

2.2  Reviewing the Organization’s Strategy

I have already emphasized that I am no expert in strategic planning, and because

this is not a book about strategic planning, this section will be short. My main
objective with this section is to point out that the defined strategy of an organiza‑
tion is another ­high-­level element that contributes to setting the direction for the
business process improvement efforts. The strategy, or rather the broad awareness
of the strategy throughout the organization, should make sure that improvement
projects are not allowed to go against it, as in the following example, or take off in
any direction, in complete anarchy.


A ­medium-­sized European airline had traditionally catered primarily to the

­business-­travel segment, offering a wide range of destinations and depar-­
tures with high service levels. During the 1990s, the ­low-­cost competitors
entered the market, typically first targeting the tourist market, but gradually
also capturing increasing ­business-­travel market shares. With tougher price
competition, the pressure for cost savings mounted, and several projects
were initiated. Several changes were implemented over a couple of years:

• The departure frequency to many destinations was significantly

• Several ­business-­class lounges were closed
• Gate staffing was reduced, leading to a situation where business
travelers could no longer go directly to the departure gate and
change their booking there
• Newspapers and meals included in the fare were eliminated or dras-­
tically reduced

As a whole, the airline became much more like its ­low-­cost competitors,
but still had somewhat higher prices. For the traditional business travel-­
ers, this made the airline a much less attractive supplier. There was a clear
position in the top management that the business segment should still be a
prioritized customer group, especially since it represented the most profit-­
able customers. In reality, through these ­so-­called improvement projects,
many passengers were driven to choose other airlines that still offered
­business-­class service at traditional prices.
18  Chapter Two

Just as for the stakeholders, reviewing the organization’s defined strategic direction
can be a worthwhile exercise, both to make sure the improvement projects are in line
with it and to uncover areas where new improvement projects should be considered.
Before moving on to a small selection of tools and techniques that can be used
during strategy reviews, I should modify myself slightly. I might have given the
impression that all improvement initiatives should be carefully assessed in terms
of strategic fit before moving forward. Although I generally discourage too much
freedom in starting projects locally without any coordination with other projects
and priorities, there are exceptions. There are many examples of outsider or “ren‑
egade” projects that have altered or redefined the strategy of the organization. If
you never allow any projects that fall outside the current strategy, you might very
well lose such opportunities. This is very much in line with the famous work by
Mintzberg (1994) about emergent strategies, a pattern of action that develops over
time in an organization in the absence of a specific mission and goals, or despite a
mission and goals. It is naïve to believe that all improvement efforts will be aligned
with strategy, and perhaps they should not, either. Some of them might be break‑
through projects that ultimately shape a new strategy or create large gains.
I will conclude by presenting a few tools and techniques that can be used
when formulating and reviewing strategies. They are all established approaches,
and some or all are probably well known to you already. If so, let this be a quick
repetition of the following:

• SWOT analysis
• Competitive forces analysis
• Strategy map

2.3  SWOT Analysis

Identifying elements inside the organization and its surroundings within the per‑
spectives of strengths, weaknesses, opportunities, and threats (SWOT), the SWOT
analysis is probably one of the ­best-­known simple strategic techniques that exists.
Its purpose is simply to create an awareness of forces that will impact the orga‑
nization in the future. By understanding these forces and making them known
throughout the organization, better strategic decisions can be made and the whole
organization will be better prepared for future developments.
Conducting a SWOT analysis is technically simple. It mainly entails brain‑
storming within the four perspectives, preferably in a group with different dis‑
ciplines and areas of the organization represented. “Technically simple” was
deliberately used since I know from experience that performing a good SWOT
analysis can be difficult, especially in terms of coming up with those elements that
are relevant. This takes training.
Understanding the Organization’s Stakeholders and Strategic Direction  19

Stepwise, the approach is as follows:

1. Compose a team to undertake the analysis, drawing on different levels

of competence within the organization. Consider supplementing the team
with external representatives.
2. For each of the four analysis perspectives, brainstorm issues that seem
3. Compile the analysis results, using a simple table or matrix.
4. Discuss which of the issues identified is believed to have the strongest
influence on the organization’s future.
5. Outline strategies or actions to deal with the findings.


A small town has seen its population drop steadily over the last few years,
and its administration considers developing a new, fairly large residential
area to counter this trend. The likely location of the area is adjacent to
existing residential areas, where there is already an existing infrastructure
in place, but it will require extending two current roads and converting parts
of a park/green area used by many families in the town for recreational
activities. To avoid charging ahead with the plans before truly understand-­
ing the possible consequences, the town council has asked that a SWOT
analysis be performed, both for the town itself as a home to families and
for the specific development plan.
The results from the analysis are summarized in Figure 2.5.

Strengths Weaknesses
Slow-paced town with high quality Few employment opportunities
of life Long distance to larger town/city
House prices lower than larger centers
towns and cities in the area The new residential area will ruin a
The new residential area will highly popular recreational area
utilize the existing infrastructure

Opportunities Threats

Rumors about about a new road Aging population makes the town
close by less attractive for young families
New inhabitants could lead to more Expected protests from
industry environmental groups against the
development plan

Figure 2.5  SWOT analysis results.

20  Chapter Two

The conclusion from the group was that additional employment oppor-­
tunities/business activity in the town would be important to develop to
make sure that the new residential area became attractive. Furthermore,
the possibility of replacement green areas to appease protests against the
plan should be looked into.

2.4  Competitive Forces Analysis

The SWOT analysis is a very broad analysis intended to identify any trends or
developments that could affect the organization. A more specific analysis into the
competitive position of the organization and its likely development can also be
useful as a precursor to ensuing improvement projects. Porter developed one rec‑
ognized approach during the 1980s (for more information, see Porter’s books from
1980, 1985, and 1990 on competitiveness). The analysis is based on the competi‑
tive forces that influence an organization, as shown in Figure 2.6.
Briefly explained, these five forces are:

• Competitors that could claim market shares from the organization and other
effects from competition, for example, price cuts or increased service level
• Buyers’ bargaining power to obtain better terms or switch suppliers, thus
impacting the organization, as well as possible backward integration by cus‑
tomers to threaten the position of the organization
• Suppliers’ bargaining power to increase prices or worsen terms for the orga‑
nization, as well as possible forward integration by suppliers to threaten the
position of the organization

Threat of new entrants

Bargaining Bargaining
Rivalry among
power of power of
existing competitors
suppliers buyers

Threat of substitute
products or services

Figure 2.6  Five competitive forces.

Understanding the Organization’s Stakeholders and Strategic Direction  21

• Threats of new entrants that can become competitors, depending on entry

barriers, switching costs for customers, and so on
• Threats of substitutes that could displace the organization’s products or ser‑
vices, depending on the match with current offerings, buyer switching pro‑
pensity and costs, and so on

Some have argued that one force or dimension is missing in the model, namely,
the power and influence of other stakeholders in the environment. These could be
actors like regulatory bodies that often decide standards or requirements, banks and
other creditors, shareholders, and so on. Whether five or six forces are included,
the purpose of the analysis is to create an understanding inside the organization of
possible threats that could materialize in the future and weaken the competitive
The steps in the analysis are very similar to that of the SWOT analysis:

1. Compose a team to undertake the analysis, drawing on different levels

of competence within the organization. Consider supplementing the team
with external representatives.
2. For each category, brainstorm issues that seem relevant.
3. Discuss which of the issues identified is believed to have the strongest
influence on the organization ahead.
4. Outline strategies or actions to deal with the findings.


A supplier of automotive parts had over the years built up a highly suc-­
cessful customer portfolio encompassing about half of the European car
brands. The parts were brake components, made from brass, that were
assembled into the brake system together with parts from many other sup-­
pliers. Recently, rumors were going around that other suppliers of simi-­
lar components were experimenting with composite materials. The use of
composite materials would significantly reduce the weight of the parts, an
important consideration for all carmakers. To decide whether to pursue
such a strategy, a competitive analysis was performed.
A small team was put together, and it decided to follow the five forces
model. A selection of the more significant forces identified is presented.

• Competitors that in general offer better terms to customers (con-­
tracts up for tender every third year)
22  Chapter Two

• Competitors that develop composite parts, thus offering lower

weight (price consequences unknown)
• Competitors that move toward becoming more comprehensive brake
system suppliers, including the parts supplied by the company

Buyers’ Bargaining Power

• Customers becoming even more aggressive in negotiations,
demanding further price reductions during the life of the contract
(the higher the risk, the more standard the product)
• Customers awarding the company’s contract to competitors

Suppliers’ Bargaining Power

• Few issues; most suppliers were local and depended on the suc-­
cess of the company

New Entrants
• If the industry moves toward composite brake materials, suppliers
with expertise in this material could target this market segment

• Ongoing development work in many places in the world could lead
to the successful introduction of electronic brake systems that could
replace ­fluid-­based systems

With all these forces on the table, the company realized that it had come
to a turning point. Carmakers were constantly pressured to reduce emis-­
sions, and weight was playing an ­ever-­more important role, irrespective of
the propulsion technology used. The likelihood of either plastic compos-­
ites or electronic systems replacing brass components in the near future
was high, and there seemed no alternative to going down that road. As a
consequence, a major improvement project was launched, with the aim of
developing the first approved composite solution.

2.5  Strategy Map

Developed by Kaplan and Norton (1996), the strategy map was briefly discussed
in the groundbreaking book about the balanced scorecard. A more comprehensive
treatment is found in the book dedicated to strategy maps (Kaplan and Norton,
2004). The concept of a strategy map is to force the organization to put words to
both its end strategic goals and the road for getting there. Making use of the per‑
Understanding the Organization’s Stakeholders and Strategic Direction  23

spectives defined in the balanced scorecard concept—financial, customer, internal

business processes, and learning and growth—an organization charting its strate‑
gic course should define milestones that must be achieved along the way to reach
the final objectives. A generic strategy map is shown in Figure 2.7.
The steps for developing a strategy map are as follows:

1. Compose a team to undertake the analysis, drawing on different levels

of competence within the organization. Consider supplementing the team
with external representatives.
2. Start by defining the overall, ­long-­term strategic objectives.
3. Go down one level in the map and identify which strategic goals must be
achieved in the financial perspective to reach the overall goals.
4. Continue this way through the other perspectives, always asking what
must be achieved to fulfill the overall goals and the goals of the above
5. Compile the strategy map based on the strategic objectives identified at
each level, linking these together to form strategic paths.

Strategic Strategic Strategic Strategic

subject subject subject subject

Strategic objectives

Figure 2.7  Generic strategy map.

24  Chapter Two

A completed strategy map should be helpful to the organization, both to clarify

for its members the direction to pursue and to derive more specific improvement
efforts required.


A shipping company was experiencing trouble regarding its environmental

performance. Large customers required reports about different environ-­
mental issues (fuel consumption, emissions, bilge water discharges, and so
on), and they complained that the company didn’t have a clear enough pro-­
file in this area. There had even been warnings that unless things improved
in future contract negotiations, contracts might be transferred to other sup-­
pliers with a more consistent approach to environmental management.
This prompted the company to review its entire strategy, which up
to now had very much been focused on low costs and providing reliable
logistics services. With the aid of an external consultant, a ­small-­business
development team sat down to design a strategy map that would espe-­
cially address the environmental aspects of the company’s development.
The strategy map is shown in Figure 2.8.

environmental performance

Premium price for Environmental performance

Financial environmental performance at no extra cost for the company

Customer Customer proud of supplier

Internal processes Consistent environmental New technology


Learning and Broad environmental Technology

competence training development program

Figure 2.8  Strategy map for the shipping company.

Understanding the Organization’s Stakeholders and Strategic Direction  25

As a result of the exercise, two specific improvement projects were

launched: one to train almost every single employee in environmental
issues and one to develop new environmental technology to help reduce
fuel consumption and emissions from the ships.

As the examples have hopefully demonstrated, the organization’s strat‑

egy strongly impacts which improvement efforts are suitable at any given time.
Although strategic planning might seem foreign to people working primarily with
quality improvement, having an eye on this part of the organization is therefore
useful. The next chapter will discuss how to ensure that you have a good under‑
standing of the organization’s current business processes before embarking on
more specific improvement projects.
Understanding Your Current
Business Processes

uring a lecture about business processes, a student asked me whether a
company consists of just business processes or also has the traditional
departments. At that time I had to answer: “Both, of course!” (Ten
years later, for some organizations, the truth would in fact be that they consist
only of business processes and have abandoned the old departments.) Viewing
organizations as consisting of processes offers many advantages, and such an
approach is also historically correct. This entire book deals with how to con‑
duct improvements and is based on improvement efforts directed at business
processes. This chapter deals with the core concepts of business processes and
how to understand them.

3.1  From Processes to Departments to Business Processes

At some point in history, people began coordinating their actions to be able

to perform more extensive or more complex tasks than what could be accom‑
plished alone. We must assume that such coordination, the first type of what
might be termed organizations or enterprises, was initiated by a demand. For
example, when a farmer hired helpers for tasks such as plowing, sowing, and
reaping, it was because the work was too extensive for the farmer to manage
on his own. The focus was on satisfying a demand for food by performing the
necessary processes. In other words, the actions were process oriented. The
workforce was not split into specialists in plowing, sowing, and harvesting;
rather, the hired help represented a pool of hands that carried out the necessary
Soon the extent and complexity of the tasks expanded further, and the num‑
ber of people in different coordinated enterprises increased. For as long as
enterprises with more than a few employees have been in existence, they have

28  Chapter Three

probably been organized into departments. It was no longer feasible to main‑

tain a workforce where everybody performed each task. The tasks became so
complex that the individual worker had to specialize. Therefore, the logical
step was to form departments consisting of individuals with similar areas of
expertise. This expanded into a tradition with a solid foothold in all types of
organizations—commercial, public, and nonprofit. Until a few years ago, this
way of organizing was highly dominant and had completely replaced the origi‑
nal way of aligning the organization more directly according to the work to be
performed. Organizing people and work into departments provided, and still
provides, some benefits:

• People were allowed to specialize within their field of expertise, thus devel‑
oping a highly refined set of skills
• Such departments could often more easily attract other specialists of the
same training
• Costs from centralizing various functions (for example, finance, personnel,
maintenance) were lowered
• The workplace was more secure; employees knew where they belonged and
which tasks they were supposed to perform
• The organizational structure was more clearly defined and could easily be
drawn and presented

Thus, we can safely claim that at least some modern organizations consist of both
functional departments and horizontal business processes (I say “at least” since


Purchasing Manufacturing Engineering Finance

Dept. Dept. Dept. Dept.

Inbound logistics

Processes Order handling

Product development

Figure 3.1  The contradiction between vertical departments and horizontal processes.
Understanding Your Current Business Processes  29

some organizations, as already mentioned, have completely done away with depart‑
ments and have organized their employees inside processes). Figure 3.1 depicts a
typical organization, with its vertical departments and horizontal processes that run
through these departments (Andersen and Pettersen, 1996). Lately it has become
obvious that this contradiction between mode of organization and tasks has created
several problems.
As soon as people are established within a square in an organizational
chart with departments like those shown in Figure 3.1, it often seems as if the
lines of this square become solid boundaries within which these individuals
must remain. Communication across the borders is limited, and members of a
department will perform only those tasks for which their department is respon‑
sible. Processes that run across departments and require contributions from
many organizational units naturally suffer. Employees inside a department are
told where their priorities should lie, and many of them end up focusing on sat‑
isfying their superiors. And this, as Jack Welch is rumored to have said, results
in a ­less-­than-desired focus on the customer: “with their face to the boss and
their ass to the customer.”
Consider a process such as the one depicted in Figure 3.2. It’s a simple pro‑
cess that starts with a salesperson receiving an order from a customer and ends

Sales dept. Order processing dept. Accounting dept.

1. Salesperson takes order 2. Order is verified 3. Credit is checked

and sends to regional and checked for
headquarters accuracy

Stock control

4. Stock availability
Logistics & distribution Production dept. is checked

7. Order is shipped 6. Order is assembled 5. Order confirmation and

details are sent to factory

Figure 3.2  A simple business process.

30  Chapter Three

when the order has been shipped (or ultimately when the products have reached the
customer safely). If executed in a normally performing organization, this process
should take anywhere from a few hours to a few days.
When this process is played out across a number of typical departments, it
looks like the version shown in Figure 3.3. All of a sudden, the process has to
traverse several departmental boundaries. For each such crossing, it somehow
ends up in a priority queue at the next department. Research has shown that
a good rule of thumb is that each such boundary crossing costs one week in
lost time. In this process there are five crossings, easily adding five weeks of
execution time—much longer than what it could have been in a streamlined
Each department seeks to maximize its influence and authority while optimiz‑
ing its performance level. The result is usually that the whole is far from being
more than the sum of the individual parts, and in the worst case, far less. Each
department suboptimizes within its area of responsibility, which in turn leads to
conflicting objectives and actions between departments. The total performance
level of the organization is naturally in line with this (Rolstadås, 1995).

Sales dept. Order processing dept. Accounting dept.

1. Salesperson takes order 2. Order is verified 3. Credit is checked

and sends to regional and checked for
headquarters accuracy

Stock control

4. Stock availability
Logistics & distribution Production dept. is checked

7. Order is shipped 6. Order is assembled 5. Order confirmation and

details are sent to factory

Figure 3.3  The same simple business process, now with department boundaries.
Understanding Your Current Business Processes  31


A company that manufactures elements for the building industry delivers

regular stock goods, standard goods that are not stored, and special vari-­
ants. One of the company’s main problems has been that the demand for
production capacity has been very unstable: sometimes it’s far above the
capacity limit, and other times it’s well below. In periods of high demand,
delivery times have dramatically increased, which is a problem in the
In the company’s monitoring of the sales department, a central perfor-­
mance measure was fulfillment of the sales quota for each period. This led
to a situation in which a salesperson, when the quota for a period had been
met, put all later orders “in the drawer” for the next period. This way, the
employee made sure he had a reserve that could help fill his next quota if
sales should be slow.
This, of course, contributed to the uneven demand for production
capacity and the ­sometimes-­prolonged delivery times. At the same time,
important information was withheld from the planning department that, in
turn, could have given better production plans. The salesperson’s opti-­
mization of his own performance level, as it was being measured by the
company, led to an overall performance level for the company that was far
from the best achievable.

These types of problems are the basis for the last decade’s change from view‑
ing the company as a number of departments to focusing on the business processes
being performed. Several issues make this a logical transition:

• Every process has a customer, internal or external, and focusing on the pro‑
cess ensures better focus on the customer
• The value creation with regard to the end product takes place in horizontal
• By defining process boundaries and the customers and suppliers of the
processes, better communication and ­well-­understood requirements can be
• By managing entire processes that run through many departments, rather
than managing individual departments, the risk of suboptimization is
• By appointing process owners, who are responsible for the processes, the
traditional fragmentation of responsibility often seen in a functional orga‑
nization is avoided
32  Chapter Three

• Managing processes provides a better foundation for controlling time and

• Repetitive business processes represent the best opportunity for the organi‑
zation to relentlessly polish performance by continually looking for ways
to improve them

Studies have also been undertaken that attempt to document the results achieved by
an orientation toward business processes in an organization. One interesting study
reported by McCormack in 2001 was based on a 32-question survey given to 100
international and U.S. companies to determine any relationship between business
process orientation (BPO) and the factors of:

• Business performance
• Interfunctional conflicts
• Interdepartmental connectedness
• Esprit de corps (team spirit)

The data were analyzed using correlation analysis and regression analysis, generat‑
ing the following findings:

• Companies with strong measures of BPO showed better overall business

• Companies with strong measures of BPO tended to have better esprit de
corps, better connectedness, and less interfunctional conflict

The results of the analysis had strong significance levels and represent strong
support for the benefits of business process orientation. What, then, is a business

3.2  Definition of a Business Process

One basic definition of a process is:

A logical series of related transactions that converts input to results or


To separate a company’s processes from any other forms of processes, the word
business has been added to form the term business process. A business process can
be defined in the following way (Ericsson Quality Institute, 1993, 22):
Understanding Your Current Business Processes  33

• A chain of logically connected, repetitive activities that

• utilizes the organization’s resources to
• refine an object (physical or mental)
• for the purpose of achieving specified and measurable results/products for
• internal or external customers.

In a similar vein, business process reengineering pioneers Hammer and Champy

(1993, 35) defined a business process as “a collection of activities that takes one
or more kinds of input and creates an output that is of value to the customer.” An
emerging view might be along the lines of, “A number of roles collaborating and
interacting to achieve a goal.”
Irrespective of which definition one subscribes to, most authors emphasize
the repetitiveness of business processes as an important characteristic. Although
I have been unable to locate the reference, I know at some time I read statistics
on the preparation time of a Big Mac. As far as I can recall, it had been reduced
by several hundred percent since its introduction. The reduction most likely is a
result of new technology in food preparation equipment and a careful tweaking of
every step of the process until it comes close to what is humanly possible. If the
process weren’t repeated so many times a day across the planet, there wouldn’t
have been such a tremendous platform for this tweaking. The fact that the process
will continue to be repeated for many years to come makes it worthwhile to invest
in further improvements, and all the repetitions provide a constant test bed for
But not only processes that are repeated to this extreme will benefit from a
careful design and ­ improvement-­oriented experimentation. Some years ago, we
laid down paving in our driveway using stones of two by four inches. Although we
don’t have a very big driveway, the stones amounted to about 14 tons. Subtracting
all the preparatory work and sealing afterward, simply carrying the stones to the
right spot, putting them down in the right pattern (we used gray and red stones),
and knocking them tightly into position took me about four days. In the beginning,
I simply grabbed a few stones, dropped them on the ground, and put them in their
places, believing this would be a few hours’ work. When I realized how slow the
process was, I started thinking how I could speed it up. I experimented with a few
different approaches—sometimes carrying enough stones for a couple of rows,
other times sorting the stones beforehand and distributing them along the edge.
Although this kept my mind going, it wasn’t exactly rocket science. To make it
more interesting, I started timing how long it took to lay down one row of stones.
Astonishingly, on day two of the job, I went from 21 to 12 minutes. Had I just
plodded on as I did in the beginning, the whole operation would probably have
taken two extra days. For this little “enterprise” the savings were significant, even
34  Chapter Three

though I worked only four days. Think about what a difference it could make for a
professional paving company if it could make similar optimizations.
Encouraged by this success, I have tried to apply the same principle at later
occasions, although (fortunately) I have not yet come across another task that took
four days of repeating the same process. And every time, it is obvious that the
effort bears fruit. Before painting 23 ­clip-­on window crossbars a few months ago,
I spent a few minutes considering how to best design the process. I figured out
which way would give me the best access, where to place the paint bucket to reach
it easily, and in which order to paint the different squares to avoid having to change
my grip too often. Tweaking the process further as I went along, I am sure I easily
saved myself three to four hours of painful work spent mostly on my knees on a
hard concrete garage floor.
Ultimately, such optimizations are a combination of a design that is well thought
through in advance of the process (sequence of activities, how to perform each step,
which equipment to use, and so on) and controlled experimentation and tweaking
of the process. In fact, processes that are carried out only a few times, perhaps even
only once, will benefit from advance planning. For more complex tasks, this is in
fact project planning, which is an established process. A word of warning is perhaps
also warranted. I once interviewed a quality assurance manager at a large construc‑
tion project. During the interview, he proudly showed me binder after binder of
procedures carefully outlined in flowcharts as well as spelled out in text (sometimes
in ­if-­then statements). If this didn’t make me wonder whether this guy had taken
things to the extreme, a remark toward the end of the interview did. He confessed
that keeping up this high level of professionalism in process design almost required
practicing outside of work. At home, he had several procedures for making fam‑
ily life more efficient, including one for “meal execution” and one for “bedtime
preparations.” I couldn’t help wondering whether he had a procedure for what was
to follow successful completion of the bedtime preparations!
Another main point in the definition of a business process is that it has a cus‑
tomer, either external or internal. Thus, almost all activities within a company can
be seen as a business process or part of a business process. What are the most cen‑
tral business processes, then?

3.3  Classification of Business Processes

There are a number of ways of classifying business processes. Many leading com‑
panies, in terms of process orientation, have conducted thorough analyses of their
own enterprises and compiled lists of their central processes. For example, both
Xerox and IBM have made lists of a different number of processes. These lists are
specially designed for the tasks carried out in these companies.
Understanding Your Current Business Processes  35

At the same time, different interest organizations have done a similar job, but
on a more general basis. The purpose for these organizations has been to create
lists that are general enough to be used by a large number of companies. Some con‑
tributors in this group are the International Benchmarking Clearinghouse (IBC),
the European Foundation for Quality Management (EFQM), and the Automotive
Industry Action Group (AIAG).
Modeling enterprises and defining general sets of business processes has come
to be a field of its own, occupying many academics. For example, researchers at the
University of Plymouth have defined a hierarchy of business processes that spans
five levels and is very extensive. The processes are divided into three main groups:
operate, manage, and support.
A somewhat simpler and more practically oriented approach can be found in
the work performed in the Norwegian TOPP (Productivity Program of the Tech‑
nology Industry) program, a research program on productivity managed from
NTNU (Norwegian University of Science and Technology) and its research partner
SINTEF in Trondheim. To create a foundation for developing methods for ­self-
­assessment and benchmarking, a framework of business processes was developed.
In line with the thinking of Porter’s value chain, the processes were divided into
primary and support processes:

• Primary processes are the central and ­value-­creating processes of the enter‑
prise. They run straight through the company, from receiving supplies from
vendors to activities on the customer side.

• Support processes are not ­value-­creating processes directly, but rather activ‑
ities needed to support the primary processes. They include activities like
financial and personnel management.

The symbolism of this view is shown in Figure 3.4. The primary processes rep‑
resent the core activities that ultimately generate income and allow a company to

Manage- Mainte-
ment nance

Purchasing Logistics Invoicing

Finances HR

Figure 3.4  Primary and support processes.

36  Chapter Three

survive. The importance of the support processes is, however, evident in the fact
that the primary processes could not be executed without them. Together they form
a necessary symbiosis.
To clarify, or perhaps complicate, things further, some of the support processes
can been extracted into a new class, termed development processes. These are
processes that are supposed to take the value chain and its primary and support
processes to a higher level of performance. Examples are product development, per‑
sonnel training, and supplier development. In the ENAPS (European Network for
Advanced Performance Studies) project, funded by the European Commission, a
generic set of processes was developed on the basis of these classes (see Figure 3.5,
ENAPS, 1997). Here, the primary processes were named “business processes” and
split into four main processes, each with subprocesses. The classes of secondary
and development processes were grouped under the heading “secondary processes”
and renamed “support” and “evolution.”
There have been other attempts at classifying processes in alternative ways,
and although these are not included here, they are probably quite accurate as well.
The main point of these exercises is to create a ­high-­level understanding of which
types of business processes exist in an organization.

ENAPS Generic Framework

Business processes Secondary processes

Product development Support

• Product research • Financial management
• Product engineering and design • Human resource management
• Process engineering and design • Information management
• Co-engineering • Maintenance
• Internal control of health,
Obtaining customer environment, and safety
• Market development Evolution
• Marketing and sales • Continuous business process
• Tendering improvement
• Product research
Order fulfillment
• Production technology research
• Procurement and inbound logistics
• Human resource development
• Production planning & control
• Supplier base development
• Manufacturing and assembly
• Development of external relations
• Distribution and outbound logistics
• Strategic planning
• Order processing

Customer service
• After-sales service
• Product take-back

Figure 3.5  Business processes in ENAPS.

Understanding Your Current Business Processes  37

3.4  Identifying the Organization’s Business Processes

Before documenting one or more business processes, a prerequisite is that the

business processes of the enterprise have been identified. This can at times be
rather difficult, as it is rarely obvious which processes are undertaken by different
departments in a functional organization. Two complementary routes to such an
understanding can be utilized (Peppard, 1998). The most direct way is to simply
generate a list of the business processes that an organization is believed to encom‑
pass. Such a job will often be based on existing process descriptions or procedures
written for ISO 9000 certification or similar purposes.
A more rewarding and systematic route is to map out the following:

• The strategy of the organization

• Its stakeholders, that is, organizations, institutions, or people affected by or
with a vested interest in the organization and its business processes
• Expectations with regard to products or services delivered by the
• The business processes that produce these products or services and support
and enable the delivery of them

The relationships between these elements are shown in Figure 3.6.

Hopefully, you realized that the first three items are the same elements that
have been discussed so far in the book. By going through this list, it is much easier
to point out the business processes that are or should be carried out by the orga‑
nization and that are necessary in fulfilling the expectations of its stakeholders.
Ideally, your organization should have exactly those business processes that are
required to satisfy the strategy defined and its stakeholders. Anything beyond these
could be seen as waste, unless they are support processes necessary for ­performing

shape encapsulated in

defines delivers

Stakeholders Processes

hold deliver

satisfy consolidate

Figure 3.6  Business processes derived from strategy and stakeholders.

38  Chapter Three

the primary processes; anything less could lead to dissatisfaction on the part of
some stakeholders.
By nesting backward from the output required by the stakeholders to the pri‑
mary and support processes and input into these, several strings of business pro‑
cesses emerge. Even if this approach does not cover every conceivable process
ever performed by the organization, this is OK. Processes not encountered when
backtracking from the stakeholders’ expectations are hardly crucial in satisfying
the stakeholders. Thus, if omitted, they will rarely be missed.

3.5  Business Process Modeling

A general rule for improving something is that it is necessary to know the current
state of things. This is very true for business processes as well. If you do not know
how the process looks and works today, it will be very difficult to know which
improvement initiatives can be started and whether they will work. Documenting
one’s own process should therefore always be the first step in any improvement
activity. Modeling, in general, carries with it several other advantages:

• Modeling requires confrontation and clarification where disagreements or

ambiguities prevail, and it does not allow vague assumptions.
• Modeling leads to a general increase in the level of specification; assump‑
tions, surroundings, and objectives must be explicitly expressed.
• A model also allows more advanced simulations or scenario testing and can
be used as a tool in ­what-­if assessments.
• A business process model represents a documentation of deliberately
designed processes. It can be used to demonstrate to customers, regulatory
bodies, or other actors that the organization employs a systematic approach
to process design and fulfills requirements posed. For customers, especially
industrial or potential, the model can be used as a marketing tool to con‑
vince them that your business processes are sound and reliable.
• Many types of standards and certification require extensive process model‑
ing, including the ­process-­based year 2000 version of ISO 9000.
• If modeling is done right, that is, in collaboration among many or all of
those involved in the business process being modeled, it stimulates involve‑
ment and enthusiasm.
• The dialog connected with the modeling process creates learning and a better
understanding of the process and its boundaries. This might be the most impor‑
tant benefit; after an extensive business process modeling effort in an organiza‑
Understanding Your Current Business Processes  39

tion, employees are much more aware of their role in the bigger scheme of things
and how their effort contributes to the overall objectives of the organization.

Business processes can be documented at two different times:

• One by one in connection with improvement projects involving the specific

• All at once at the start of a general “improvement journey”

You might select an approach where a process is not documented until a project or
some other activity is started to improve it. In such cases, process documentation
is the first activity of the project and serves several purposes:

• A common understanding within the improvement team of the content of

the process is created, including its activities, results, and who performs the
different steps
• The scope of the process is defined, as are the boundaries to adjacent
• If desired, it is possible to define more specific problems within the main

This approach works quite well in practice. There is no need to document many
processes at the start, as this is carried out when the need arises. For smaller organi‑
zations lacking resources or for organizations where the business processes change
rapidly, this is probably the preferred approach.
The other way is to initiate the improvement journey by going through the entire
organization and modeling all or most of the business processes. As mentioned in sec‑
tion 3.3, this might involve everything from 15 to 100 business processes. It is not pos‑
sible to cover every single process performed within the enterprise; the purpose is to
document the most important ones. The advantages of this approach are as follows:

• By involving a large number of employees in this work, a positive atti‑

tude and improvement motivation that are useful in later projects are often
• After undertaking such an exercise, top management will have formed a
very good overview of the organization and the need for specific improve‑
ment projects
• Insight into the individual processes that is created during the documenta‑
tion work often makes it possible to pinpoint elements of individual pro‑
cesses that can be improved
40  Chapter Three

To the extent that the time and resources necessary for such an introductory pro‑
cess modeling can be mustered, this is probably the most valuable approach. Many
organizations today have done this exercise of modeling for most of their processes.
For many, it has been part of the path toward ISO 9000 certification; for others,
it’s simply a useful analysis. However, bear in mind that process models and pro‑
cedures in quality handbooks often reflect an ideal situation, not always the real
situation. In improvement projects, it is essential to start from the real situation to
be able to identify problem areas. Furthermore, most processes are to some extent
dynamic, often rendering existing descriptions inaccurate or erroneous. Checking
the quality of such material before using it can save much extra work. The objec‑
tive must always be to document the process as it is performed today, not the way
it was supposed to be performed.
Before moving on to more specific techniques for modeling business pro‑
cesses, I should point out that it might not be as straightforward as I have perhaps
made it seem so far. There is, in fact, a hierarchy of models at different levels that
could be employed, as shown in Figure 3.7 (which I should point out represents the
most complex situation; you might say a worst case).
Let me quickly introduce each of these:

• The stakeholder model has already been discussed (see section 2.1), and
it is merely a graphical representation of the stakeholder analysis. In this
hierarchy, it gives the ­highest-­level view of the organization.
• The industrial value system, often termed value chain, more specifically
defines the organization’s position in the larger chain of ­supplier-­customer
relationships that form the chain, from raw material suppliers to end ven‑
dors of products or services to the final customer. Every organization is part
of one or more such value systems, and understanding your position in them
is an important insight to bring along in the process modeling task. I will not
pursue the topic of value systems/value chains/supply chains any further in
this book, as it is simply too comprehensive to allow a fair treatment here.
• The overall process model is perhaps the first true business process model
in the hierarchy, but it does not represent an individual process. Rather, such
overall models attempt to illustrate how the individual processes are linked,
in chains or networks, inside the organization. They can often make the big
picture clearer than if you jump directly to single processes. There are many
ways to construct an overall model; the one shown in the figure merely offers
an example. Some models of this type are quite detailed, indicating flows that
exist between the different processes; others simply portray them together in
the same chart. I will not discuss this level of modeling in much more detail
here, but you should decide whether you will include this level of modeling
and, if so, in which style. How to identify the business processes that should
go into the model, however, was briefly discussed in section 3.4.
Understanding Your Current Business Processes  41

Stakeholder analysis

Authorities Owners Management

Suppliers Customers

institutions Competition

Employees Environment partners

Industrial value system


Overall process model

Primary Marketing Order

Purchasing Manufacturing Distribution Invoicing
processes and sales processing

Support HR
Maintenance Safety, health, and environment
processes management

Development Product Market Supplier

processes development development development

Relationship map

Relevant Quality
Supplier Process X
process control

Negotiations Quality check and agreement approval

Identify suppliers, price/offer

Process X negotiate and sign

agreement Request for

Need for new or

revised agreement

Agreement Request offers and Negotiations for articles for

Signed agreement which no agreement exists Supplier
evaluation perform negotiations


Place order with

Defined Receipt of product
Process X supplier according Follow-up of order Payment
article? or service
Yes to agreement

Follow-up and Payment for

Order response Invoice delivery

Supplier Supplier Supplier

Figure 3.7  The hierarchy of process modeling.
Individual flowcharts

Marketing Product
department R&D department planning Manufacturing
article? or service
Yes to agreement

Follow-up and Payment for

Order response Invoice delivery

42  Chapter Three

Supplier Supplier Supplier

Individual flowcharts

Marketing Product
department R&D department planning Manufacturing

Initiation of

Design of product concept and

suggestion for technical solution

Development of Process Trial

specifications planning manufacturing

Process certification

Product ready for


Procedure xx

1. Step a
2. Step b
3. Step c
4. Step d
5. Step e
6. Step f
7. Step g

Figure 3.7  Continued.

• In between the overall process model and the individual flowcharts, some
find it helpful to include a relationship map. This portrays the main inter‑
faces and linkages among units inside the organization itself and external
actors that interact to perform the business process.
• Individual flowcharts, which are the most common type of process models
and the ones that are most useful in process improvement work. There are
several different styles of modeling, and these will be discussed in more
detail shortly.
• Finally, I have also included the type of ­text-­based written procedures that
were more common some years ago. When developed into a high level of
Understanding Your Current Business Processes  43

detail, these are probably the most precise process descriptions. For example,
they often include detailed instructions about which button to press in a com‑
puter program or how to hold a mechanical tool. However, they are often
difficult to understand quickly, and they rarely contribute to the overall under‑
standing of a process. Thus, I generally do not worry too much if this level is
not covered when the aim of the modeling is improvement purposes.

The following sections present in more detail some of the most commonly used
and, in my opinion, most useful modeling approaches:

• Relationship mapping
• Traditional flowchart, including the IDEF0 standard
• Cross-functional flowchart
• Flowchart divided into process segments
• Several-leveled flowchart
• Flowchart with statistics

3.6  Relationship Mapping

Before we can draw a detailed flowchart of a process, it is often necessary to create

an overall picture of the individuals who are part of the process and their relation‑
ships with one another and the rest of the world. This is especially true for more
extensive and more complicated processes involving a number of individuals or
departments. Is the objective, for instance, to document the process of order receipt
and delivery of goods to the customer? If so, it can be quite a challenge to put in
place the separate steps of the process just like that. In such a situation, relationship
mapping is an ideal first step.
In contrast to an ordinary flowchart, a relationship map does not consider
activities or the sequence of such. The map is constructed by placing on a blank
sheet the different units, department, or individuals expected to participate in or
impact the process. For the process of order receipt and delivery, logical partici‑
pants would be the sales, planning, production, and procurement departments
internally, as well as customers and suppliers. Furthermore, we might assume that
the finance department and external transport companies would be included. A
general rule is that it is better to include too many elements, as irrelevant ones will
naturally be eliminated throughout the process. It is also possible to draw such a
map on several levels so that each department can be further detailed on a lower
level without blurring the overall picture.
44  Chapter Three

After establishing the potential participants in the process, each relationship

among them is analyzed to define its type. Different types of arrows are suitable
for this purpose. Elements without any relationships to other elements are removed
from the map. In the end, the map is redrawn and will give a good overview of
relationships between participants and stakeholders in the process.
Figure 3.8 shows a relationship map for the sample process. The types of
arrows used in this map are only suggestions; there are no standards in this area.
It should also be pointed out that the task of conducting such a relationship
mapping and the preceding activities of constructing a flowchart and other tasks
related to process documentation must be carried out in a group that includes
the most central participants in the process. The objective is to improve and
adjust the process model until it reflects how the process is performed today.
On the other hand, you should be careful not to spend too much time creating
a very detailed and accurate depiction of the process. It might very well be that
a reasonably good model achieved quite easily is better than one that demands
many resources to generate but is perfect. This can be seen in the context of the
later use of the process model.
To summarize, the procedure for constructing a relationship map is:

1. Identify the internal and external organizational units that play a part in the
business process.
2. Construct a diagram where these units are placed in logical positions
toward one another.

Own company

Sales Planning

Manufacturing Purchasing


Flow of goods Information

Order Discussion/negotiations

Figure 3.8  Example of a relationship map.

Understanding Your Current Business Processes  45

3. Go through the entire diagram and consider if there is a relationship

between any two units. For each relationship identified, indicate this in
the diagram using a type of arrow that corresponds to the type of link that
4. After completing this analysis, remove from the diagram any units that do
not have links to other units, unless it is a point to demonstrate specifically
that no links exist.
5. If the diagram becomes difficult to interpret, rearrange it by moving units
with extensive links closer to each other.


A large international corporation was organized with one central manu-­

facturing site covering all of Europe and local dealerships with their own
inventories of finished goods in a number of countries. There were clear
indications that the supply process, including communicating needs from
the local dealers and distribution to them, was not working satisfactorily.
Thus, the company started a project to improve the material flow, begin-­
ning with the main manufacturing site. However, it proved difficult to form
an overall impression of the flow of information and goods, and it was thus
decided to start the project by drawing a relationship map.
The first step was to collect information from all involved parties about
the most important transactions they were involved in:
• The local dealers based their demand prognoses on ­ twice-­yearly
sales consultations with their main customers
• Using the aggregated information from these consultations, the
dealers transmitted their expected demand for the different product
families to the manufacturing unit
• Here, information from all the dealers was further aggregated into
rough prognoses for the next six months
• These formed the basis for negotiations with suppliers about frame
agreements for the following period
• The local dealers issued detailed orders every month with a fixed
delivery time from the manufacturing site of three weeks
• The ordered products were either taken from a small ­finished-­goods
inventory or manufactured before they were sent by truck to the
The local dealers were invoiced only when the goods had been sold to end
customers. This produced the relationship map shown in Figure 3.9.
46  Chapter Three

about frame Goods

Goods manufacturing site Goods

Prognoses Prognoses
and orders and orders
Local Local
dealers Goods dealers
Sales Sales
Goods Goods
consultations consultations

End End Local End End

customers customers dealers customers customers


End End
customers customers

Figure 3.9  The resulting relationship map for the supply process.

3.7  Traditional Flowchart

A flowchart is a graphic depiction of the flow of activities in a process. The use

of flowcharts reinforces the fact that it is much easier to understand something
presented graphically instead of described by words. Put differently, an image is
worth a thousand words.
There are many ways of drawing a flowchart (Andersen and Pettersen, 1996).
The most basic is simply to use different symbols to represent activities, and arrows
to illustrate the connections between the activities. There are a number of variants
in the symbols used, including everything from complex pictures to simple boxes
and lines. One way is not necessarily better than another; the point is simply that
the users must understand the symbols. Some commonly used symbols are:

Start or finish point

Step or activity in
the process

Decision point

Input or output

Start or finish point

Step or activity in
the process
Understanding Your Current Business Processes  47
Decision point

Input or output


Manual operation

Storage or waiting

In addition, it is possible to indicate on the side of the symbols in the flowchart

what equipment or resources are being used and under what conditions the activ‑
ity is being performed. A flowchart for the sample process discussed earlier might
look like the one shown in Figure 3.10.
This chart could have been made more detailed—for instance, by including the
suppliers in the process, the negotiations with the suppliers and customers, and so
on. Nevertheless, it illustrates the principles for drawing a flowchart.

Customer demand




Production Procurement
plan needs






Figure 3.10  Flowchart for the supply process.

48  Chapter Three

The steps to design a flowchart are as follows:

1. Define the boundaries for the business process to be modeled, especially

the start and end points. Both should be ­well-­defined outputs or events.

2. Starting from the end point, identify the activities carried out in the process,
along with important outputs generated, periods of waiting, and so on.

3. Construct the flowchart graphically by placing items in correct sequence,

moving from the end of the process backward. In cases of decision points
or places where the process branches out, try to orient the diagram so that
activities flow in the most suitable direction, that is, from left to right and
top to bottom.

4. In cases of disagreement in the modeling team, take time to resolve any

issues so that the final model truly reflects a common understanding.
Remember that such disagreements often represent good opportunities
for clarifying potential conflicts or misunderstandings.

5. When the team is satisfied that the most important elements of the process
have been captured and placed in the correct sequence, it is a good idea to
redraw the flowchart to increase readability.


A group of secretaries in a public office experienced problems with docu-­

ments and other materials that, once they were filed, were difficult to find
again. Among the employees, there was a hunch that some people used dif-­
ferent methods for filing than others. It was therefore decided that the way
the process is carried out—and how it really should be carried out—should
be defined exactly. For this purpose, they wanted to use a flowchart.
Armed with a white board and pads of small yellow stickers, the sec-­
retaries gathered in a meeting room. It was soon evident that they all per-­
formed approximately the same steps when filing documents, but the
criteria for where material was filed varied quite extensively among them.
After a heated debate, they were able to agree on both the process and the
filing criteria. The flowchart in Figure 3.11 is the result of the exercise.
Understanding Your Current Business Processes  49

Material for

Type of
Decisions and material? Letters, faxes,
resolutions and e-mails

File under
File under File under
“cases under
“closed cases” “correspondence”

Criterion Criterion Criterion

alphabetically date case

document list


Figure 3.11  Flowchart for the filing process.

When presenting the tool of flowcharting, I would be negligent not to men‑

tion IDEF0, often referred to as IDEFØ, that is, ­IDEF-­zero. IDEF0 (where IDEF
is an acronym for Integrated DEFinition method) is best described as a standard
for modeling decisions, actions, and activities of an organization or system. It was
derived from a ­well-­established graphical language called the Structured Analysis
and Design Technique (SADT). Because of its consistency in its notation and rules
for how flowcharts are designed, it is quite a powerful approach. However, I find
that the resulting flowcharts, especially for more complex processes, become quite
difficult to read. They are correct in the sense that all flows are captured, but seeing
the big picture can often be difficult. Therefore, I think IDEF0 is best suited for
­high-­level modeling of a process.
In an IDEF0 diagram, four types of links between items in the chart are defined:

• Controls: requirements, laws, company policy, and so on, that dictate the
behavior of the step in the process
• Mechanism: resources that carry out or support the process
50  Chapter Three

• Input: information, materials, and so on, that initiate the process

• Output: the result of the process (transformation of input)

These four connections are graphically linked to one predetermined side of the
square representing the process step, as shown in Figure 3.12.
Figure 3.13 shows a typical IDEF0 flowchart, and I’d like to point out that
this is a fairly simple process that has been modeled. Even so, you can already


Input Process Output


Figure 3.12  Standard IDEF0 diagram design.

Management & control Evaluation feedback Market

(internal, EC, audits) C3 & recommendations C4 feedback
RC milestones C2

Technical WP R1
annex Define
I1 reqs + specs

P. 5 03

LC Models
models V1 04
A112 models v1
P. 6

WP R3 scenarios
Define 01
& demos 02
P. 7

Projects M1 M2 Partners

Figure 3.13  Sample IDEF0 diagram.

Understanding Your Current Business Processes  51

see that the diagram has become quite complicated. Compared with the simpler,
albeit perhaps technically not as correct, flowchart shown in Figure 3.10, this one
is more difficult to relate to and understand the logic of. Although I might sound
overly critical toward the IDEF0 standard, I have found the approach less useful in
­real-­life business process modeling.
The notation of IDEF0 in Figure 3.12 is useful for guiding the modeling effort.
Other notations also exist, such as the turtle approach. The turtle diagram was
designed to remind people who are modeling business processes to capture the fol‑
lowing relevant information: inputs, outputs, equipment required, skills required,
work instructions, and performance measures.
All such standards have been designed to ensure that the flowchart is logi‑
cally constructed. One problem with standard, simple flowcharts is that it can be
difficult to see who performs which tasks. This information is better presented in
­cross-­functional flowcharts, which is the next topic.

3.8 ­ Cross-­Functional Flowchart

As has been mentioned, an ordinary flowchart mainly describes what activities

are performed in a process. A cross-functional flowchart shows who performs the
activities and which functional department they belong to—from which the name
arises (Andersen and Pettersen, 1996). Figure 3.14 shows an example, where the
ordinary flowchart in Figure 3.10 has been supplemented with more detailed infor‑
mation. Each department has been placed inside a “swim lane,” and activities per‑
formed by the department are located in this column.
Adding these pieces of information does not take much time compared with
the job of describing the sequence of activities itself, but it contributes to a much
clearer representation of the process. It is therefore generally recommended to
use a ­ cross-­functional flowchart. Another advantage of this type of flowchart is
that it allows you to analyze when responsibilities in the process shift. As was
discussed earlier in this chapter and shown in Figure 3.3, when a business process
traverses a department boundary, delays in the process are usually introduced. A
­cross-­functional flowchart makes it easy to see how many times and where this
happens, thus allowing for improvements by either redefining responsibilities
or placing steps performed by the same department in sequence. Frequent shifts
between departments or actors in the process also indicate a danger that the overall
process responsibility is highly fragmented, increasing the likelihood that no one
will take charge for the performance of the process.
I should also mention that the swim lanes can denote external players in the
process, physical locations where steps of the process are executed, or other units.
It is also possible in a ­cross-­functional flowchart to display additional information.
Along the vertical axis, or the horizontal axis if the process is drawn in a landscape
format, information such as the following can be added to the chart:
52  Chapter Three

Planning Procurement Manufacturing Distribution

Customer department department department department








Figure 3.14  Example of a ­cross-­functional flowchart.

• Time spent so far in the process

• Incurred costs thus far in the process
• Value added
• Degree of completion

With these additional details, the flowchart is able to convey much more informa‑
tion than just the pure sequence of activities in the process. To construct a ­cross-
­functional flowchart, proceed as follows:

1. Define the boundaries for the business process to be modeled, especially

the start and end points. Both should be ­well-­defined outputs or events.
2. Starting from the end point, identify the activities carried out in the process,
along with important outputs generated, periods of waiting, and so on.
3. For each of the items in the process, determine which organizational unit
is in charge of executing the item.
4. Construct the flowchart graphically by creating swim lanes or columns
that correspond with the organizational units involved, preferably in such a
sequence that units with close cooperation are located next to each other.
5. Continue by placing items in the correct sequence and in the correct swim
lane, moving from the end of the process backward. In cases of decision
Understanding Your Current Business Processes  53

points or places where the process branches out, try to orient the diagram
so that activities flow in the most suitable direction, that is, from left to
right and top to bottom.
6. In cases of disagreement in the modeling team, take time to resolve any
issues so that the final model truly reflects a common understanding.
Remember that such disagreements often represent good opportunities for
clarifying potential conflicts or misunderstandings.
7. When the team is satisfied that the most important elements of the process
have been captured and placed in the correct sequence, it is a good idea to
redraw the flowchart to increase readability.


A large company had noticed during the last few years that it was quite
expensive to generate the necessary information for financial reporting to var-­
ious public institutions, including tax authorities. To gain an overview of who
did what in the process and where costs were incurred, a ­cross-­functional
flowchart was constructed. The resulting chart, shown in Figure 3.15, was
well suited for this purpose and helped reduce costs considerably through a
redistribution of responsibilities for different tasks. Effort was made to ensure

Incurred Authorities Financial manager Clerk All departments

0 Report call

100 of call

Identification of
250 necessary

600 Collection of

3900 Submission
of information

4300 Compiling of



Figure 3.15  ­Cross-­functional flowchart for the reporting process.

54  Chapter Three

that the necessary information from various departments was systematically

collected, rendering infrequent, ­all-­out efforts unnecessary.

Please keep in mind that the more information that is added, the more com‑
plex the flowchart typically becomes. Dividing it into logical areas is one way of
increasing the readability.

3.9  Flowchart Divided into Process Areas

Even flowcharts not designed using the IDEF0 notation can become too complex to
be easy to work with. The chart in Figure 3.14 is perhaps not a good example, but if
the process to start with is long and complex, or much other information is added to
the flowchart, it might become difficult to read. Two possible ways of dealing with
this are (1) divide the flowchart into distinct process areas and (2) split the flowchart
into several smaller charts (the second approach is dealt with in section 3.10).
Dividing the chart into process areas is simply a matter of identifying logical
segments of the process that can be grouped together. These can then be set apart
from other areas by drawing them in different colors, placing them inside rect‑
angles with borders separating them from other process areas, or shading the area
of the process. An example of this is shown in Figure 3.16. Although this is not a
complex process in itself, it illustrates the principle. This is a process consisting of
three distinct segments: sea transport, coordination of land transport, and the actual
land transport. By placing these inside three separate rectangles, they are logically
kept apart, and each can be viewed separately within the big picture. For processes
with tens, if not hundreds, of steps, this is quite effective.
Since this is simply a variant of the traditional flowchart, a more detailed
example of its use has not been included. The steps to create such a chart are simi‑
lar to those of a regular flowchart:

1. Define the boundaries for the business process to be modeled, especially

the start and end points. Both should be ­well-­defined outputs or events.
2. Starting from the end point, identify the activities carried out in the process,
along with important outputs generated, periods of waiting, and so on.
3. Construct the flowchart graphically by placing items incorrect sequence, mov‑
ing from the end of the process backward. In cases of decision points or places
where the process branches out, try to orient the diagram so that activities flow
in the most suitable direction, that is, from left to right and top to bottom.
Understanding Your Current Business Processes  55

Land logistics Land transporter

Finished vehicles logistics National autologistics

Follow up status
and invoices for
Land transport status

Create and Transport

negotiate contracts
Distribution network
transport routes handling
Deep sea
Freight lines details

Send vessel
Book transport
Timer Land transport
volume forecast

Figure 3.16  Example of a flowchart divided into process areas.

4. In cases of disagreement in the modeling team, take time to resolve any

issues so that the final model truly reflects a common understanding.
Remember that such disagreements often represent good opportunities for
clarifying potential conflicts or misunderstandings.
5. At this point, examine the flowchart for any distinct phases or areas of the
process. If useful, these can be separated into their own areas, using either
shading or rectangles to create boundaries around them.
6. When the team is satisfied that the most important elements of the process
have been captured and placed in the correct sequence, it is a good idea to
redraw the flowchart to increase readability.

3.10 ­ Several-­Leveled Flowchart

Flowchart complexity can also be remedied by breaking the chart into several hier‑
archical levels, for both ordinary and ­ cross-­functional flowcharts (Andersen and
Pettersen, 1996). The principle is that the top level shows only the main activities,
which are numbered 1.0, 2.0, and so on. This top level gives a good overview of
the entire process without blurring the major aspects with many details. Figure 3.17
shows a somewhat simplified version of the chart in Figure 3.14 as a level 0 chart.
To present the main activities in more detail, another flowchart on the level
below this one is created. For activity 3.0, the more detailed chart is shown in
56  Chapter Three

Customer demand






Satisfied customer

Figure 3.17  Level 0 chart for the process.

Figure 3.18. The individual steps at this level are logically linked to the level
above by numbering them 3.1, 3.2, and so on. If there is a need for further
detail, more levels are simply added. With this technique, it is possible to pre­
sent the overall process using only a level 0 chart. As the need occurs, details
are displayed for the individual steps of the process through ­lower-­level charts.
This makes the technique very powerful through combining a lucid presenta‑
tion format with details where needed. If the flowcharts are stored and accessed
electronically, perhaps as part of an intranet accessed through a Web browser,
the hierarchical representation is well suited for navigation down through the
levels. Simply clicking on a step in a chart on a higher level takes you to the
more detailed chart for that step.
A set of flowcharts at different levels is created through the following steps:

1. Define the boundaries for the business process to be modeled, especially

the start and end points. Both should be ­well-­defined outputs or events.
2. Starting from the end point, identify the main, ­ high-­level activities car‑
ried out in the process, along with important outputs generated, periods of
waiting, and so on.
3. Construct the flowchart graphically by placing these ­high-­level items in
correct sequence, moving from the end of the process backward. In cases
of decision points or places where the process branches out, try to orient
Understanding Your Current Business Processes  57

Procurement list



an order

Receiving order

Registering a
delivered order

Parts for

Figure 3.18  Level 1 chart for procurement.

the diagram so that activities flow in the most suitable direction, that is,
from left to right and top to bottom. Activities in this ­upper-­level chart are
numbered in sequence from start to finish.
4. In cases of disagreement in the modeling team, take time to resolve any
issues so that the final model truly reflects a common understanding.
Remember that such disagreements often represent good opportunities for
clarifying potential conflicts or misunderstandings.
5. At this point, examine the flowchart for any ­high-­level activities that would
benefit from having details presented in a separate flowchart. For each of
these, separate charts are created by following steps 1–4. These are num‑
bered in such a way that they link the activities to the correct step in the
­upper-­level chart. If required, further flowcharts of even more detail can be
6. When the team is satisfied that the most important elements of the process
have been captured and placed in the correct sequence, it is a good idea to
redraw one or more of the charts to increase readability.
58  Chapter Three


To ensure that all departments used the same procedure for storing the
necessary information for financial reporting, the company from the last
example decided to document the ideal information flow of the process by
using ­several-­leveled flowcharts. The flowchart in Figure 3.15 constituted
the level 0 chart, and the step that represented the activity taking place in
all departments was numbered 4.0. Because the reporting required dif-­
ferent information from different departments, ­department-­specific proce-­
dures were produced at level 1. For the procurement department, the chart
is shown in Figure 3.19.

Monthly registering of:
• Procurement value
• Supplier distribution
• Country of origin distribution
• Delivery times

Quarterly registering of:
• Spent workforce hours
• Costs for wages
• Value-added tax paid

Year-end registering of:
• Total purchasing value
• Supplier distribution
• Country of origin distribution

All reports are
generated and stored
in spreadsheets

Figure 3.19  Level 1 chart for reporting.

In cases where processes branch out into several possible parallel paths, and
where it is useful to understand the importance of each path, statistics can be added
to the flowchart.

3.11  Flowchart with Statistics

One problem I have encountered many times when modeling processes is that
obscure exceptions or special cases require the same, if not more, attention than
Understanding Your Current Business Processes  59

the most common flow of events. Very often, you will find that a certain process is
performed according to a standard sequence. For some customers or some prod‑
ucts, or due to other exceptions, there will be smaller or larger deviations from this
norm. When modeling this process, the modeling team will insist on including the
This makes the flowchart more complex than what is strictly necessary. Some‑
times you’d be better off simply modeling the main flow and disregarding these
exceptions, as this will be sufficient for understanding the logic of the process. In
other cases, the exceptions might be what cause problems in the process, and they
must be captured. In any case, adding some sort of load statistics or weights to the
different flow possibilities can be helpful. These indicate how often the main and
alternative flows are used. The approach is best demonstrated using an example,
but let me first briefly list the steps to use when making such a chart:

1. Define the boundaries for the business process to be modeled, especially

the start and end points. Both should be ­well-­defined outputs or events.
Keep in mind that this type of chart is best suited for processes where there
are two or more alternative flows.

2. Starting from the end point, identify the activities carried out in the pro‑
cess, along with important outputs generated, periods of waiting, and so

3. Construct the flowchart graphically by placing items in correct sequence,

moving from the end of the process backward. In cases of decision points
or places where the process branches out, try to orient the diagram so that
activities flow in the most suitable direction, that is, from left to right and
top to bottom.

4. In cases of disagreement in the modeling team, take time to resolve any

issues so that the final model truly reflects a common understanding.
Remember that such disagreements often represent good opportunities for
clarifying potential conflicts or misunderstandings.

5. At this point, determine where in the flowchart the statistics could be use‑
ful, typically for parallel and alternative flows through the process. Use
existing statistics or collect the required statistics about usage of the alter‑
native flows. The statistics are inserted into the chart to indicate how often
each alternative is used.

6. When the team is satisfied that the most important elements of the process
have been captured and placed in the correct sequence, it is a good idea to
redraw the flowchart to increase readability.
60  Chapter Three


A small branch office of a bank had started a project to redecorate the

offices. To determine the best layout of the refurbished office, it was impor-­
tant to understand the flow of customers in the bank. The main questions
discussed were how much capacity, and thus space, to allocate to the
main reception, which would be the first point of contact for customers,
and how to structure the back offices, where customer representatives
specializing in various areas would be located.
During one week of service with the current office layout, all customer
activity was logged. A flowchart depicting the physical flow of the custom-­
ers through the bank was constructed, and data showing the number of
customers taking the different routes were added. The resulting chart is
shown in Figure 3.20, and it helped the bank decide to increase the capac-­
ity (and level of experience) of the front reception. Many people visiting the
bank had appointments, and they were simply shown into the offices of the
people they came to see. Additionally, many had requests that could be
settled in the reception area. If the requests could not be handled there, the
customers were referred to the customer representatives, taking up valu-­
able time that the representatives could have used to pursue sales leads.

Example continued

In concluding the treatment of business process modeling, I should proba‑

bly mention that other means of modeling processes have come into use lately.
Some find that simple symbols are less intuitive in flowcharts. To make process
charts more readable, using more detailed drawings or even photographs can
help. Photos of actual machines, people, or locations can help liven up any
Some organizations have successfully used video to document processes. It’s
especially helpful for those processes where a very precise sequence of opera‑
tions or performing steps exactly the right way is important. For such processes,
video can offer much more detail than any flowchart or written procedure. Think
of all the small operations involved in replacing the transmission in your car, and
how confusing schematic drawings of such an operation can be. Having access to
video of the process being performed, allowing rewinding when necessary, would
naturally be much more helpful for a car mechanic uncertain about the details of
the process.
I have even heard that a telecom company has recorded various repair
and maintenance procedures for cell phone base stations on video and com‑
pressed the files to their limit. Should service people find themselves far into
the wilderness and facing a repair problem not anticipated, the video clip can
Phase 1: Reception Phase 2: Initial assessment Phase 3: Sales
26 Customer not relevant
No customer representative customers for the bank office
available/appointment 15 Directed
customers to self-
solution used
Initial assessment Simple recommended
Referred of needs at service requests 90
desk solved customers
Directly to Assessment of Needs
Referred cash counter needs in cash desk assessed

Customer comes
to bank
Assessment of Needs
Not needs “daily bank” assessed

Customer assessed
Customer goes to referred
service desk Not Referral of customer
Assessment of
146 customers needs in credit
to service desk Referred department
Assessment of
needs in savings

Need for
External referral services
of customer

Referral to
Needs 10
insurance, property,
assessed customers

Referral to
Needs 3
assessed customers
Understanding Your Current Business Processes  61

Figure 3.20  Flowchart with bank customer movement statistics.

62  Chapter Three

be downloaded to the cell phone and viewed. A hospital has taken a slightly
different approach by videotaping all operations. If a surgeon experiences a
problem during the surgery, she or he will afterward record voice narrative on
top of the video clip to describe the problem, how it was handled, whether the
actions taken were successful, and so on. The clips are indexed with descrip‑
tive keywords and stored in a database. When another surgeon prepares for a
certain type of operation, all video clips pertaining to it can be retrieved and
reviewed. This way, the surgeon will have fresh information about possible
problems and how to deal with them.

3.12  Paper and Pencil or PC?

A highly relevant and general question for any type of process modeling and con‑
struction of flowcharts is what medium should be used. Is it better to stick to the
old tools of paper and pencil or be more up to date and draw on a PC? I believe the
answer for this one is both.
In the first phase, trying to agree on how the process truly looks, it is definitely
an advantage to use paper. In the type of plenary sessions common in this phase, a
flip chart or other type of large drawing surface hung on the wall should be used.
Usually, many opinions surface, and the flowchart will be a dynamic representa‑
tion that is constantly changing. A technique that enables quite painless adjustment
to these dynamics is using colored sticky notes to represent process elements. As
the chart is being developed, the notes can be moved around, and the bother of
erasing and redrawing is avoided.
For this purpose, a computer is of comparatively little use. First of all, the
screen is too small. Even if it is enlarged by using an overhead projector, it is still
awkward working on a PC in this type of meeting. Furthermore, I’ve learned from
experience that there is always a danger that the software tool itself takes attention
away from the process and the modeling task.
In the next round, however, it is necessary to capture the flowchart from the
less manageable, large format used during its development. The flowchart must be
stored, copied, and modified. For these purposes, the computer is clearly highly
suitable. There are a number of software tools available on the market that make
the construction and adjustment of flowcharts very easy.
After the flowchart has been completed, but also between meetings during
the process modeling, a PC is used to draw the flowchart. The chart can then be
printed, copied, and distributed. In addition, the file itself can be sent by e-mail
to others who might need the chart (for instance, benchmarking partners). The
software allows the user to update the chart as meetings are held. Afterward, the
chart can be electronically stored, updated, and changed as the business process
Understanding Your Current Business Processes  63

changes over time. It is difficult and bothersome to do this maintenance by

hand, as it involves a lot of erasing and rewriting each time the chart is updated.
Such a ­two-­step approach, where the more suitable medium for different pur‑
poses is used, is by far more favorable than categorically selecting one or the
other for the entire job.
Using Performance Measurement
in Business Process Improvement

ne argument for the importance of process modeling is that to improve
something, you need to know the current state of things. Correspondingly,
the general argument for performance measurement is that to improve a
process, you must know how well it is performing today. This chapter gives a brief
overview of the most important principles of performance measurement.

4.1  The Role of Performance Measurement in Business

Process Improvement

Trying to run an organization without access to relevant and pertinent information

on performance is like trying to drive a car, fly an airplane, or command any other
type of vehicle without any instruments. The information provided by the differ‑
ent types of gauges, warning lights, and alarms in a typical instrument panel in a
vehicle gives you the decision basis needed: the oil pressure lamp lighting up tells
you to check the cause immediately and probably add oil, the speedometer tells
you your speed, and a buzzing means you’re driving over rumble strips and you
should make sure you’re not veering off the road.
Although the performance information you get in an organization is not as up
to date and does not require such immediate action, it is still invaluable. An airline
with delays on a flight needs to know exactly how many seats are available on later
connections for passengers who will miss their originally booked flights. A sales
clerk in a shoe store might find that the store has sold out of a pair of boots in a cus‑
tomer’s size, but by checking the inventory of nearby stores in the same chain, he
can reserve a pair at a different store for the customer. When a purchaser in a manu‑
facturing company sees that the delivery precision of a supplier has been gradually
slipping during the last two months, she knows it’s time to either approach the
supplier and demand improvements or look for an alternative supplier.

66  Chapter Four

The list of examples is endless, and there are huge differences in how the
information is provided, how it is used, and so on. There is no doubt that such
performance data provide extremely useful information for deciding to initiate an
improvement project. When investigating how performance measurement is used
in organizations, we came up with a long list of different application areas (Ander‑
sen and Henriksen, 2004):

• Monitoring the overall performance levels of the organization

• Setting a strategic direction and using measurement to ensure adherence to
this direction
• Using average or absolute performance levels to perform detailed opera‑
tional planning of activities and processes
• Using performance history to develop cost estimates for products or
• Basing planning on ­up-­to-date performance data
• Establishing foresight/early warning through monitoring leading perfor‑
mance indicators
• Undertaking regular supplier assessments
• Exploiting performance measurement for altering the behavior of indi‑
viduals, groups, or whole organizations and thus used to promote desired
• Establishing incentives through focusing on certain performance parameters
and using performance data as a basis for bonuses or other rewards
• Using performance measurement to determine what processes need
• Evaluating improvement projects to monitor whether they deliver their
• Exploiting performance data for marketing purposes
• Sharing data with customers, for example, as an approach to documenting
that performance is taken seriously
• Benchmarking

As you can see, most of these have some relevance for improvement work, and we
could add to this list the fact that existing performance data about a process can
also give useful insight into the causes of inferior performance. This can again be
used in an analysis phase in an improvement project to understand which aspects
Using Performance Measurement in Business Process Improvement  67

of the process need changing. In summary, I see a performance measurement sys‑

tem as an integral part of a business process improvement system. By this, I don’t
literally mean that you need a clearly defined system for either, but rather that any
organization that takes continuous improvement seriously needs to have in place
procedures for periodic measurement of important performance aspects. In some
cases, this will be an advanced ­ IT-­based system; for other organizations, it can
simply mean collecting customer satisfaction data every six months.

4.2  Implementing a Performance Measurement System

Although the term performance measurement system has come into general use
in management literature, a system can in this context be of any size and com‑
plexity. Thus, this section briefly deals with how to ensure the existence of some
kind of systematic performance measurement in your organization to support your
improvement efforts.
I should, however, mention that during the last few years, more advanced
­computer-­based systems are being used for performance measurement. In my
view, many organizations have invested far too much time and money in imple‑
menting systems so complex that they often don’t work, and when they do, there
is no real need for all their capabilities. According to Spitzer (2007), 10 percent
of a performance measurement system is technology; the rest is how the system is
used: the context of measurement, its focus, and the integration and interactivity of
measurement. I hope I have made myself clear that this chapter of the book is not
a sales pitch for such systems, but rather a reminder that performance information
is useful for improvement purposes.
A few years ago, a colleague and I wrote a book on designing a performance
measurement system. We outlined an ­eight-­step process for designing and imple‑
menting any type of performance measurement system, simple or highly complex.
Although I assume you will not decide to design such a system in the middle of a
business process improvement project, it might be worthwhile to see the steps such
a process can entail. The design process is illustrated in Figure 4.1 (Andersen and
Fagerhaug, 2001).
The eight steps of the design process are:

1. Understanding and mapping business structures and processes. This is the

introductory step of the design process, whose main purpose is to force those
setting out to design a performance measurement system to think through
and reacquaint themselves with the organization, its competitive position, its
environment, and, not in the least, its business processes. You will probably
realize that this is similar to what has been dealt with previously in this book:
to make sure that improvement efforts are seen in light of the strategy of the
organization and its stakeholders.
68  Chapter Four

2. Developing business performance priorities. This step is about under‑

standing stakeholders and aligning the measurements with their needs and
expectations. Again, this is identical to what I have talked about as being
an introductory exercise to an improvement effort.

3. Understanding the current performance measurement system. If a perfor‑

mance measurement system already exists, then this step is simply a mat‑
ter of reviewing whether parts of this can be reused in the new system.

4. Developing performance indicators. The most important element of your

­state-­of-the-art performance measurement system is the set of performance
indicators you will use to measure your organization’s performance and
business processes. The purpose of this step is to populate the performance
measurement system with an appropriate amount of relevant and precise
performance indicators that are able to drive performance.

5. Deciding how to collect the required data. The main purpose of this activ‑
ity is to arrive at solutions for collecting the necessary data for the defined
performance indicators. These range from manual data collection to auto‑
matic data capture to picking the required information out of existing IT

6. Designing reporting and performance data presentation formats. In this

step, you will decide how the performance data are presented to the
users; how they should apply the performance data for management,
monitoring, and improvement; who will have access rights to perfor‑
mance data, and so on. To some extent, the design of the presentation
formats for the performance data can be compared to the design of the
dashboard in a car, where you have to make judgments regarding ergo‑
nomics, the human mind’s abilities for perception and data interpreta‑
tion, and what signals and cues the performance measurement system
should provide.

7. Testing and adjusting the performance measurement system. Here, the

system is tested so that elements that do not work as planned can be

8. Implementing the performance measurement system. In this step, the sys‑

tem is finally put to official use.

It is beyond the scope of this book to deal with this process in detail, so I will leave
this process here and instead discuss more specifically the performance indicators
that form the heart of such a system.
Using Performance Measurement in Business Process Improvement  69

Step 1 Step 5
Understanding and mapping Deciding how to collect
business structures and processes the required data

Step 2 Step 6
Developing business Designing reporting and
performance priorities performance data presentation


Step 7
Step 3
Testing and adjusting the
Understanding the current performance measurement system
performance measurement system

Step 8
Step 4 Implementing the performance
Developing performance measurement system

Figure 4.1  The performance measurement system design process.

70  Chapter Four

4.3  Performance Indicators

To give some insight into the world of performance indicators, it can be useful to
outline some typical dipole characteristics.

4.3.1  Hard versus Soft Indicators

Hard indicators are pure facts that can be measured directly, whereas soft indi‑
cators are less tangible conditions that must be measured indirectly. Another
set of terms is quantitative measures and qualitative indicators. The time it
takes to carry out something or how much it costs are typical hard indicators.
Quality, expressed as satisfaction of needs or attitudes, is one example of a
typical soft indicator. Some differences between hard and soft indicators are
listed in Table 4.1.
Hard indicators are by far more widely used; soft indicators are seen by
many as being so inaccurate that they are rarely useful. On the other hand, as
Deming (1986) underlined, the most important numbers are often not known.
Management by numbers is one of the deadly diseases that have ruined many
enterprises in the Western world. Customer satisfaction is a good example of
a soft performance indicator that is best expressed as the customers’ attitudes
toward the product or service being delivered. In fear of this being too difficult
to measure accurately, many companies have tried to define customer satisfac‑
tion as the number of complaints or warranty costs. To assume that those who
do not complain are satisfied is at best naïve and at worst completely wrong.
The conclusion is that both hard and soft measures are necessary to give a
complete picture.

Table 4.1  Differences between hard and soft indicators.

Hard indicator Soft indicator
Objective reference Observer bias
Accurately known Surrogate indicator
Hierarchical Multivariable situation


In many cases, when trying to measure softer dimensions of perfor-­

mance, it is necessary to approach the problem through the use of sur-­
rogate indicators. If, for example, there is a need to take a soft measure
of quality of the atmosphere in a meeting room, this can be quite dif-­
Using Performance Measurement in Business Process Improvement  71

ficult. It is possible to measure specific conditions such as temperature,

humidity, and air circulation. The problem with these measures is that
no optimal value can be easily defined, as this will heavily depend on
the individual participants in the meeting. A number of surrogate indica-­
tors can be used in this case:

• The number of people active in the discussion

• The number of ideas or suggestions generated
• The number of people leaving the room for various reasons

Even if they are not direct translations of what was originally attempted
to be measured, such indicators can give indirect interpretations about its
performance level.

4.3.2  Financial versus Nonfinancial Indicators

Financial indicators include both basic and derived indicators of a financial char‑
acter and/or using monetary values as the measurement unit. Such indicators are
almost always hard, but are usually the result of some type of calculation. A few
such measures are shown in Table 4.2.
Focusing on such financial indicators is obviously an important part of the tra‑
ditional way of running and managing an organization. Financial indicators were
often seen as synonymous with performance due to the direct link to the company’s
financial result. The increasing understanding that competitive advantage is more
closely linked to operational conditions such as quality, service, environmental
issues, and so on, has often been met with a redefinition of such operative dimen‑
sions to financially based indicators.
Nonfinancial indicators are common denominators for performance indicators
that have measurement units other than monetary value. They can be both hard and
soft, as shown in the following examples:

• Delivery precision
• Defect rate
• Number of complaints
• Customer satisfaction
• Quality of work life
• Innovation rate
• Brand image strength
72  Chapter Four

Table 4.2  Examples of financial indicators.

Financial indicator Calculation (simplified)
Profit margin Total sales – total costs
Value added Sales – input goods
Turnover of capital Sales/total capital

As for hard and soft indicators, the conclusion with regard to financial and nonfi‑
nancial indicators is that both are needed. Problems arise when one of the catego‑
ries is left out of the equation.

4.3.3  Leading versus Lagging Indicators

Again, these dipole types of indicators are not necessarily conflicting; rather, they
are two types of indicators that should act in a balanced interplay. These two dimen‑
sions are, by the way, often used to illustrate the differences between Western and
Japanese thinking when it comes to management. Western management culture
emphasizes results and measures accordingly, that is, using measurement systems
focused on measuring achievements.
In line with traditional Japanese attitudes, the most important part is per‑
forming the process in a sound manner, which will in turn give the desired
results. This is reflected in the Japanese measurement systems, where far more
emphasis is put on process measures—measures that describe certain important
characteristics of a process and that are assumed to have an effect on the desired
results. An example is the number of meetings held in a ­cross-­functional team;
the typical corresponding Western measure would be the number of changes
Kaplan and Norton (1996) coined the terms leading indicators and lag‑
ging indicators in connection with the balanced scorecard, which is one specific
approach to performance measurement. Lagging indicators are aimed at register‑
ing achieved results. By their very nature, they appear quite late in the ­cause-­and-
effect chain along a business process, when that process has produced its output
(or often has not, which will cause the alarm to sound). The idea is that measuring
results is too late; rather, you should measure issues, which will give you an earlier
warning that the process is not performing as intended and thus will not produce
the intended results down the line.
If you listen to certain types of athletes—especially those active in individual,
­concentration-­based sports such as high jumping, speed skating, or bowling—
you will hear them talk about something similar. In interviews before important
events, they will often say that their strategy is to focus on their tasks and the
Using Performance Measurement in Business Process Improvement  73

Too late!


Cost Flexibility

Time Env.

Knowledge Business Market Financial

development processes performance result

Figure 4.2  Performance measurement as an early warning system.

process, and the results will come. If a bowler is dead set on bowling a strike,
chances are high that he or she will fail. But if the bowler focuses on the ­in-­run,
the arm movement, and the hook technique and makes sure that these are done
well, he or she will not be so tense from the pressure of a strike and the likelihood
of getting one will increase.
Figure 4.2 illustrates this early warning and attention principle in an orga‑
nization. A simplified ­ cause-­and-effect chain starts with the development of
knowledge inside the organization. This knowledge, in turn, is employed to
develop and perform business processes, which then create products or ser‑
vices that achieve some level of performance in the marketplace. This leads to
a financial result, either good or bad. Typically, there is a considerable time lag
from the initial knowledge development to the financial result. The organiza‑
tion will have had plenty of time to make many bad decisions before you can
read about them in the financial statement. Therefore, the accounting system is
not a good tool for monitoring performance. Focusing on market performance
gives you an earlier warning, even earlier if you measure your business process
performance. Perhaps the earliest warning comes from monitoring your basic
knowledge development inside the organization, and it is often a year or two
ahead of when the accounting red light goes off.
74  Chapter Four


A manager of a large agricultural collective in the former Soviet Union won

the prize for the most productive collective three years in a row. The perfor-­
mance measure used was the number of kilograms of meat produced per
year. The fourth year, he shot himself. He had no breeding stock left.

These different dipole types of performance indicators show that performance

must be measured in a balanced manner, which is the main message of the bal‑
anced scorecard concept. It also means covering a number of different performance
dimensions. If you measure only time consumption in your business processes,
there is a danger that employees will sacrifice quality or politeness toward custom‑
ers to achieve this one dimension. Measure only quality and you can end up with
a sluggish process that strives for sheer perfection all the time, much more so than
what is really necessary. Important performance dimensions include the following:

• Time, with regard to speed of delivery, execution, and so on, is becoming

more and more important (Stalk and Hout, 1990)
• Quality, measured as the more specific factor of defect rate, but also as the
product’s or service’s ability to satisfy needs and expectations
• Cost, as one important dimension of the performance picture but not the
entire picture, as it has traditionally been
• Flexibility, which is also a dimension of growing importance; it is expressed
as the extent to which tailored solutions can be delivered or volume fluctua‑
tions can be easily handled
• Environmental impact, which is becoming crucial when discussing the
overall performance of an organization
• Safety, for employees and others who come into contact with the organiza‑
tion and its business processes
• Business ethics, which over the last few years has proved an extremely
important aspect, where failure to uphold ethical standards has brought
down several ­well-­known companies

The next step is to decide which performance targets are critical to achieve and
which improvements this requires.
Creating a Business Process
Improvement Road Map

efore commencing an improvement process, it is important to have a plan.
Randomly starting off in an arbitrary direction will rarely lead to good
results. This chapter deals with the improvement road map in general and
more specific tools that can be used to create it.

5.1  Improvement Project Planning

You have perhaps noticed that I have referred to improvement projects many times
but have not elaborated on the topic any further. I will not spend much time dis‑
cussing it here, either, but it is worth noting that most improvement efforts are
normally carried out as projects. It is possible to envision less extensive improve‑
ment efforts, perhaps involving only one or two persons, not being defined as a
project, and in some cases it might be the right approach. However, for slightly
more complex undertakings, I definitely recommend organizing and managing the
improvement work as a project.
This also means resorting to normal project planning practices, that is, defin‑
ing objectives, deriving required activities to attain the objectives, estimating dura‑
tion and cost of activities, and assigning responsibilities. For all these issues, there
are countless good project management books available. Thus, this chapter does
not deal with traditional project planning or management.
Rather, the purpose here is to address how to ensure that the investments you
make in improvement efforts are aligned with strategic priorities, with the needs
of stakeholders, and with current performance shortcomings. There are several
tools and techniques available for this type of analysis. Fortunately, these also

76  Chapter Five

c­ omplement one another by addressing different aspects or viewing available data

from different angles. The tools that will be covered here are:

• Trend analysis
• Spider chart
• Performance matrix
• Criteria testing
• Quality function deployment (QFD)

5.2  Trend Analysis

Trend analysis, as the name implies, is simply an analysis of how results or perfor‑
mance develops over time. By comparing current results with past performance, it
is possible to form an opinion of the direction of the development and sometimes
even extrapolate to predict future levels. The latter is rarely very accurate, however,
and the main application is to spot trends to enable staying on top of any perfor‑
mance development, be it positive or negative. This is part of the early warning
dimension discussed earlier, as an indication of a negative trend usually appears
before the performance level reaches an unacceptably low level, thus allowing you
to counteract the development with improvements at an early stage.
There is really nothing fancy or complex about trend analysis—the main point
is to include a sufficient amount of historical data to create a credible trend image.
If not enough earlier measurements are included and minor fluctuations are inter‑
preted as a trend, you could make the unfortunate mistake of tampering with a
process that is displaying only normal variation.
Stepwise, this means:

1. Decide which performance dimensions or indicators to include in the anal‑

ysis and whether to construct one shared diagram for all the data or several
individual diagrams. The advantage of including several indicators in the
same chart is that it allows identification of any covariance between the
indicators; that is, they develop either in rhythm with or in ­anti-­phase of
each other. This can be important information, as there might be a ­cause-
­and-effect relationship between them.
2. When constructing the trend chart(s), time is assigned to the horizontal
axis, which places the most recent measurements at the right side of the
3. The performance level is tracked on the vertical axis by plotting the avail‑
able data to form trend lines.
Creating a Business Process Improvement Road Map  77

Delivery time

Number of new products


Number of bids won


Figure 5.1  Trend analysis for three example measures.

Figure 5.1 shows an example of a simple trend chart, where three measures
have been included. Measures that display a negative trend will obviously be
relevant candidates for improvement. As will be demonstrated a little later,
however, it is not only the development over time that impacts this decision,
but also how important the business process is for the competitiveness of the


A ­medium-­sized company that strongly emphasized both product devel-­

opment and efficient manufacturing of goods, mainly for export, had been
conducting performance measurement for quite some time. The company
was not certain that the measurements gave a correct picture of the situa-­
tion. Central performance measures had traditionally been:

• Manufacturing costs per unit, both for components and for finished
products, measured as the total running costs for machinery and
equipment divided by the number of manufactured units. This mea-­
sure had, in previous years, shown a steady increase, and much
effort had been put into reducing it.
• The costs for purchased parts present in the end product. This
measure had also been increasing, and many suppliers had been
replaced, without giving the desired results.

The company decided to start measuring a number of key performance

aspects every six months. Some important performance indicators were:

• True manufacturing costs per unit, measured as the manufacturing time

multiplied by an hourly rate for the machine or equipment in question
78  Chapter Five

• Price development for purchased parts

• The number of new components specified during product

After two years of measurement, a trend analysis of these measures

showed a completely different situation, as shown in Figure 5.2.
It appeared that the total volume, expressed as the number of products,
had decreased somewhat, as one product could now perform more than
one function. Thus, the total running costs for the manufacturing equip-­
ment had been distributed over fewer units, which resulted in the believed
increase in manufacturing costs. The fact was that regular improvements
in this process had decreased the costs. At the same time, the old per-­
formance measure of price development for purchased components had
increased because the product developer was now using more purchased
parts in the products. In fact, the prices for comparable components had
been continually reduced. The reason why the manufacturing costs and
the prices for purchased parts had not decreased even more was the con-­
tinual introduction of new components during product development. All
these matters were revealed through the analysis, and the improvement
effort was directed toward the product development process.

Manufacturing costs

Price development for purchased parts



New components during product development

6/05 12/05 6/06 12/06

Figure 5.2  Results of the trend analysis.

5.3  Spider Chart

Whereas trend analysis renders it possible to compare the current performance

level with previous measurements, the spider chart is an analysis tool offering
additional capabilities for graphically displaying your performance data. It is more
of a general chart type that has many applications, but in this respect, there are
primarily two useful ways to employ a spider chart:
Creating a Business Process Improvement Road Map  79

• To gain a quick overview of the performance levels for a number of different

performance indicators simultaneously, mainly to find which are in order
and which are lagging

• To compare the organization’s own performance level with that of other

organizations—a graphical presentation of benchmarking data

Figure 5.3 shows an example of a spider chart for the business process of
product development. Each spoke in the chart represents one performance indi‑
cator for this process. Obviously, it is also possible to let the spokes repre‑
sent processes and design one chart to represent an overall view of the entire
To construct a spider chart, follow this procedure:

1. Collect data from market analyses, surveys, competitor analyses, and so on.

2. Assign one variable to each spoke in the chart.

3. Divide each spoke into logical segments by using a separate unit of mea‑
surement for each variable. The farther from the center of the chart, the
higher the performance.

Performance level
Time to of the most important
market competitor

% new
Etc. products

Cost per
new product

Etc. Market

Own performance

Figure 5.3  Example of a spider chart.

80  Chapter Five

4. Plot the performance data for each variable along the correct spokes, using
different colors or symbols to separate data points from those of different
5. Draw lines between the data points for each organization to generate per‑
formance profiles.
6. Identify the variables that show the largest gaps between your organization
and the benchmarks.

By plotting the performance level of both your own organization and one or more
other organizations, you can see how good your organization is. Depending on
where the gap to the competitors is largest, you can select the business processes
that should be improved.


After realizing where improvements were needed, the company in the pre-­
vious example wanted to compare its status with the status of its com-­
petitors. Through gathering information from different sources, including
brochures and statistics from the competitors themselves, a spider chart
for comparison was drawn. The result is shown in Figure 5.4.

Own performance

Competitor B

Competitor A

New components Procurement

specified prices

Figure 5.4  Spider chart for the central performance measures.

Creating a Business Process Improvement Road Map  81

This confirmed that the efforts to reduce the manufacturing costs had
given results, but there was still a potential with regard to procurement
prices, perhaps by entering into more ­long-­term relationships with a few
suppliers. Furthermore, the performance level pertaining to standardization
of components in new products was dismal. Thus it was correct to start
improvement initiatives within this process.

5.4  Performance Matrix

The main problem with employing the spider chart for deciding which performance
indicators or business processes to improve is that all entries included in the chart are
assigned equal weight. From the spider chart, the mechanical approach is simply to
conclude that the process needing improvement has the largest gap to the benchmark‑
ing partners. To balance this picture, the performance matrix adds the importance factor
to the analysis. This supplements the two preceding tools, which focus on performance
alone. Using the performance matrix with the other tools can avoid wasting resources
to improve processes that are performing poorly but that are also not very important.
A generic performance matrix is shown in Figure 5.5. The matrix is divided
into four quadrants, with importance placed along the horizontal axis and the cur‑
rent performance level along the vertical. The performance indicators to be ana‑
lyzed are plotted in the matrix according to the performance data and a subjective
evaluation of their importance. The meaning of the quadrants is as follows:

• Unimportant (low importance, low performance): The performance level is

low, but the low importance renders it unnecessary to put any resources into
• Overkill (low importance, high performance): The performance level is
high, but this is of less consequence because the business processes in this
quadrant are not especially important to the organization’s competitiveness.
Thus, this is not a candidate for improvement.
• Must be improved (high importance, low performance): This is the obvious
area for starting improvements. The business processes that fall within this
area are important, but the current performance level is low.
• OK (high importance, high performance): A golden rule is that areas where
the performance is already good should also be improved. However, the busi‑
ness processes that in addition to being important are not being performed
well today (Must be improved) should be improved first. If no processes fall
within that quadrant, processes in the OK quadrant can be relevant candi‑
dates for improvement efforts.
82  Chapter Five

Current performance
Overkill OK

1 Unimportant Must be improved

1 5 9


Figure 5.5  A generic performance matrix.

The steps in using a performance matrix are the following:

1. Construct an empty chart by placing importance on the horizontal axis and

current performance on the vertical axis and dividing both axes into nine
segments of equal size.
2. Decide which business processes or other items to include in the analysis.
3. Place each business process in the chart according to its position along the
two axes, using symbols to identify each factor.
4. Divide the chart into four quadrants approximately at the middle of each
axis. If many items are clustered in one area, place the division lines far‑
ther to one side.
5. Determine which factors fall within the different quadrants.


A manufacturer of heavy mechanical equipment for marine applications had

identified six critical issues measured through ­ self-­assessment. All mea-­
sures showed a potential for improvement, but there were not sufficient
resources to start with all six of them at once. The company found it difficult
to decide where to start the improvement activity. To prioritize, a perfor-­
mance matrix was constructed that included the following six measures:

1. Advanced and flexible product design

2. Delivery time
Creating a Business Process Improvement Road Map  83

3. Delivery precision
4. Price
5. Accessibility for maintenance and repairs
6. Product design and finish

The matrix in Figure 5.6 helped the company decide to concentrate on

improving delivery time and delivery precision.

5 1
Current performance

5 3

1 5 9


Figure 5.6  Performance matrix for the six performance measures.

5.5  Criteria Testing

Before the criteria testing tool itself is presented, the term critical success factor
(CSF) should be defined. One definition is:

A limited number of factors that to a large extent impacts the organiza‑

tion’s competitiveness and its performance in the marketplace.

Typical examples of such CSFs are the prices, quality, and special features of the
organization’s products or services; reputation; and so on. If asking the question,
what is it that our customers truly value about our enterprise and helps maintain
them as customers? the answer will usually constitute the CSFs.
If, however, one tries to improve these directly to increase the competitiveness,
it can be difficult to determine where to concentrate the efforts. For example, deliv‑
ery precision might be an enterprise’s main competitive advantage, but the reason
84  Chapter Five

for what causes this and how it can be improved is likely not obvious. Criteria test‑
ing is a ­matrix-­based tool that assists in determining this. A typical form used for
criteria testing is shown in Figure 5.7.
The procedure for using criteria testing is as follows:

1. Place the identified CSFs, typically three to five, in the upper field of
the matrix. If desired, assign each of these a different weight factor that
expresses relative importance. In Figure 5.7, weight factors from one to
three are used, but other numeric values can be used as well.
2. In the ­left-­hand field of the matrix, place all possible business processes
assumed to have an impact on these factors.
3. For each business process, assess its impact on each of the CSFs. Again,
the example uses impact factors from one to three, where one is low impact
and three is high impact.
4. Multiply the impact factor by the weight factor of the CSF and place the
product in the correct matrix cell.
5. For each business process, add up the products and place the total in the
­right-­hand column of the matrix. This numeric value indicates the col‑
lective impact of the business process on the complete set of CSFs. The
higher the score, the better reason to improve this process, as this will give
the highest overall effect on the organization’s CSFs.

It should also be pointed out that the assessments made during the criteria testing,
both to assign weight factors and to determine the impact factors, are quite subjec‑
tive. The numeric values obtained at the end of the test should be viewed only as

CSF 1 2 3 4 5 Total
Processes Weight 3 1 1 3 2 score

Process 1 3 1 2 9 4 19
Process 2 9 3 1 3 2 18
Process 3 9 2 3 6 6 26
Process n 3 2 2 3 6 16

Figure 5.7  Matrix for criteria testing.

Creating a Business Process Improvement Road Map  85

directional and not as absolute answers. The more people who participate in the
test, the more valid the results are assumed to be.


A chain of hair salons initiated an improvement project to rectify its decreas-­

ing sales over a long period. The following CSFs were identified:
1. Price was important, but only as long as it was in line with that of
competing salons
2. Good appointment availability, that is, getting an appointment at
the desired time
3. The possibility for establishing a ­ customer-­stylist relationship.
Because of some recent staff changes, this had been difficult over
the past two years, and the customers had reacted negatively
4. Additional services such as a solarium and manicures, as requested
by many customers
These factors were weighted, and a criteria testing was performed. The
result is displayed in Figure 5.8. The conclusion was to concentrate the
effort around the three processes that ended up with the highest scores:
recruiting, scheduling, and accepting reservations.

CSF 1 2 3 4 Total
Business processes Weight 1 3 3 2 score

Providing haircuts and other services 3 3 0 0 6

Scheduling 1 9 6 0 16
Accepting reservations 0 9 6 0 15

Recruiting 1 3 9 6 19

Keeping the staff 1 0 9 0 10

Planning additional services 0 0 0 6 6

Purchasing accessories 3 0 0 2 5

Figure 5.8  Criteria testing for a hair salon chain.

5.6  Quality Function Deployment

The quality function deployment (QFD) tool was developed to represent a ­customer-
­oriented approach to product development. For this use, it is a methodology for
86  Chapter Five

structuring customer needs, expectations, and requirements and translating these

into detailed product and process specifications. The principles can, however, also
be used for a number of other problems, including improvement planning. Before
showing how QFD can be applied to improvement planning, the theory behind the
tool is presented. To start with, the product development process consists of several
sequential phases (Akao, 1990):

• Transforming customer requirements into a product concept

• Transforming the product concept into a product design
• Transforming the product design into a process design
• Transforming the process design into production documentation

Each step of this process must adhere to the original customer requirements. The
basic structure of QFD is a relational matrix at different stages of the process, as
shown in Figure 5.9.
What constitutes the goals of the analysis, which in the first phase of product
development are customer requirements and expectations. How expresses the means to
reach these goals, which are technical product concepts in the product development’s
first phase. In the next phase, these will form the Whats, and How will represent
detailed design solutions for the product concept. If a weight factor for each element
of What is multiplied by a grade indicating how well each How element satisfies
the requirement, an indicator for the performance of each How element is generated,
which is placed in the field How much. This way, each phase of the product develop‑
ment process is linked as a chain of relational matrices, as shown in Figure 5.10, to
ensure that the voice of the customer is transmitted throughout the entire process.
Additional information can be added, thus creating the chart popularly known
as the “house of quality,” as shown in Figure 5.11.
The QFD process is conducted by entering data into each room of the house of
quality. As was mentioned, What represents the external requirements—in prod‑
uct development these are the customer’s product requirements. A weight factor
expressing the element’s importance is attached to each element in What, which



How much

Figure 5.9  The basic structure of QFD.

Creating a Business Process Improvement Road Map  87


require- Product
ments design

Product Process
concept design

Product Production
design documents


Figure 5.10  A chain of QFD charts.

vs. How

How Why

What Points

How much

Figure 5.11  The house of quality.

renders it possible to emphasize some requirements more strongly than others.

Why represents the challenges facing the organization—for example, competitors’
solutions to these customer requirements. This can be expressed through a type of
benchmarking of different competitors’ offers and their quality or performance. An
example is shown in Figure 5.12.
After determining how the customer requirements can be fulfilled—that is,
How—the relational matrix linking What and How is completed. To make the
matrix as clear as possible, it is usually preferable to use as few types of rela‑
tions as possible. A set of commonly used symbols is shown in Figure 5.13. In
the same manner, the roof of the house of quality forms a relational matrix to
be used for investigating whether there are any relationships among the different
­elements of How. In this matrix, it is possible to indicate both positive and negative
relations—that is, factors that work together or that create ­trade-­offs or conflicts.
Some common symbols for this matrix are shown in Figure 5.14. For each How,
the weight for the determined relationship to the individual element of What is
88  Chapter Five

Poor Good
1 2 3 4 5 6 7

Our product
Company A
Company B

Figure 5.12  Benchmarking of others’ offers.

multiplied by the corresponding factor for importance for each requirement ele‑
ment. All products are summarized and placed in the lower field of the chart, How
much. Elements of How with a high score in this field should be preferred to others
if all elements cannot be incorporated into the product.
In improvement terms, QFD is a technique useful for ensuring that the overall
requirements and the organization’s strategy are maintained throughout the entire
improvement planning process, the same way the voice of the customer is main‑
tained throughout the entire product development process under ordinary use of
the tool. For this use, the steps of the process are the following:

Relation Symbol Weight

Weak 1
Medium 3
Strong 9

Figure 5.13  Symbols for the relational matrix.

Relationship Symbol

Strong positive
Strong negative

Figure 5.14  Symbols for the roof matrix.

Creating a Business Process Improvement Road Map  89

1. Identify the requirements to be used in the analysis and place these in the
Whats field of the house of quality. The requirements typically stem from
strategic decisions, stakeholder expectations, or market surveys.

2. Assign importance factors to each of these requirements.

3. Identify the possible business processes or other elements that are candi‑
dates for improvement projects, and place these in the Hows field.

4. For each How element, assess its impact on each of the What elements and
assign an impact factor, normally 1, 3, or 9.

5. Multiply the impact factor by the importance factor for the What element,
and write the result in the appropriate place in the relational matrix at the
heart of the house of quality.

6. If using the roof of the house of quality, go through each cell in the matrix
at the intersection of two What elements. Analyze whether the two ele‑
ments influence each other, and if so, place the appropriate influence sym‑
bol in the cell.

7. If using the benchmarking part of the house of quality, identify relevant data
that can be used and establish performance profiles for each data source.
Go through each What element and indicate the score of each organization
being benchmarked, including your own, on the performance scale. Com‑
plete the profiles by drawing a line along the various data points.

8. For each How element, add up the scores down the column of the rela‑
tional matrix and write the sum at the bottom of the Absolute importance

9. Calculate the percentage of the sum of the scores each How element
accounts for, and write the number below the scores in the Relative impor‑
tance row.

10. If using the assessment of how difficult it will be to undertake an improve‑

ment project for the different business processes, assign a difficulty factor
to each How element in this row.

11. Divide the scores for Absolute importance by the difficulty factor for each
How element.

12. Calculate the adjusted Relative importance scores.

90  Chapter Five


In a move to reduce the costs incurred for the mother company, PBI,
an internal IT department in a large company, was forced to take on
external customers. To position itself for this new market situation, the
IT department decided to use QFD to identify those business processes
it would need to excel at in the future. The resulting QFD chart is shown
in Figure 5.15.
First, the performance requirements were mapped using stakeholder
analysis and through reviewing the redefined strategy for the unit. Each
requirement was assigned an importance rating, and a limited competi-­
tive benchmarking was completed from insight into two competitors in the
external market.
Next, the identified business processes were inserted into the How
field. Impact symbols were assigned to each of the requirement/business
process combinations of the relationship matrix (empty fields indicate
no impact). The products of the importance ratings and impacts were
generated and summed up for each business process by using the cor-­
responding numerical impact factor for each of the impact symbols. Just
to make sure this is clear, let me explain this in more detail for one of the
business processes. For user contact and complaints handling, the cal-­
culation is as follows:

• There is an impact factor of 9 for average system ­up-­time, which

carries an importance rating of 10; this gives a product value of
• An impact factor of 1 for annual cost reduction, multiplied by the
importance rating of 5 gives a value of 5.
• An impact factor of 9 for external customers’ system ­ up-­time
requirements, multiplied by the importance rating of 9 gives a
value of 81.
• A factor of 3 for annual profit margin, multiplied by the importance
rating of 4 gives a value of 12.
• An impact factor of 9 for image development, multiplied by the
importance rating of 10 gives a value of 90.
• The sum expressing how much this business process is able to
impact the collective set of performance requirements is 278.
• Finally, relative importance figures were calculated in the last field of
the diagram, giving an impact rating of 18 percent.

The resulting relative importance figures for the entire set of business
processes made it much easier for the team to prioritize which busi-­
ness processes to focus on: operations, user contact, and customer
Business processes
or other activities
Own company
Company A
Company B

Performance requirements

1 2 3 4 5
Financial management


Portfolio planning and service design

Customer communication and sales
Human resource management

User contact and complaints handling

PBI Average system up-time of 97% 10

customers Annual cost reduction of 3% for all major services 5

Management/ Reasonable/innovative remuneration and benefits 9

employees Good working environment 7
System up-time requirements 9


customers Vendor respectability 5
Owners Annual profit margin of 6% 4
Develop image as reliable and cost-efficient supplier 10
Constantly introduce new technology/solutions 8
Capturing external clients for 25% of the revenues 5

Provide an attractive place of work 7

Absolute importance

Relative importance

Figure 5.15  QFD chart for an internal IT department.

Creating a Business Process Improvement Road Map  91
92  Chapter Two

After seeing the different tools that aid in prioritizing improvement ideas, you can
probably appreciate how they all supplement each other. They differ in focus from
trend to competitors’ performance to importance and more comprehensive sets of
requirements, and they are quite different in terms of complexity. As a contrast to
this chapter, which strongly focused on a selection of tools, the next chapter will
deal with “softer” issues in business process improvement.
Organizing for Business Process

t this stage in the book, and certainly judging by its title, you might have
formed the impression that I believe a good tool will fix any problem in
improvement work. Although I believe in the qualities of a good tool, I
hope I haven’t gone quite that far yet. There are a number of traits that set an orga‑
nization adept at continuous improvement apart from a merely adequate one. This
chapter is a “softer” one, focusing on some of these traits: improvement skills and
training in them, organizing for improvement, and stimulating an ­improvement-
­oriented organizational culture.

6.1  Business Process Improvement Skills

I’d like to revisit the brief discussion at the start of the book about the apparent
contradiction between continuous improvement and breakthroughs. While I often
find the frequent sports analogies in management literature to be misplaced, I can‑
not help but introduce one here.
Think of improvement capability as a muscle. As you know, a muscle must be
regularly exercised to stay fit and be able to perform. Athletes know this very well,
which is why they keep exercising, paying special attention to those muscle groups
that are critical in their type of sport. It would be highly unlikely to hear any world
or Olympic champions assign their victories to three months of doing nothing but
eating and smoking prior to the event.
The same argument can be made for the “improvement muscle.” If it is left dor‑
mant for long periods of time, it will be difficult to revive every time it’s needed.
For example, think of health clubs just after New Year’s. They are crowded with
overweight people with resolutions to get fit. You can almost hear their muscles
cranking back into life. If any of these individuals manage to stay the course, it will
take months before their muscles are functioning smoothly. Quite the opposite of

94  Chapter Six

a fit athlete with toned muscles and strong general form. Thus, when starting an
improvement project, make sure your improvement skills are ready to go.
To achieve this, some preconditions must be fulfilled:

• A sufficiently large portion of the organization must be trained in business

process improvement and involved in improvement projects. This is not a
skill reserved for top- or ­middle-­level management. The best improvement
projects I have observed always either include team members from all levels
of the organization or are run exclusively by employees at lower levels in
the organizational hierarchy.
• There must be a sufficient amount of improvement projects going on at any
given time so that employees are regularly involved in improvement work.
If three years pass from the time a person is part of an improvement project,
she or he will have to start almost from scratch every time. This is both a
mindset and a utilization of skills that takes time to get comfortable with, so
never falling out of it is the preferred option.
• The organizational structure must allow employees at all levels the freedom
to get involved in improvement efforts. This means freeing up time for par‑
ticipation in specific improvement projects and establishing arenas where
people can meet to develop ideas and share experiences.
• There must be incentives for engaging in improvement work. The organiza‑
tion depends on individuals or groups of people who initiate new improvement
projects, have enough passion for the ideas to see them through to approval,
and execute them properly. Usually, this comes in addition to regular respon‑
sibilities in the organization, and unless there are rewards for this extra effort,
people will not do it or will be quickly weaned from this type of behavior.

The latter two of these will be dealt with in later sections. For the first two, it’s
really a question of whether you want to have a constantly fit team of improve‑
ment players available for the next match, or you are content with having to train a
bunch of couch potatoes each time the need arises. In assuming the latter is not too
appealing, the only feasible mechanisms at hand are training and practice.
Training can take many forms and shapes. Traditional courses can be engag‑
ing and powerful, or they can be boring and easily forgotten. Training can take
place through ­experience-­sharing sessions among people who have been involved
in improvement projects. To allow people more freedom about when and where to
acquire knowledge, various e-learning approaches can be used. There have been
experiments to develop applications that resemble computer games and that are
used for transferring more tacit skills, and whose engagement factor makes them
very attractive. Irrespective of which training approaches are used, I would ide‑
ally like to see every single employee receive at least a modicum of insight into
improvement work.
Organizing for Business Process Improvement  95

6.2  Organizational Modes That Support Business Process


Because this book is not about organizational design, I will not outline the differ‑
ent philosophies for constructing organizations and what these entail. Rather, I will
address a few issues that have a direct bearing on the opportunities and climate for
business process improvement.
At issue is to what extent the business processes of an organization are visible
and appear in the organizational structure. We have already seen how traditional
departmental silos hamper flow and cooperation along business processes, but I
did not really offer an alternative. There is, in fact, an alternative—namely, struc‑
turing the organization along its business processes.
This can take many different shapes, but it usually involves some type of process
ownership in the organization. A common problem in organizations still focusing on
departments or just starting to orient themselves toward business processes is that
very few people feel responsible for the organization’s processes. There are usually
limited definitions of where one process ends and the next one starts; no one has been
given the responsibility of delivering the product or service from the process to the
customer, on time and of the right quality. At best, someone assumes responsibility
for the process, but then often more than one person does so simultaneously. The
result is that local “emperors” can be found around the organization, ruling over
segments of the process, but without having anyone fully responsible for the entire
process. A distinguishing feature of processes without ownership is that no improve‑
ments are initiated unless there exists a clear order from management to do so.
Process ownership means appointing people as owners of the individual busi‑
ness processes. Management must do the appointing, and a general rule is that the
person who has the highest organizational authority in the process or who impacts
the largest portion of the process should be its process owner. A process owner is
given a variety of responsibilities:

• To initiate changes to the process.

• To continually measure the performance level of the process.
• To initiate improvements in the process.
• To establish some type of steering group for the process that will work
together to improve it. This steering group will normally consist of people
who perform tasks within the process, but suppliers and customers of the
process are also relevant members.

After a company introduces process orientation and process ownership, the organi‑
zational chart will change character, from vertical silos with department managers
to horizontal processes with process owners. Examples of the two extremes are
shown in Figures 6.1 and 6.2.
96  Chapter Six


Marketing Sales Logistics Production Procurement Financial Personnel

manager manager manager manager manager manager manager

Department Department Department Department Department Department Department

co-worker #1 co-worker #1 co-worker #1 co-worker #1 co-worker #1 co-worker #1 co-worker #1

Department Department Department Department Department Department Department

co-worker #2 co-worker #2 co-worker #2 co-worker #2 co-worker #2 co-worker #2 co-worker #2

Department Department Department Department Department Department Department

co-worker #3 co-worker #3 co-worker #3 co-worker #3 co-worker #3 co-worker #3 co-worker #3

Figure 6.1  Traditional organization with vertical silos.

Organizing for Business Process Improvement  97


Process owner
Manufacturing and delivery of products to customers
Process co-worker Process co-worker Process co-worker Process co-worker

Process owner Process owner Process owner

Order handling Manufacturing Delivery and installation
Process co-worker Process co-worker Process co-worker

Process owner Process owner

Obtaining customer commitment Invoicing
Process co-worker Process co-worker Process co-worker

Process owner Process owner Process owner

Product development Personnel development After-sales service
Process co-worker Process co-worker Process co-worker

Figure 6.2  Organizing along business processes with process owners.

It should be pointed out that there will often be a hierarchy of business processes
and corresponding process owners. As can be seen in Figure 6.2, the extensive
process of manufacturing and delivery of products to customers has one process
owner coordinating ­high-­level issues related to this process. At the same time, this
process consists of several shorter processes at the level below, for example, order
handling, manufacturing, and delivery and installation. The individual process
owners for these subprocesses are responsible for their isolated processes, but they
must coordinate their activity with the process owner of the overall process.
I have seen examples of organizations that have taken this model to the extreme,
that is, disbanding all traditional departments and assigning all employees to one
or more process departments. Process departments are organizational units that
have been assembled not according to the principle of grouping people of similar
expertise, but rather to contain all the different types of expertise required to per‑
form the process in question. These units are headed by the process owner, who has
the human resources management responsibility toward the team members, and
sometimes they are shared by two or more process owners in cases where members
belong to more than one process team.
Although this organizational model carries inherent advantages, there are dis‑
advantages as well. Among these is a tendency not to have room for ­full-­time posi‑
tions in many processes, thus ending up sharing resources between processes. The
98  Chapter Six

line of authority toward these employees is often blurred; in many cases process
owners fight over employee capacity in situations of high workload. Furthermore,
some process team members claim that they miss colleagues with the same back‑
ground as themselves with whom they can discuss professional problems.
Quite a few organizations have tried to strike a balance between the two
extremes by designing a type of matrix organization. These still maintain the tra‑
ditional functional departments, where the department managers hold the human
resources management responsibility over the department employees. At the same
time, the business processes have been clarified, often even drawn into the orga‑
nizational chart, overlaying the departments. All employees in the departments
have further been defined to belong to one or more business processes, giving the
process owners authority to decide on tasks and how their time is spent. From what
I have seen and heard, this does not solve all problems; there are still cases of dif‑
ferent managers pulling employees in their own direction, but this model can be a
reasonable compromise.
These different models dictate primarily how employees are organized in daily
operations. Here, we are more concerned with how they function during improve‑
ment efforts. And if there is one lesson I’ve learned after working with many dif‑
ferent organizations over the years, it is that improvement work is a team sport.
Teams bring much more knowledge, ideas, and creativity to the table than does an
individual, and there is nothing as frustrating as struggling with an improvement
project alone. Frankly, there are no reasons why one single person should carry out
an improvement project.
There is no secret to composing improvement teams; it is a matter of finding
the right people and giving them the necessary mandate for their project. Regard‑
ing members, the team should fill the following roles:

• Team leader, that is, someone responsible for calling meetings, monitoring
the progress of the project, and making sure that minutes are prepared and
that necessary information is gathered and distributed. Many of these tasks
can usually be delegated, but the responsibility still remains with the team
leader. It is quite common that this role is given to the person in the team
with the highest position in the organizational hierarchy.
• Link to management, either through management representation in the team
or by giving someone authority to represent the management.
• Process owner or the person who in other ways is responsible for the whole
or most of the process to be improved.
• Other people involved in the process, preferably from a ­cross-­section of the
organizational units the process runs through. It is extremely important to
make sure that the team does not consist solely of people located above or
on the side of the process to be improved. If those who will later have to
Organizing for Business Process Improvement  99

change their routines are not involved in the team, making the necessary
changes might be hard to do.
• An internal or external customer of the process. Along with the process
view comes a commitment to viewing the customer of the process as the
most important part of the puzzle. By involving the customer in the team,
it is possible to make sure that the customer’s views are adhered to in the
improvement activity.
• An internal or external supplier to the process. In the same way that any
process has a customer, any process also has a supplier that provides some
sort of input. The supplier is to some extent a part of the factors that shape
the performance level of the process through the input being supplied, and
should therefore be included in the improvement activity.
• Possibly some kind of external assistance in cases where the improvement
tool to be used is new to the team.

It is possible for one person to take on several roles. From experience, a team
should not have more than six or seven members, but this depends on the scope
and complexity of the project. At the same time, care should be taken so as not to
establish a team that is too small, as this might hamper its ability to perform its
tasks, and it could undermine a broader sense of ownership of the project’s results.
The following requirements for candidates should be posed when selecting team

• Time to participate actively and wholeheartedly in the work. Instead of

including key personnel who never have time to participate, you should
choose people who will make an effort in the project. On the other hand,
there are usually some people you cannot do without, and these individuals
will have to set aside the necessary time.
• Competencies and knowledge about both the organization and the relevant
business processes, as well as training in the tool expected to be used. Alter‑
natively, training in the use of the tool must be provided at the outset of the
• Motivation to work on improving the selected process. A person selected
against his or her will is an unsuitable member of an improvement team.
• The ability to cooperate, listen, and communicate, as this is a typical team
effort that is less suited for the traditional silent loner.
• Credibility and respect in the organization, to ensure an impact when pre‑
senting results from the project and proceeding with effective implementa‑
tion of improvements.
100  Chapter Six

Finally, a warning must be issued against recruiting members on the basis of the think‑
ing, “You have nothing better to do; go participate in this improvement project.”


A manufacturing company with approximately 400 employees started an

improvement project connected to the procurement process and its link to
the more general logistics processes. An improvement team for the project
was established, and it consisted of the following people:

• Procurement manager
• Logistics manager
• Production planning manager

In addition, external individuals were hired to be part of the team. This

composition had several negative consequences:

• These three managers traveled quite extensively and were generally

rather busy, which made it difficult to arrange meetings or to have
any work done between meetings.
• The fact that they were quite high up in the organizational hierarchy
meant that they were less involved in the operative execution of pro-­
cesses. Thus, the knowledge about the processes within the team
was too low to enable any detailed documentation of the process or
generation of improvement suggestions.
• Each member of the team had a large degree of responsibility within
her or his own area and felt personally responsible for the perfor-­
mance level within it. It was therefore difficult to admit weaknesses
and problems, and very few constructive improvement suggestions
achieved broad support from the team.

It would have been wise to allow more operative personnel into the team.

In addition to these general improvement team guidelines, there are some more
permanent groups whose tasks are to continually work with improvement initia‑
tives. More specifically, two such groups are described here: quality circles and
­cross-­functional teams.

6.3  Quality Circles

Quality circles originated in Japan and were highly popular when the phenom‑
enon was first described in Western management literature in the 1970s and
Organizing for Business Process Improvement  101

1980s. They might still be used in Japanese organizations, but I have not been
able to uncover any statistics or documentation about this. Many academics
and consultants tried to import quality circles to the United States and Europe,
and some companies, after finding quality circles to be highly worthwhile, still
use them.
A quality circle can be said to consist of (Aune, 1985: 11):

• A group of people from the same work area (they have the same type of job
and experience the same problems) who;
• Under the guidance of a circle leader (in the beginning preferably a foreman
or the like);
• Voluntarily participate in regular meetings during normal working hours,
approximately one hour per week;
• To, according to their own priorities, identify, analyze, and solve problems
within their own work area, and;
• In writing and orally present suggestions for solutions with cost estimates to
a person authorized to decide on implementation.

The objectives of the work in quality circles are, on purpose, defined to be twofold:

• To strengthen the competitiveness of the organization through creating

• To create a good basis for development of the individual employee’s cre‑
ative skills and allow the use of these skills in practical improvement work

The standard approach is to establish several quality circles, each having its own
circle leader. Above these is usually a facilitator who can assist if some of the cir‑
cles get stuck and who also functions as a point of contact for the management of
the organization. Quite often, we can also find a steering committee for the circles
that coordinates the entire activity in this area. The committee usually consists of
members from management. When these different elements are put together, the
organization of the quality circle work can look like Figure 6.3.
I mentioned that a number of organizations in the United States and Europe
tried quality circles during the 1970s and 1980s. Studies have shown that there
are a number of reasons why quality circles have not worked too well in the

• The circles were not part of a larger quality program

• The circles never became an integrated part of the work in the organization
• The circles were introduced before the general consciousness level about
quality had become sufficiently high
102  Chapter Six

le Circ
Circ le

der Lea
Lea der













ad er
er ad
Leader Le
Ci e
rc rcl
le Ci


Figure 6.3  Organization of quality circles.

• The companies were more concerned with the financial gains achievable
than the working mode itself
• The work in the circles was perceived to be too bureaucratic, and all the
imposed rules seemed artificial
• As the circles succeeded in solving the smaller problems locally, they ran
out of problems they could attack

The last explanation is especially relevant in terms of how improvement work

is organized. There is often an evolution in organizations that take continuous
improvement seriously, from first addressing local problems within organizational
units to escalating to more complex problems. These are often attacked using
­cross-­functional teams.

6.4 ­ Cross-­Functional Teams

Such complex problems are rarely so unidimensional that they can be solved within
one organizational unit. In mature organizations that have worked with improve‑
ment for a while, it is not unusual that, at some stage, the remaining problems are
of a ­cross-­functional nature, thus requiring ­cross-­functional improvement efforts.
True improvement of such problems often results from efforts in the “white” zones
Organizing for Business Process Improvement  103

of the organization chart, that is, at the interfaces between departments or busi‑
ness processes. The answer to this challenge can often be ­cross-­functional teams
that possess broader competencies and, not in the least, organizational positions to
work on the case.
A ­cross-­functional team can be defined as follows:

A group consisting of members from different functional departments or

areas of responsibility and often also from different hierarchical levels.
The purpose of the team is to solve problems involving several organiza‑
tional units or business processes.

Cross-functional teams can be formed on an ad hoc basis, where a team is formed

to solve or investigate a specific problem and is thereafter dissolved, or they can
be permanent, like the quality circles. The normal and recommended size of such
teams is between four and eight people. There are also rules that govern the work
of ­cross-­functional teams:

• Meetings are held only when there is 100 percent attendance

• Meetings are held in “sacred” places, that is, not on the home turf of any of
the members and somewhere where there are no interruptions
• Meetings are held during regular working hours
• Before the team is established, management must agree to follow up any rec‑
ommendations made by the team, subject to realistic budgetary constraints
• Experts can be summoned according to needs

It should be expected that during the first meeting after the establishment of the
team, the work will be rather theoretical and characterized by the fact that the team
and its members are trying to find their places. Such discussions can seem a waste
of time and not leading to any solutions or improvements, but they help educate
the members and allow them to get to know each other. For each problem attacked
by the team, the result should be a list of improvement actions with related cost
estimates, a schedule for implementation, and indications of expected results.

6.5  Stimulating an Improvement Culture

When having the privilege of seeing different organizations from the inside and
observing how they work with improvements, I find that the single most defining
characteristic in truly ­world-­class improvers is the way the organizational culture
supports continuous improvement. There is almost a tangible quality to the atmo‑
sphere in these organizations that spells optimism, positivism, and a ­never-­ending
104  Chapter Six

quest for something better. I should also emphasize that this climate does in no way
seem desperate or frantic; these organizations have the ­ self-­confidence to know
they are already good, but they genuinely believe they can do even better. It is a
culture that instills confidence in those who come in contact with these organiza‑
tions, and more often than not inspires them into trying to copy it. Of course, this
is not easy. This type of culture cannot be implemented the following week; it is
carefully developed and nurtured over a long time (if anything, the only thing you
can do quickly is ruin such a culture by starting to behave inconsistently).
Studies have reported finding this type of organizational culture in highly suc‑
cessful organizations. The study Good to Great (Collins, 2001) identified a num‑
ber of companies that for a period of at least 50 years had been highly successful,
along with a control group of companies in the same industries that had performed
only averagely. The study team found that the successful companies shared some
similarities (which were found to contribute to either stimulating progress or pre‑
serving the core of the companies):

• BHAGs (big, hairy, audacious goals; stimulates progress)

• Cultlike culture (preserves the core)
• Try a lot of stuff and keep what works (stimulates progress)
• Homegrown management (preserves the core)
• Good enough never is (stimulates progress)

The latter of these capture the spirit I described earlier of ­world-­class ­improvement-
­oriented organizations and the relentless pursuit of ­ever-­higher performance levels.
So how do you create and maintain such a culture? This is a question impossible
to answer. There are probably a number of ways you can arrive there, all of which
have no quick fix. I cannot offer you a list of steps, either, but I hope the following
items can at least play an important role for organizations trying to stimulate such
an ­improvement-­oriented organizational culture:

• Encourage employees at all levels in the organization. When things are

working fairly well, with decent results being produced, it is tempting to
rest on your laurels. In such times, employees who launch criticism, point
out things that are not perfect, or simply suggest possible improvements can
be seen as killjoys. Realizing that following up such ideas will require time,
money, and energy, it is easy to kill the idea by dismissing it or asking that it
be postponed. This is a ­fail-­safe way of killing any improvement spirit that
might have developed. Always be positive toward such ideas and treat them
with the respect they deserve. In the end, it might very well be that the idea
was a bad one or that pursuing it was not cost effective. But what’s impor‑
tant is that you have shown that ideas are welcome and taken seriously.
Organizing for Business Process Improvement  105

• Celebrate victories. When improvement projects are completed, you should

mark the occasion. Those pioneering the project should be “shown off” to
the rest of the organization, and the results achieved should be announced.
The positive energy created through such success stories is invaluable.
• Never allow the organization to become overly content or complacent. In
celebrating the victories, make sure that the organization is not lulled into a
false sense of security and contentment. There is no organizational culture
that produces so few improvements as a complacent one, perhaps with the
exception of one that is complacent without any reason for being so.
• Expose the organization to continuous benchmarking. There are fewer
ways that are more effective in combating organizational complacency than
benchmarking. Comparing performance levels of your organization with
relevant benchmarks can shake any overly ­self-­confident organization out
of this unfortunate state. If you are hovering around three weeks average
delivery time for a certain type of product and you demonstrate to the orga‑
nization that a competitor consistently delivers in one week, chances are
good that improvement ideas will surface shortly (unless you allow people
to argue that the benchmark is not suitable or to give you numerous excuses
why you could never achieve similar results).
• Reward improvement achievements. Organizational rewards come in many
forms: honor and visibility, various perks, money (bonuses or pay raises),
and promotions. All these can help motivate people to improvement efforts,
including those who have already achieved results and are rewarded for this,
and those who see that creating improvements pays off. If you have a perfor‑
mance measurement system in your organization, data from this system can
often be useful as input into reward systems. A word of warning is warranted,
though. Researchers argue about the effects of different types of rewards,
some even claiming they achieve very little. Unfairly awarded rewards might
have negative effects on motivation, and awkwardly designed reward sys‑
tems can encourage undesired behavior throughout the organization. This is
especially true for individual rewards, which can stimulate an egocentric or
selfish culture that hampers teamwork and business process thinking.

At this point in the book, I am happy to announce that I have presented all the
introductory material preceding the actual business process improvement toolbox.
We will now move on to the more detailed improvement tools that you can use to
create these coveted improvements.
The Business Process
Improvement Toolbox

long with the business process improvement toolbox, this chapter includes
an overview of the different phases in improvement work and the tools
that belong to each phase. A far more detailed treatment of each tool is
given in the ensuing chapters.

7.1  The Need for a Toolbox

It is inherent in the nature of tools that their purpose is to offer support. For exam‑
ple, if you want to fix your car, you select and use the tools that will get the job
done. The desired result is a car that runs, and it doesn’t matter whether a wrench,
a pliers, or a screwdriver was used. In the same way, the desired results for organi‑
zations are improved performance and competitiveness. For this purpose, they can
use whatever tools work, as long as they give results. It is thus far more important
to know how to select the right tool and use it to reach the set goals than to know a
large number of different tools.
At conferences, seminars, and other forums, claims have been made regarding
the use of a particular tool. These claims are often followed by an intense debate.
Such statements and allegations show that too much emphasis has been placed
on the tool itself and not on its purpose. The fact is that all the tools presented in
this book pose different requirements to the user, they serve different purposes,
their use and results depend on the situation in which they are applied, and they
give different results. Thus, it is not possible to allege that any one tool is best.
All these tools should be seen as belonging to a large and ­well-­equipped tool‑
box available to the organization striving to improve. They should supplement each
other and function in symbiosis. Which tool is used in a specific situation depends
on the characteristics of the organization using it and the situation in which it is
applied. First, the situation and problem at hand must be defined, and then a suitable

108  Chapter Seven

tool selected for it. If we have only one tool—say, a hammer—it is amazing how
quickly all problems come to look like nails. The purpose of this book, therefore, is
to present an extensive toolbox for the business process improvement handyman.

7.2  A Business Process Improvement Model

I presented a framework in Figure 1.2 where I tried to illustrate the different fac‑
tors that are required for or influence business process improvement. We have now
dealt with all the items in the framework except for the core activity of improving
business processes. To make it more manageable, it is useful to divide the larger
task into a number of phases that improvement projects typically go through. This
can be done in many different ways, and I assure you it is not entirely easy to cre‑
ate a logical flow of activities in such a model. I have tried to reorganize the stage
model presented in the first edition of the book, ending up with the one presented
in Figure 7.1.
One of the problems in designing such a model, which by nature must be por‑
trayed as linear, is that such processes are rarely so neat. Improvement teams will
find themselves skipping phases, retracing their steps, reiterating work, and so on.
I hope you will see these stages merely as illustrations of some distinct groups of
tasks or mindsets during such projects. The seven phases of this model have the
following objectives:

1. Develop improvement priorities based on an overall performance under‑

standing. This is a preliminary stage that precedes specific improvement
projects, and the aim is to understand which business processes will be
important to improve, given the organization’s situation, stakeholders, and
2. Understand the current business process and the performance shortcom‑
ing in question. This is normally the first step in an improvement project,
where the objective is to understand how well the process works at the
moment and what seems to characterize the problem experienced.
3. Collect data about the performance shortcoming. Business process
improvement work is at its best when it is based on facts and data, and the
purpose here is to gather relevant information for the ensuing analyses.
4. Analyze the performance shortcoming. In this stage you apply different
techniques to try to understand the true nature of the problem causing the
­less-­than-desired performance.
5. Generate ideas about causes of the performance shortcoming and possible
improvements. This is a creative phase to generate a broad range of ideas
The Business Process Improvement Toolbox  109

7. Implement the

6. Develop improvements to
remedy the performance

5. Generate ideas about causes

of the performance shortcoming
and possible improvements

4. Analyze the performance


3. Collect data about the

performance shortcoming

2. Understand the current

business process and the
performance shortcoming
in question

1. Develop improvement
priorities based on an overall
performance understanding

Figure 7.1  Stages in business process improvement work.

about what causes the performance shortcoming, but it is also a selec‑

tion of tools that can be used in other phases of the improvement process.
In Chapter 10, I have included some techniques for choosing among the
resulting ideas and creating consensus in the improvement team.
6. Develop improvements to remedy the performance shortcoming. To many,
this is the exciting stage of the process, where the objective is to create
new solutions or process designs that will eliminate the performance
7. Implement the improvements. While perhaps sounding easy compared
with finding ways to improve the process, implementing lasting change
can often be the hardest part of an improvement project.
Table 7.1  Tools for different phases in the improvement work.
Develop current process Collect data about Analyze the Generate ideas
performance and performance the performance performance and choose Develop Implement
priorities shortcoming shortcoming shortcoming among them improvements improvements
Stakeholder Relationship Sampling Critical incident Brainstorming Streamlining AΔT analysis
analysis with mapping
Kano model
110  Chapter Seven

SWOT analysis Traditional Surveying Pareto chart Brainwriting Idealizing Tree diagram
flowchart and process
program chart
Competitive Cross- Check sheet Cause-and- Crawford slip QFD Force field
forces analysis functional effect chart method analysis
Strategy map Flowchart Problem Five whys Nominal group Statistical
divided into concentration analysis technique process control
process diagram
Trend analysis Several-leveled Scatter chart Six thinking Six Sigma
flowchart hats
Spider chart Flowchart with Histogram Paired Business
statistics comparisons process
Performance Relations Benchmarking
matrix diagram
Criteria testing Matrix diagram
QFD Is–is not
The Business Process Improvement Toolbox  111

7.3  The Tools in the Toolbox

Within each of these phases are a number of tools and techniques that I have
found to be useful and possible to apply without a PhD. This selection of tools
can certainly be discussed. Some people will find that there are tools missing that
they think should be there, and others will probably try a tool and find that they
don’t like it. I accept this, and I am certainly open to feedback and suggestions for
The world is, fortunately, not sufficiently square that all the tools can be singu‑
larly defined for each criterion and thus placed in the right category. The classifica‑
tion has been carried out according to the most prominent features of the tools and
in complete confidence that exceptions exist.
The presentation of the tools throughout the book will follow the order of the
stages, starting with stage 3, as the first two have already been dealt with. For each
stage, Table 7.1 lists the tools and techniques included. Depending on which phase
the project is currently in and which task you want to solve, this table can serve as
a guideline for what tools are available. For further advice on which tool will work
in a given situation, I can offer no additional help other than to experiment. There
is no doubt that some of these are more complex, require more skills to apply, and
take longer to complete; additionally, the degree of change or innovation they con‑
tribute to also varies. However, it seems impossible to draw up general rules about
these issues. I suggest trying different ones and seeing which work for you.
Tools for Collecting Data about the
Performance Shortcoming

t this stage of improving a business process, you will typically possess a
process model of some sort for the business process you have undertaken
to improve. This process has been selected because it displays a perfor‑
mance shortcoming of some sort. You might even think you know fairly well what
the problem is and how it can be fixed. If so, a word of caution is in order: ­spur-
­of-the-moment inspiration is rarely the best starting point for your improvement
project. If there is one thing most authorities in the area of quality management
agree on, it is the importance of ­fact-­based decision making—knowing the situ‑
ation at hand well enough to make sure you proceed in the best direction. In this
chapter, you will learn about different tools and techniques for collecting data and
information about the performance shortcoming you are facing. There are four key
approaches and tools in this group:

• Sampling
• Surveying
• Check sheet
• Problem concentration diagram

8.1  Sampling

Sampling is not a specific tool per se, but rather a useful principle to know about before
collecting data. While it might sound trivial, data collection can be a ­time-­consuming
and expensive task. Sampling is simply a way of economizing the ­data-­collection pro‑
cess. The purpose is to allow collection of a smaller amount of data while maintaining a
data set with sufficient validity to draw conclusions about the situation being studied.

114  Chapter Eight

There are many ways to determine how the sample should be collected. Some
of the most common types include the following:

• Random sampling: In this method, random numbers determine which data

to draw from the full population. You can find such random numbers in
published tables, by using a computer to generate them, or simply by throw‑
ing dice. An example is pulling out customers 4, 11, 19, 21, 34, and so on,
among the first 1000 to be served during the month and asking his or her
• Systematic sampling: In many cases, it is difficult or even impossible to
pick out random samples. In systematic sampling, data are collected at fixed
intervals of time, numbers, length, and so on. For example, every 20 min‑
utes the number of customers waiting in line is counted.
• Stratified sampling: For some business processes, there are differences
between categories within the entire population. In this case, you should
purposely collect data from each of the categories so that the samples rep‑
resent the categories in correct proportion to one another. If a company has
seven salespeople, customer satisfaction scores can be collected from the
customers of each, relative to the number of customers each serves.
• Cluster sampling: If you know that the population is stable and without
much variation over time, cluster sampling is the simplest way of defining
the sample. In this case, a group of the units is taken to represent the whole
population—for example, the entire batch of parts produced in one hour to
represent the entire week’s production.

As I’ve already mentioned, sampling is not one unified tool, which means it is dif‑
ficult to define generic steps. Rather, sampling is used to support other tools, but
there are some important issues to keep in mind:

1. Assess the nature of the population to be sampled to decide on a suitable

sampling approach (take into account the homogeneity of the population,
any clustering of data, and so on).
2. Decide on the sample size, which can be a tricky task. For help with this
issue, consult available standards that specify sample size and frequency.
Such standards are available as military standards, ANSI standards, and
ISO standards. For a list of relevant standards, see http://www.variation.
3. Collect the sample of data according to the chosen sampling approach.
4. Calculate simple figures such as averages and means so you can test
whether the sample is a reasonable representation of the population.
Tools for Collecting Data about the Performance Shortcoming  115


A large hospital had experienced several outbreaks of infections, and these

had occurred in many different units across the hospital. A doctor looking
into the problem found that the presence of a certain type of bacterium
seemed to predict such outbreaks quite accurately, and that this warning
came in time to take measures to avoid or at least reduce the extent of the
It was therefore decided to start testing for this bacterium, with the
hope of being able to stop many outbreaks before they could take hold.
In the beginning, the idea was to test extensively in every unit, but it soon
became clear that the costs and disruptions were almost a worse problem
than the consequences of the occasional infection outbreak.
A statistician was consulted and asked to design a sampling plan that
would be less of a burden on the personnel and patients, but at the same
time would be effective in warning of future infections. She came up with
a cycle plan in which each unit would do tests once every second week.
After this procedure was implemented, the test proved capable of pre-­
dicting several outbreaks, and many of them were thwarted altogether by
speedy action.

8.2  Surveying

Sampling gives you advice about how you can collect less data and still be reason‑
ably sure you capture the full picture of the performance shortcoming. In many
cases, the data you need to collect are precise and specific—for example, the
dimensions of a physical product, the color of a chemical test on a color scale, or
the costs of different aspects. However, in many business processes, information
about people’s attitudes, feelings, or opinions is more important. Such information
cannot be collected using tools suitable for physical dimensions or pH values. For
these matters, surveying is used to collect data.
Before looking at the steps for conducting a survey, it should be mentioned
that there are several different ways of conducting surveys. First of all, the basis
for a survey is normally a questionnaire of some sort. The questionnaire outlines
the questions to be asked of the respondent, and depending on how the answers are
collected, the questionnaire can come in many different forms.
There are at least two different ways of obtaining answers from the respon‑
dents. First, the questions can be presented in writing and the answers collected the
same way. For this approach, you can use a questionnaire that is mailed or handed
out personally, or you can use an electronic one, sent by e-mail or published on a
Web site. Second, the questions can be orally transmitted, and the respondents can
answer back orally. This can take place in traditional interviews, either ­face-­to-face
116  Chapter Eight

or over the phone, or managed by a computer that reads questions to respondents

over the phone and records their answers. In all types of surveys, large amounts
of data can be collected relatively easily and inexpensively. While questionnaires
filled in by the respondents generate the most data per dollar, interviews tend to
give data of higher quality.
To conduct a survey, follow these steps:

1. Clearly define the objective of the survey and how the data will be used
2. Determine what information is required to achieve this objective
3. Decide how the survey will be undertaken—that is, written (via mail, fax,
e-mail, or the Internet) or verbal (by telephone or in person)
4. Develop the questionnaire, keeping in mind issues such as type and
sequence of questions, understandability, language, grouping of questions,
brevity, and so on
5. Test the questionnaire to ensure that all questions are easy to understand
and can measure what they are intended to
6. Identify the sample of respondents
7. Perform the survey according to the chosen approach


A private retirement home had seen the number of vacant rooms steadily
increase over the last few months, and informal talks with both residents
and their next of kin indicated that many were dissatisfied with the services
offered. However, these brief talks, often just when passing each other in
the hallway, gave little specific insight into the cause.
The management decided to undertake a more thorough survey to gain
a better understanding of how it could turn the trend. It was decided that
it was necessary to survey both the residents and their relatives, especially
since the latter group often acted as legal guardians of the residents and
made the decisions about moving them.
Two questionnaires were developed, one for the residents and one for
the relatives. Both were short, two pages, and asked the respondents to
rate factors like the comfort of the rooms, cleanliness, quality of the meals,
medical care, ­in-­house activities, excursions, price, and so on. In a period
of two weeks, about 30 questionnaires were collected from each respon-­
dent group. Some findings stood out:
Tools for Collecting Data about the Performance Shortcoming  117

• Meals were considered nutritional but rather boring, with little vari-­
ety from week to week, and this was one reason why quite a few
residents ate less than they should.
• The residents in poorest health found few recreational activities
that suited their abilities, and those in good health said the same. It
seemed the activities had been designed for an average that didn’t
• Relatives found the visiting hours to be too limited, often making it
difficult to find time to come see their relatives.
• Invoicing three months’ rent each quarter meant each bill was rather
large and caused problems with payment.
For many of these issues, changes could be made that would cost very
little money. A more extensive meal rotation cycle was developed with
the help of a local restaurant, the activities program was changed to
accommodate the fitness levels of all residents, visiting hours were
extended, and bills were issued monthly. A few months later, there were
no vacant rooms, and both the residents and the relatives seemed much
more content.

8.3  Check Sheet

A check sheet is a table or form used for registering data as they are collected. One
of the main applications is registering how often different problems or incidents
occur. This provides important information about problem areas or probable causes
of errors, and thus provides a good foundation for deciding where to concentrate
during improvement.
The approach is normally as follows:

1. Agree on what events are to be recorded. These must be clearly defined

to avoid doubt whether an event truly occurred. It is usually also smart to
include a category of “other” to capture incidents that are not easy to oth‑
erwise categorize.

2. Define the period for data recording and a suitable division into intervals.

3. Design the check sheet to be used during recording. An example is

shown in Figure 8.1, where space is allocated for recording each event
and summarizing both within the intervals and for the entire recording
118  Chapter Eight

Total number of
Problem Week 1 Week 2 Week 3 per problem
A 27
B 4
C 23
D 53
E 9
F 2
G 2
H 14
Total number of
problems per week 36 55 43 134

Figure 8.1  Example of a check sheet.

4. Perform data collection during the agreed period. In advance, it is necessary

to make sure that everyone taking part in the data collection has a common
understanding of the task, so as to achieve consistency in the data material.
5. When data collection is completed, analyze the material to identify events
displaying a high number of occurrences. These will contribute to the
prioritization of what specific areas within the chosen business process
should be emphasized in the ensuing improvement work. A suitable visual
aid for this analysis is a Pareto chart (see Figure 9.2).

The danger of overlooking one or more categories of problems should be men‑

tioned. If we are exclusively looking for events of the type defined in the check
sheet, other problems might occur but fail to be recorded because the attention is
directed only toward the expected problems. This is partly countered by the inclu‑
sion of an “other” category, but this issue should still be kept in mind.


A ­ medium-­sized company in the electrical installation industry annually

submits a large number of bids for jobs for both private and industrial
customers. The company was not satisfied with the portion of bids that
resulted in jobs, and therefore wanted to follow up and improve the pro-­
cess of generating bids.
To obtain an overview of the reasons why the company was not assigned
more jobs, the check sheet shown in Figure 8.2 was designed. Each time
a bid was rejected by a potential customer or the bid of a competitor was
Tools for Collecting Data about the Performance Shortcoming  119

preferred, the customer was asked to explain the reasons. The responses
were entered into the check sheet during a period of three months.
As the sheet indicates, price was not the dominating reason for rejec-­
tion, contrary to what the company had believed. Rather, the problem was
very often a lack of flexibility with regard to when the job could be carried
out. The result was that a new system for monitoring the availability of the
service workers was designed. In addition, it was made clear in every sin-­
gle bid that the work could be started on very short notice and could very
well be carried out in several separate periods, all at once, during evenings,
or whenever. In the long run, this resulted in a radical increase in the portion
of bids won, and at substantially better prices.

Total number of
Cause of occurrences
lost bid January February March per cause
Too high price 7
Poor quality 4
Low flexibility 40
Poor impression
during inspection
Low technical
Total number of
causes per month 23 20 21 64

Figure 8.2  Check sheet for recording the causes of lost bids.

8.4  Problem Concentration Diagram

The check sheet presented in the previous section is a practical tool to collect data,
especially about occurrences. When periods of time, areas of occurrence, or some
other suitable headings have been defined for the columns in the check sheet, it can
easily be used to record the desired data. However, in some cases, it is not that easy
to label the columns properly.
Say, for example, a rental car company is interested in determining where
scratches and dents on its cars are most often located. To collect such data, a check
sheet could perhaps be used, listing categories as hood, roof, trunk, different doors,
and so on, as possible locations. However, the data could be too rough. Whenever
a door has been indicated, there is no telling where on the door the damage was
done. In such a case, a problem concentration diagram can offer more insight, and I
am sure you have seen an example of such a diagram if you have ever rented a car.
120  Chapter Eight

When you pick up the car, the keys are typically in a small paper bag that contains
fields for recording the time of return and the mileage (granted, there is a transition
away from these ­paper-­based forms to electronic scanning upon rental car return,
but the example is still valid), as well as a diagram of the car from different angles.
If the car already has some damage, this is marked in the diagrams, and any new
damages are duly recorded upon return. These diagrams can then be analyzed to
see where scratches and dents show up most often, possibly as a basis for installing
extra protection of these areas—for example, door protectors.
It is possible to combine a check sheet with a problem concentration diagram.
If the check sheet is suitable for data collection, the data recorded can be trans‑
ferred to a problem concentration diagram to be analyzed further. In this case, the
diagram is perhaps more of an analysis tool that can be used when investigating the
performance shortcoming in more detail.
The steps in using problem concentration diagrams are the following:

1. Clearly define what events are to be recorded, and assign a symbol to each
type of problem or event
2. Define the period for data recording and suitable intervals
3. Design the diagram by drawing a map of the system, object, or area
4. Populate the diagram with problem registration data through recording
problem occurrences directly into the diagram


A ­medium-­sized city had been constantly criticized over the last few years
for inadequate street lighting. People felt insecure in areas where the light-­
ing was poor, and the media claimed that both crimes and traffic accidents
could be linked to dark or poorly lit areas of the city.
The city’s technical operations department was in charge of the street-­
lights. It was equally frustrated with the problem, both because of the extra
work from the constant repairs and because of all the negative attention
to the situation. The paradox was that it had never changed as many light-­
bulbs or fixtures per month as it did in the last two years, effectively having
increased productivity significantly by appointing a dedicated streetlight
team with specialized vehicles at its disposal.
There were extensive discussions about what caused so many bulbs
or lamps to fail. Vandalism had been ruled out early on; only a few broken
lamps were due to rock throwing or gunshots. In many cases, the bulbs
simply went out much sooner than their estimated life expectancy; in other
cases, the fixtures showed signs of burning. The bulbs were of the same
brand, purchased from the same manufacturer as always, and seemed of
consistent quality.
Tools for Collecting Data about the Performance Shortcoming  121

Truly puzzled by this enigma, the streetlight team decided to record

every defective lamp that was fixed. This was done using a problem con-­
centration diagram in the shape of a large city map with all streetlights
marked. All the repair teams had such a map in their cars, and whenever
a lamp was fixed, it was marked on the map. The teams recorded some
400 defective lamps, and all lamps were displayed in the same map, with
columns showing the number of lamps with defects in each area of the city.
A limited and simplified cutout of this map is shown in Figure 8.3.
This small part of the map may not show this very well, but the full
version displayed extreme differences in concentration of defective lamps.
Many areas had virtually no broken bulbs or fixtures, but others seemed
to suffer from downright epidemics. At first, this gave the team no further
clues as to what caused these anomalies, but almost by accident, an engi-­
neer from the local power company came upon the map. He immediately
saw that the map quite accurately outlined the areas where new generators
had been installed over the last two to three years. From tests the power
company had undertaken, he knew that the voltage levels in these areas

Town hall

Gas station


= Defective lamps
Train station

Figure 8.3  Cutout of city map with streetlight problem concentration.

122  Chapter Eight

often ran higher than with the old generators. Further inquiries revealed
that the fixtures used in the streetlights were vulnerable to elevated voltage
levels, causing them to overload and break or the bulb to burn out.
There was no ­short-­term solution to this problem, because there was
no easy way of reducing the voltage delivered to the lamps. However, when
fixtures were replaced, a new type was installed that was less sensitive. In
the long run, this should reduce the problem.

The data available about the performance shortcoming, whether they are col‑
lected through the use of one of these tools or already existing, are now ready for
Tools for Analyzing the
Performance Shortcoming

here are many possible tools that could be included in this section of the
book. Therefore, I have selected a range from widely used, rather simple
tools to less ­well-­known techniques:

• Critical incident
• Pareto chart
• Cause-and-effect chart
• Five whys analysis
• Scatter chart
• Histogram
• Relations diagram
• Matrix diagram
• Is–is not analysis
• Fault tree analysis

9.1  Critical Incident

Critical incident is a technique that can be used for identifying a process, sub‑
process, or problem that should be improved (Lawlor, 1985). After you’ve com‑
pleted this task and are in progress with an improvement project, it is also useful
for figuring out what is causing the performance shortcoming.

124  Chapter Nine

It is a quite open and frank way of seeking information about organizational

problems. The technique relies on honest responses from those involved, and a
prerequisite is that the participants are completely free to express their views. In
this respect, it is important that management display the right attitude to avoid
censorship or withholding of information in fear of consequences from being too
The technique works as follows:

1. First, the participants in the analysis are selected, and this should include
people actively involved in the business process to be improved.
2. Next, the group of participants is asked to answer questions such as the
• Which incident last week was the most difficult to handle?
• Which episode created the biggest problems in terms of maintaining
customer satisfaction?
• Which incident cost the most in terms of extra resources or direct
The purpose is to focus on the critical incidents that, in one way or another,
created problems for the employees, the organization, or other stakehold‑
ers. The period covered by the questions can range from a few days to
several months. It is, however, not favorable if the period is too long, as it
might be difficult to determine the most critical incident simply because
many incidents can qualify as candidates given a sufficient length of
3. The collected answers are sorted and analyzed according to the number
of times the different incidents have been mentioned. A graphical repre‑
sentation format might very well be used for this purpose. The incidents
occurring most often, such as the critical ones, are obvious candidates for
prevention of recurrence.


A large corporation started a project to improve customer service in its staff

of 15 switchboard operators. It was difficult to decide where in this area to
start, so it was decided to try the technique of critical incident.
All the operators were asked to describe the most difficult situations
they had experienced the last month. The resulting events were sorted
by frequency and presented in a chart, as shown in Figure 9.1. As the
Tools for Analyzing the Performance Shortcoming  125

chart shows, the most critical incidents were “the wanted person does not
answer” and “the operator does not know who should handle the request.”
This resulted in efforts focused around designing a system to keep track
of the individual employee and drawing up clearer rules for who should
handle which requests.

Number of


15 14


5 4

The wanted person does not answer,
and no absent note has been entered

The operator does not know

who should handle the request

The call disappears

during transfer

Calls that have been directed are

not transferred to the right extension

Other employees interrupt by

coming into the office

Figure 9.1  Critical incidents for the switchboard operators.

9.2  Pareto Chart

The Pareto chart is based on the Pareto principle, formulated by the Italian
mathematician Vilfredo Pareto during the 1800s. He was concerned with the
distribution of riches in society and claimed that 20 percent of the population
owned 80 percent of the wealth. Translated to modern quality terminology,
the Pareto principle states that most of the effects, often around 80 percent,
are from a small number of causes, often only 20 percent. For example, usu‑
ally 80 percent of the problems related to purchased material are caused by
126  Chapter Nine

20 percent of the suppliers. Even more importantly, 80 percent of all costs

connected to poor quality or generally low performance is from 20 percent of
all possible causes. A healthy approach is therefore to start the improvement
work by attacking these 20 percent, often labeled “the vital few.” This does not
imply that the remaining 80 percent should be ignored; “these important many”
should in due time also be addressed. The Pareto principle only says something
about the order in which the problems should be attacked.
The Pareto chart is a tool used to graphically display this skewed distribu‑
tion, the 80–20 rule. The chart shows the causes of a problem sorted by the
degree of seriousness and expressed as frequency of occurrence, costs, perfor‑
mance level, and so on. The causes are sorted by placing the most severe on
the left side of the chart, rendering it quite easy to identify the vital few. To
enable portraying additional information in the chart, it is common to include
a curve showing cumulative importance. This is depicted in the general chart
in Figure 9.2.
The following are the most common steps in using a Pareto chart:

1. Define the performance shortcoming and the potential causes of it.

2. Decide which quantitative measure to use when comparing the possible
causes. As has been mentioned, this measure might be how often the dif‑
ferent problems occur or consequences of them in terms of cost or other
3. Use existing data or collect the necessary data.
4. Place the causes from left to right along the horizontal axis in descending rela‑
tive importance. Draw rectangles of heights that represent this importance.




Figure 9.2  A general Pareto chart with a line for cumulative importance.
Tools for Analyzing the Performance Shortcoming  127

5. Mark the data value on the left vertical axis and the percentage value on
the right, and draw the curve for cumulative importance along the top edge
of the rectangles.

A quick inspection of the chart can now answer questions like, what are the two
or three main causes of the low performance level of this process? or how large a
portion of the costs can be attributed to the most vital causes? This information can
be used for actively directing the improvement efforts toward areas that are likely
to produce the best effects. Pareto charts can easily be constructed using standard
spreadsheet software.


The installation company in the check sheets example also drew a Pareto
chart displaying the collected data. Instead of how often different reasons
were given for why the company’s bid did not win, the value of the lost job
was assigned to the vertical axis. The resulting chart is shown in Figure 9.3.
The chart confirmed that the low flexibility expressed to the customers
regarding how quickly and at what time of day the work could be done not
only was the most frequent reason for failing to win bids, but also caused
the biggest losses in terms of ­lost-­job value. It was, in other words, right to
aim for a new system for planning the work in the company.

Lost-job value


Low flexibility

High price

Poor impression
during inspection

Poor quality

Low technical

Figure 9.3  Pareto chart for value of lost jobs.

128  Chapter Nine

9.3 ­ Cause-­and-Effect Chart

The cause-and-effect chart is one of the classical and most widely used tools in
quality management, and not without reason. Many view the tool as being some‑
what ­old-­fashioned and sturdy in a rather boring sense, but it is truly a useful tool
with many powers. The main purpose is, as the name implies, to identify possible
causes of an effect. The effect being analyzed can be an experienced problem or a
future ­hoped-­for state where the causes no longer occur.
There are at least two types of ­cause-­and-effect charts:

• Fishbone chart
• Process chart

The fishbone chart is the traditional way of constructing such charts, where the
main product is a chart whose shape resembles a fishbone. The principles for this
chart are shown in Figure 9.4. There are two ways of creating the chart:

• Dispersion analysis, where the effect being analyzed is drawn on the right
side of the chart, at the end of a large arrow. Main groups of probable causes
are drawn as branches to the arrow. For each branch, all possible causes are
• Cause enumeration, where all probable causes are simply brainstormed and
listed in the order they were generated. When this has been done, the causes
are grouped and drawn in the fishbone chart.

The end product is the same regardless of the approach.

Main Main
category category

Cause Cause
Cause Cause
Cause Cause to be
Cause Cause solved
Cause Cause (effect)

Cause Cause
Main Main
category category

Figure 9.4  The structure for a fishbone chart.

Tools for Analyzing the Performance Shortcoming  129

The main steps for producing a fishbone chart using dispersion analysis, which
is the more common approach, are the following:

1. Assemble a suitable group of individuals who possess the necessary

knowledge about the area to be analyzed.
2. Clearly describe the effect for which causes are sought. This effect is often a
low performance level for one of the business processes of the organization.
3. Using a white board or some other large medium, draw the effect at the
end of a large arrow. The point is to set aside enough space for the gener‑
ated causes, not symmetry or nice drawing effects.
4. Identify the main categories of possible causes of the effect and place these
at branches emanating from the large arrow. For physical processes, some
common main categories are:
• People
• Machines and equipment
• Materials
• Methods
• Measures
• Environment—culture, organizational structure, physical environment,
and so on
Similarly, for service processes, the traditional categories are:
• People
• Processes
• Framework conditions
• Work environment
5. Brainstorm all possible causes and place these in the suitable area of the
chart. Emphasize brief and succinct descriptions. Proceed through the
chart one main category at a time, but also include suggestions that belong
to categories other than the one currently being treated. Causes that belong
to more than one category are placed in all relevant positions. It is often
required to redraw the chart after the first version has been completed.
6. Analyze the identified causes to determine those that should be addressed fur‑
ther. Remember that the purpose is to cure the problem, not the symptoms.
Main Main Main Main
category category category category
130  Chapter Nine

Cause Cause Cause Cause

Cause Cause Cause Cause
Problem Problem
Cause Cause to be Cause Cause to be
Cause Cause solved Cause Cause solved
Cause Cause (effect) Cause Cause (effect)

Cause Cause Cause Cause

Main Main Main Main
category category category category

Start Step 1 Step 2 Step 3 Step 4 Finish

Main Main Main Main

category category category category

Cause Cause Cause Cause

Cause Cause Cause Cause
Problem Problem
Cause Cause to be Cause Cause to be
Cause Cause solved Cause Cause solved
Cause Cause (effect) Cause Cause (effect)

Cause Cause Cause Cause

Main Main Main Main
category category category category

Figure 9.5  Process chart with fishbone charts for each process step.
Tools for Analyzing the Performance Shortcoming  131

The process chart variant of the fishbone chart is more directly aimed at improve‑
ment of business processes. The main steps in the process to be improved are
drawn. For each step of the process believed to create problems or contribute to a
low overall performance level, a fishbone chart is constructed that deals with all
potential causes of why this step performs less than expected. For each of these
individual fishbone charts, the process is the same as outlined earlier for construct‑
ing such a chart.
After individual charts for each problem step of the process are designed, a
collective analysis is undertaken to identify the causes that seem to be of the high‑
est importance. For these, solutions are sought that can reduce the negative effects
on the overall performance level of the process. The resulting chart might look like
the one shown in Figure 9.5.


A pump manufacturer experienced frequent defects in a series of pumps

delivered to a customer. Closer inspection revealed that most of the defects
stemmed from inaccurate dimensions of two shafts used in the pumps.
A group consisting of designers, the manufacturing manager, the shaft
department manager, and several operators was established. The objec-­
tive was to find and eliminate the causes of the problems. The design of
the fishbone chart in Figure 9.6 gave some clues as to where the problems
might be found.
From the fishbone chart, it was evident that the physical environment
where the shafts were manufactured was rather unsuitable for the process,
as a low temperature in the shop along with the wrong coolant resulted in
most of the dimension errors.

Employees Methods

Low motivation Unsuitable process route

hardening Dimension
Wrong coolant Poor Too infrequent
maintenance tool sharpening

Often too
low temperature
Environment Equipment

Figure 9.6  Fishbone chart for shaft manufacturing.

132  Chapter Nine

9.4  Five Whys Analysis

The five whys analysis is also known as the ­why-­why chart and root cause analy‑
sis. As these names imply, the purpose is to find the true root cause of a problem
(Andersen and Pettersen, 1996). The technique can very well be used in connec‑
tion with a ­cause-­and-effect chart to analyze each identified cause to ensure that it
really is the root cause of the problem and not only a symptom of another and more
deeply rooted cause. This can in fact be compared to peeling an onion, where each
layer is removed to reveal another layer, until the center of the onion is reached.
The procedure for conducting a five whys analysis is as follows:

1. Determine the starting point, either a problem or a ­high-­level cause that

should be further analyzed.
2. Use brainstorming to find causes at the level below the starting point.
3. For each identified cause, pose the question, why is this a cause for the
original problem?
4. For each new answer to the question, ask the question again and again
until no new answers result. This will probably be one of the root causes
of the problem. As a rule of thumb, this often requires five rounds of the
question why.

If the question is slightly altered to ask “how” instead of “why,” the technique can
be used to find root means for reaching a desired state or effect. The analysis can
be conducted in different ways. A graphically visible way of keeping track of the
different levels of causes is listing them below one another, as shown in Figure 9.7.
In this example, a manufacturing company is attempting to reduce the amount of
work in progress. The result of the analysis is that the key to reducing the work in
progress is to develop good relationships with the suppliers. If this analysis had

Low level of work in progress

Why? Maintain no stock of finished goods

Why? Short manufacturing time

Why? Run small batch sizes

Why? Frequent and swift deliveries from the suppliers

Why? Extremely good relationships with the suppliers

Figure 9.7  List representation of five whys.

Tools for Analyzing the Performance Shortcoming  133

not been undertaken, the company might be led to believe that the answer was
simply to remove the finished goods inventory, which could have had serious con‑
sequences. Alternatively, a diagram might be used to portray an entire network of
causes at different levels, as shown in the following example.


A video rental chain had experienced an ­ever-­increasing degree of customer dissatisfaction

when measuring customer satisfaction data. A simple questionnaire revealed the four most
important causes:

• Long checkout times

• Poor selection of titles
• Impolite and unfriendly personnel
• Unfavorable store location and layout

Through a five whys analysis, the chart in Figure 9.8 was constructed. Most of the problems
could thus be attributed to low wages and low management expertise.

files No computer Low management
Long checkout Manual system expertise
times invoicing
Poor employee High employee Low
training turnover wages

Poor customer Poor question and Low management

Poor title analysis selection plan expertise
Dissatisfied selection Poor employee High employee Low
customers training turnover wages

Impolite Poor employee High employee Low

personnel training turnover wages

No market analysis
Poor store performed Low management
location and layout No competitor expertise
analysis performed

Figure 9.8  Five whys analysis chart.

134  Chapter Nine

9.5  Scatter Chart

A scatter chart can be used to show the relationship between two variables. The
variables can be process characteristics, performance measures, or other condi‑
tions and are usually measured at specified time intervals. When one of the factors
increases, the other can also increase, decrease, or display only random variation.
If the two variables seem to change in synchronization, it might mean that they
are related and impact each other. For example, we might find that the number
of defects increases in proportion with the amount of overtime. However, keep in
mind that even if there is some degree of synchronized variation between variables,
it does not mean for certain that there is a ­cause-­and-effect relationship between
them; it may very well be a third variable that is causing the effects. This can be
illustrated by the fact that a scatter chart was once constructed that showed an obvi‑
ous correlation between the Dow Jones index and the water level of Lake Superior
in the years between 1925 and 1965.


Strong positive correlation Possible positive correlation


No correlation

Strong negative correlation Possible negative correlation

Figure 9.9  Scatter charts showing different degrees of correlation.

Tools for Analyzing the Performance Shortcoming  135

The degree of relationship between the variables being examined can range from
highly positive to highly negative correlation. Between these two extremes are weaker
degrees of both positive and negative correlation, as well as no correlation. Figure 9.9
shows some examples of scatter charts for different degrees of correlation.
When using a scatter chart, the following steps are usually conducted:

1. Select the two variables, one independent and one dependent, to be

2. For each value of the independent variable, the corresponding value of the
dependent variable is measured. These two values form a data pair to be
plotted in the chart. Typically, there should be at least 30, but preferably
more than 100, data pairs to produce a meaningful chart.
3. Draw the chart by placing the independent variable, that is, the expected
cause variable, on the x axis, and the dependent, the expected effect vari‑
able, on the y axis.
4. Plot the collected data pairs on the chart and analyze them. If the chart
shows no correlation, the data pairs can be drawn in a logarithmic chart.
Such a chart can reveal connections that are not visible in a chart with
ordinary axes.

Producing a scatter chart by hand, at least with some amount of data, can be tedious
work. Spreadsheet software does the job much more easily. Remember, if the chart
indicates a relationship, the variables should be further analyzed to confirm this. In
the same way, a chart showing no relationship should not automatically lead to the
conclusion that no relationship exists.


As part of efforts to create an improved state of readiness for performing

rush jobs, the electric installation company we looked at earlier wanted to
examine the relationship between the amount of jobs—and thus the need
for ready servicemen—and the number of days with thunder and lightning.
Through a quarter of a year, the company every week counted the num-­
ber of jobs and the number of days with thunder and lightning. The scat-­
ter chart that was designed (see Figure 9.10) displays a clear correlation
between bad weather and the amount of jobs. In response, the company
started considering the weather forecasts when planning capacity and the
possibility for reserve capacity.
136  Chapter Nine

Amount of jobs





Number of days with

0 1 2 3 4 5 6 thunder and lightning

Figure 9.10  Scatter chart to analyze weather and jobs.

9.6  Histogram

A histogram, or a bar chart as it is also called, is used to display the distribution and
variation of a measure of some sort. The measure can be an infinite number of differ‑
ent aspects—for example, length, diameter, duration, and costs. The same informa‑
tion can also be presented in a table, but it can be difficult to spot important patterns in
the data material. The graphic presentation format makes it easier to see relationships.
Unlike the Pareto chart, which is really a variant of a histogram that displays the fre‑
quency of occurrence for a phenomenon, a histogram portrays intervals of data.
The following list outlines the construction of a histogram. The procedure is
illustrated using an example where the diameter of a hole has been measured for
125 work pieces. The data material is shown in Table 9.1.

1. Count the number of data points in the data material and label this number
N. To produce a valid histogram, there should be at least 50 data points. In
the example, N = 125.
2. Determine the numerical distance R between the largest and smallest value
in the data material. In the example, R = 10.7 – 9.0 = 1.7.
3. Depending on the number of data points, R is divided into a number of
equally large classes, C. The number of classes can be found in Table 9.2.
For N = 125 in the example, the number of classes should be between 7
and 12. We have decided to use 10 classes, that is, C = 10.
Tools for Analyzing the Performance Shortcoming  137

Table 9.1  Measures of the diameter of a hole for 125 work pieces.
9.9 9.9 10.4 9.8 10.1 10.2 9.8 10.3 9.7 9.7
9.8 9.3 10.2 9.3 9.2 9.8 9.8 10.1 9.8 9.8
9.7 9.8 10.2 9.8 10.2 10.0 9.7 9.5 9.6 9.5
10.2 9.4 10.1 10.1 9.6 9.7 10.0 10.0 9.3 9.5
9.9 10.1 9.6 9.7 9.6 9.5 9.7 9.7 10.0 9.7
9.3 10.7 9.8 9.8 9.8 9.9 9.6 9.7 9.7 9.9
9.0 10.2 9.3 10.3 9.9 9.9 10.1 10.7 10.7 9.6
10.0 9.5 9.2 9.9 10.0 10.1 10.0 9.8 9.4 9.3
9.5 9.7 9.7 9.7 9.8 10.2 10.4 9.6 9.9
9.6 9.7 9.4 9.8 9.9 10.3 9.8 10.0 10.0
10.3 9.4 10.6 9.4 9.8 9.8 9.5 10.7 10.1
9.5 9.6 10.1 10.1 9.6 9.3 9.5 9.9 10.3
9.9 9.5 9.7 10.1 10.0 10.0 9.6 9.4 9.9

Table 9.2  Determining the number of classes, C.

Number of data points, N Number of classes, C
Fewer than 50 5–7
50–100 6–10
100–250 7–12
More than 250 10–20

4. Determine the width of each class, labeled H. This is calculated using the
following formula:

R 1.7
H= = = 0.17 ≈ 0.2.
C 10

As can be seen, the class width in the example is rounded off to 0.2, as the
width should always have the same number of decimal places as the data
points—that is, one decimal place in this case.
5. Determine the lower and upper values for the individual classes. This is
done by letting the smallest value in the data material become the lower
value for the first class. The upper value for this class is then found by
adding the class width to the lower value. In the example, the first class
will span the range from 9.0 to 9.2. The next class starts at 9.2 and covers
the values up to 9.4, and so on. Remember that the lower value always
138  Chapter Nine

includes this value, while the upper value only goes up to it. The value 9.2,
in other words, belongs to the second class, not the first one.
6. To simplify the construction of the histogram, the data material is inserted
into a check sheet. For the example, the check sheet is shown in Figure 9.11.
7. Finally, the histogram is constructed according to the check sheet. The
classes are marked along the horizontal axis and the frequency along
the vertical. The distribution among the classes is indicated by bars. The
resulting histogram for the example is shown in Figure 9.12.

Class Lower value Upper value Frequency Total

1 9.0 9.2 1
2 9.2 9.4 9
3 9.4 9.6 16
4 9.6 9.8 27
5 9.8 10.0 31
6 10.0 10.2 23
7 10.2 10.4 12
8 10.4 10.6 2
9 10.6 10.8 4
10 10.8 11.0 0

Figure 9.11  Check sheet for the example.








9.0 9.2 9.4 9.6 9.8 10.0 10.2 10.4 10.6

Figure 9.12  Histogram for the example.

Tools for Analyzing the Performance Shortcoming  139


Imagine that this histogram was designed by a company manufacturing

steel components that should have a hole with a diameter between 7.5
and 10.5 millimeters. With the help of the diagram, the company can see
that the process falls beyond the upper specification limit and will actually
manufacture approximately 3 percent defects. This information can thus
be used to adjust the process to avoid defects.

Using histograms this way is actually highly connected to statistical process con‑
trol, which is treated in section 11.4. Whereas a control chart in statistical process
control contributes to a continuous monitoring of whether a process is under con‑
trol, a histogram can reveal permanent deviations that do not appear in a control
chart. The shape of the histogram should be closely inspected and interpreted to
reveal problems in the process.
Ideally, a histogram will produce an image of the variation in the data mate‑
rial while at the same time providing a sensible level of detail. If, however, too few
classes have been defined, only a few bars will be the result, and these are not suited
for revealing any patterns. Correspondingly, too many classes will hide the pattern
because some will be empty, which results in an appearance resembling a comb.
Particular patterns in the histogram indicate some typical problems in the

• A pattern with one obvious peak shows the mean value for the process.
Depending on how much the process varies around this peak, the process is
defined to be good or bad. A diagram like the one shown in Figure 9.13 is a
process displaying a small variation width that lies centered within the spec‑
ification limits. This is a good process. The diagram in Figure 9.14 shows
a poorer process with a large variation width. If this cannot be reduced, the
process will produce results outside the limits, which can be countered by
conducting a 100 percent control.
• Furthermore, the location of the peak in diagrams with only one peak deter‑
mines whether the process is under control. As was mentioned, the pro‑
cess in Figure 9.13 is a good process displaying little variation and lying
well centered within the limits. In Figure 9.15, another process with little
variation is shown, but this one is not centered. Adjusting this process could
result in a good process.
140  Chapter Nine

Lower Upper
specification specification
limit limit


Figure 9.13  Centered process with a small variation width.


Lower Upper
specification specification
limit limit


Figure 9.14  Centered process with a large variation width.

• When the histogram displays two clear peaks, as shown in Figure 9.16, this
can be the result of different causes. Either the data points can stem from
two different sources, which should be checked, or the process’s mean value
has changed during the data collection interval.
• A ­cut-­off pattern—that is, a histogram ending abruptly and showing no
signs of tapering off—is a sign of control or selection of the results. This
is shown in Figure 9.17 and is often visible close to the tolerance lim‑
its. In this case, a control has probably been conducted, and results fall‑
Tools for Analyzing the Performance Shortcoming  141


Lower Upper
specification specification
limit limit


Figure 9.15  Process with a small variation width, but not centered.



Figure 9.16  Histogram with two peaks.

ing outside the limits have been removed from the data material. In these
cases, we must know how much the defects cost. The process should be
improved to avoid the defects.
• The previously mentioned comblike pattern is an indication that too
many classes have been defined. However, such a pattern can also be a
result of problems with the measurement equipment. If the measurement
equipment is unable to register points in the individual classes or with the
accuracy required to use so many classes, such an appearance will be the
result. In such a case, the measurement approach should be reevaluated.
142  Chapter Nine


Lower Upper
specification specification
limit limit


Figure 9.17  Histogram that is cut off.

9.7  Relations Diagram

A relations diagram is a tool to identify logical ­cause-­and-effect relationships in

a complex and confusing problem or situation. For problems or situations where
there is a network of such relationships, a relations diagram is particularly useful,
as it has the ability to visualize them. There are two types of relations diagrams:

• Qualitative relations diagram

• Quantitative relations diagram

In the qualitative diagram, both the problem and the causes at several levels
can be included, as shown in Figure 9.18. The diagram is actually quite simi‑


Cause Cause

Cause Problem Cause

Cause-and-effect Cause

Figure 9.18  The principles of a qualitative relations diagram.

Tools for Analyzing the Performance Shortcoming  143

lar to the traditional ­cause-­and-effect chart, but it is more suited for complex
To generate a qualitative relations diagram, follow these steps:

1. Isolate all factors believed to be related to the problem. Without forming

an opinion about the relationships between the factors, each of these is
freely expressed on an individual basis. Boxes can very well be drawn that
contain the factors.
2. Identify the causal relationships between the factors and illustrate these
with the help of arrows in the diagram.
3. Classify the factors depending on which role they play in the ­cause-­and-
effect situation.
4. Concentrate the improvement effort around the main causes of the


After having spent much time and money introducing a system for perfor-­
mance measurement, a company learned that the system was not widely
used and was viewed with little respect among the employees, at times
even directly sabotaged. The company spent quite some time construct-­
ing the relations diagram in Figure 9.19, and it found two causes of the

Inaccurate Poor knowledge
measurement about data collection

Nonmotivating The measurement The measures are

measures system does not work not taken seriously

Poorly defined Some measures

measures conflict

Figure 9.19  Qualitative relations diagram for a poorly functioning measurement system.

A quantitative relations diagram can sometimes be easier for determining the

role of different factors. In Figure 9.20, a general diagram is shown, constructed
according to the following approach:
144  Chapter Nine

1. Place the factors to be included in the analysis throughout the diagram,

preferably in a coarse circular shape.
2. For each factor, assess which other factors this impacts or is impacted by,
and indicate these impacts with arrows. The direction of the arrow indi‑
cates the direction of the impact—that is, an arrow pointing to factor B
from factor A means that factor A impacts factor B.
3. After all relationships have been assessed, the number of arrows is counted
and denoted in the diagram.

Depending on the number of arrows pointing in each direction for a factor, the fac‑
tor can be defined to play one of two roles:

• Performance driver or cause—that is, a factor that impacts or enables

the performance level of another factor. Another term is throttle vari‑
able. A performance indicator has more arrows pointing away from it
than to it.
• Result indicator or effect—that is, a factor that indicates a consequence as a
result of a performance driver. A result indicator has more arrows pointing
to it than away from it.

When trying to find the main cause of a problem or an effect, ultimately wanting to
identify the factor that will increase the performance if it is improved, the perfor‑
mance drivers must form the starting point. They drive the process and thus create
its performance level.

0 in 4 out

2 in 0 out Factor A
Factor E

Factor B
1 in 3 out
Factor D
2 in 0 out

Factor C
2 in 0 out

Figure 9.20  A general quantitative relations diagram.

Tools for Analyzing the Performance Shortcoming  145


If a quantitative relations diagram were used in the preceding example

instead of a qualitative one, the result would be as depicted in Figure 9.21.
This diagram gives the same information: the measurement system does
not work (7 in, 0 out), while the main performance drivers are poorly defined
measures (0 in, 5 out) and lacking measurement training (0 in, 3 out).

0 in 3 out

3 in 2 out training
1 in 2 out
measurement Poor knowledge
about data collection

1 in 1 out
3 in 1 out
The measures are
not taken seriously

Poorly defined The measurement

measures system does not work
0 in 5 out 7 in 0 out
Some measures

1 in 2 out

Figure 9.21  Quantitative relations diagram for the example.

9.8  Matrix Diagram

So far, we have looked at tools that, in different ways, contribute to identifying

relationships between different factors, often in the shape of ­cause-­and-effect rela‑
tionships. The matrix diagram has about the same objective, but the force of this
tool compared with many of the others lies in its ability to graphically portray the
strength of such relationships. Like many of the other tools, this one can also be
used in several stages of improvement work, such as prioritizing improvement
areas, identifying problems and causes, and planning.
146  Chapter Nine

Depending on the shape of the matrix, and thus the number of variables ana‑
lyzed, there are several types of matrix diagrams, as shown in Figure 9.22 (Swan‑
son, 1995):

• Roof-shaped, which was briefly mentioned in section 5.6. Unlike the other
types of matrix diagrams, where the strength of the relationships is given in
only one direction, the roof matrix classifies relationships as neutral, posi‑
tive, or negative.

• L-shaped

• T-shaped

• Y-shaped

• X-shaped

• In addition, there is a C-shaped matrix that treats ­three-­dimensional rela‑

tionships, but is rarely used due to its complexity.

The number of variables analyzed in the different formats, as well as the number
of direct and indirect (that is, through a third variable) relations, is summarized in
Table 9.3.

Variable 2
Variable 1

Variable 1

Roof-shaped L-shaped


Variable 1

Variable 2



Variable 2 Variable 1 Variable 3

Variable 3

Variable 4

T-shaped X-shaped

Figure 9.22  Types of matrix diagrams.

Tools for Analyzing the Performance Shortcoming  147

Table 9.3  Formats with the number of variables and relations for matrix diagrams.
Number of Indirect
Matrix shape variables Direct relations relations
L 2 1 0
T 3 2 1
Y 3 3 0
X 4 4 2
C 3 3 simultaneously 0
Roof 1 — —

As explained in section 5.6, a standard set of symbols is used for indicating the
strength of relations between variables. Figure 9.23 shows these symbols and the
corresponding weight factors.
The steps when using a matrix diagram are the following:

1. Select the variables to be analyzed for potential relationships.

2. Select the matrix format according to the number of variables and the
number of expected relations. Table 9.3 can aid in this task.

3. Insert the variables into the diagram.

4. Indicate relations by using the symbols in Figure 9.23. Do not be tempted

to use corresponding weight factors inside the actual matrix, as this reduces
the readability of the diagram.

5. For each row and column in the diagram, the corresponding weight factors
for each relations symbol are summarized and the sums are presented in
the diagram.

6. Variables with a high sum play an important role in the situation and should
be analyzed further.

Relation Symbol Weight

Weak 1
Medium 3
Strong 9

Figure 9.23  Relations symbols.

148  Chapter Nine


The video rental chain described in a previous example wanted to assess

the impact of the different business processes with regard to customer
needs and expectations. The analysis was performed in an L-shaped
matrix diagram, as shown in Figure 9.24.
With the diagram, the company could answer a number of questions.
If, for example, it wanted to improve the checkout time, the relevant pro-­
cesses would be the checkout process itself as well as the information
system used. Low prices, on the other hand, could be achieved through
improvements in almost all processes.


Customer Market Title Competitor Information

expectations analysis Checkout selection Advertising analysis system Training Total
Low prices 22
Good selection 34
Several copies 21
Store layout 12
Store location 3
Speedy checkout 21
Friendly staff 12

Total 18 22 21 6 16 18 24 125

Figure 9.24  Matrix diagram for the video rental chain.

9.9  Is–Is Not Analysis

One phenomenon I have observed quite a few times in improvement projects is

an idea overload. Many who regularly work with improvements become quite
adept at taking brainstorming sessions to the fullest, and, as a result, many teams
find themselves inundated with ideas about causes of low performance or possible
improvements that can be made. In such cases, the danger is that ideas are so plen‑
tiful it becomes difficult to separate the essential from the trivial.
The is–is not analysis is a tool that helps you see distinctions and clarify
what the problem is not about. Using this tool can help you understand what
is most likely to cause the performance shortcoming as well as identify which
issues are definitely not related to it. Even at a later stage in the improvement
cycle, when many ideas for improvement actions are on the table, the analy‑
Tools for Analyzing the Performance Shortcoming  149

sis can be used to distinguish between what is and what is not related to the
The steps to undertake an is–is not analysis are:

1. Produce an empty matrix of six rows by four columns and fill the header
column with the standard headings shown in Figure 9.25.
2. In the ­upper-­left corner of the matrix, state the performance shortcoming
being analyzed.
3. Fill in the second column with “is” information: what, where, when, who,
and so on.
4. In the same manner, fill in the third column with “is not” information.
5. Compare the two columns for anything odd or that stands out and place
these in the fourth column.
6. For each element in the fourth column, analyze how it could be a cause of
the performance shortcoming.
7. For the possible causes identified this way, test them by checking if they
explain all items in the “is” and “is not” columns. Those that do are likely
the real cause(s).

shortcoming Is Is not Distinction

What occurs,
what objects
are affected?

Where does
the problem

When does
the problem

Extent of

Who is

Figure 9.25  Generic table for is–is not analysis.

150  Chapter Nine


A large ski resort operates a number of different ­sub-­businesses: ski lifts,

ski equipment rental, a ski store, restaurants, and a hotel. Lately, several
accidents have been reported in connection with the ski lifts. The regulat-­
ing body granting licenses to operate the lifts was notified, and following
an inspection, the resort was given a month to investigate the incidents
and come up with an action plan to prevent further accidents.
The matrix in Figure 9.26 emerged from the analysis of the reported
incidents that occurred during the previous year.
The analysis gave clear indications that the injuries occurred only in
special areas of the resort and were caused by a certain group of guests.
After further investigation, it was discovered that the tree barrier that was
supposed to keep skiers away from the lift area had been compromised
at both lifts 2 and 5. A couple of trees had fallen, and branches appeared
to have been cut by a knife. This was enough to create an opening that
allowed a shortcut between different slopes. Normally one would have to
ski to the bottom, traverse a flat area to the next lift, and take this to the

shortcoming Is Is not Distinction

People on the No one has Injuries occur

What occurs,
ground hurt by been hurt by as a result of
what objects
objects dropped falling out of objects falling
are affected?
from the lifts the lifts from the lifts

Lifts 1, 3, and Only a small

Where does During the top
4 or lower part of the lift
the problem 50 yards of
parts of lifts area seems to
occur? lifts 2 and 5
2 and 5 be affected

95% of the Weekends

When does Very few
incidents seem to be the
the problem injuries during
occurred during problematic
occur? weekdays
weekends time period

Injured guests No injuries Injuries seem

Extent of
between 7 and among adults to afflict younger
22 years old older than 22 guests

Objects (skis, Inexperienced

Only one case of
helmets, poles) skiers with
Who is an experienced
dropped by rented equip-
involved? skier dropping
inexperienced ment cause
a ski
skiers most incidents

Figure 9.26  Ski resort is–is not analysis.

Tools for Analyzing the Performance Shortcoming  151

top. Putting up a railing or some other physical barrier would block this
shortcut and eliminate most of the problem.
The other observation was that only inexperienced skiers—more spe-­
cifically, people who rented equipment from the resort—dropped skis and
other objects from the lifts. Such objects could not hurt other guests after
the shortcuts had been closed, but this was still annoying, both to the
guests and to the staff who had to retrieve the objects. Since service per-­
sonnel on the slope had written down the names of some of the people
involved in the accidents, three of the people who had dropped skis could
be interviewed. They all complained that they received very little training
in how to put on the boots, skis, and so on, properly when renting the
gear. As a result, the boots were loose on their feet and were not properly
attached to the bindings. The process for equipment rental was changed:
one attendant picked out the correct equipment, and another attendant
explained to the guests how to use the gear properly.

9.10  Bottleneck Analysis

Bottleneck analysis derives from a large body of research contained within the
optimized production technology (OPT) theory and the theory of constraints,
developed by Goldratt (see both Goldratt and Cox, 1992, and Goldratt, 1990). One
of the key messages of these theories revolves around bottlenecks in production
systems (where a production system can deliver products or services). A produc‑
tion system consists of a number of resources that are used to transform inputs to
outputs. Resources can be people, computers, machines, cars, and so on, and they
can all be classified as one of two types:

• Bottleneck, meaning a resource with a capacity that is less than or equal to

the demand for the resource from the market
• Non-bottleneck, which has a capacity greater than the demand

Metaphorically speaking, a bottleneck can be seen as a constriction or narrowing

of the flow passage in a production system or business process, that is, a point in
the process that limits the flow and reduces it to a narrow stream:

152  Chapter Nine

A simple example illustrates this, where the process to manufacture a prod‑

uct consists of five steps, as shown in Figure 9.27. At a demand of 2000 units
per week, the process can deliver the products with its current capacity. If
demand is increased to 2200 units in a week, step three in the process has
too low of a capacity to keep up with demand and becomes the bottleneck.
Although the other four machines can take the increased demand, the output
from the entire process will be limited by this one bottleneck resource. The
bottleneck is the weakest link in the chain and controls how much the whole
process is able to produce.
The theory of constraints contains much discussion about the properties of
bottlenecks and how bottleneck and ­non-­bottleneck resources are linked. I will not
go into detail here, but I hope you understand the significance of a bottleneck as a
limiting factor in the performance of a process. From this it follows that knowing
which resources constitute bottlenecks is useful.
First, you can eliminate a bottleneck by increasing its capacity or redesign‑
ing the process to avoid using the bottleneck. This will increase the capacity of
the process but only up to the point where another resource takes over as the
bottleneck; bottleneck elimination is a continual battle. Second, identifying the
bottleneck allows you to optimize your processes to maximize the conditions
for this resource. You can divert production from the bottleneck to some other
resource to free up time. If the bottleneck is a machine or some other type of
physical equipment, you should make sure maintenance is carried out according
to plan. If a human is the bottleneck, take good care of this employee to ensure
that she or he stays healthy, enjoys the job, and so on. Further, always make sure
the bottleneck resource is kept going, perhaps by keeping a buffer stock in front
of it. And since an hour lost in a bottleneck is an hour lost for the entire process
(according to OPT rules), make sure you don’t allow the bottleneck to produce
defective parts or services.


Process flow

1 2 3 4 5

Figure 9.27  A bottleneck resource limits the flow from the entire process.
Tools for Analyzing the Performance Shortcoming  153

Nevertheless, how you deal with a bottleneck once it has been identified is not
an integral part of the analysis. The purpose of the bottleneck analysis is to identify
the bottleneck(s) of the process or production system. The following steps detail
how to identify a bottleneck:

1. Model the business process or production system, using either a flowchart

or a network diagram.
2. Identify the demand for products or services from the process or produc‑
tion system.
3. Start from the end of the process and determine the capacity need for each
resource by multiplying the volume demanded by the cycle time in each
4. Compare the capacity need with the available capacity, preferably by
showing these as bar charts with two bars for each resource in the flow‑
chart or network diagram. Resources where the capacity need exceeds the
available capacity are bottlenecks.

Notice from the definition of a bottleneck and how the analysis is carried out that
a bottleneck is only momentary. If demand composition is changed, the bottleneck
could move to another resource or the process could function without any bottle‑
necks. However, for stable processes or production systems with a similar demand
composition over time, the bottleneck often stays the same.


A public university hospital serves as an emergency ward for its town,

meaning the doctors will see anything from serious emergency victims to
­after-­hours patients with bruised knees or sprained fingers. During a normal
week, especially between 7 and 11 p.m., a line of waiting patients would
form, many of who were in pain from accidents. Some complained loudly,
and others waited in silence. The line created frustration and stress for both
patients and personnel.
Some months before, the entire hospital had gone through a project
of mapping its processes, including the emergency ward. The flowcharts
showed a set of processes that sometimes ran through many units/types
of personnel, with the typical patient following this route:

• Immediately after entering the emergency area, see the receiving

nurse to describe the symptoms and be registered.
• In cases where the receiving nurse could quite accurately diagnose
the problem right away, the patient would be sent directly to the next
logical station, for example, X-ray, blood testing, a specialist, and so
on. A nurse accompanied these patients to the next station.
154  Chapter Nine

• If the receiving nurse could not accurately diagnose the problem, the
patient was given a number and placed in the waiting area, which
was the case for the majority of the patients.
• When called, the patient would see a receiving doctor, assisted by a
nurse, where a first examination was performed to determine where
the patient should go next. In some cases—for example, stitching
small wounds—treatment could be given here and the patient might
be allowed to go home.
• Normally, the patient would be released back into the waiting area
to await her or his place in line for the next step in the treatment, be
it X-ray, blood testing, or another doctor.
• When the patient’s number came up again, a nurse would take him
or her to the next post. If this was only further testing—for example,
an X-ray—the patient would be taken back to the waiting area after-­
ward to reenter the line.
• Having received the test results, or if testing was not required, the
patient would be taken to a specialist when the patient’s number in
the queue came up.
• For most patients this was the last step, either being treated and
sent home or being diagnosed and having to come back during the
day for further treatment or analysis.

Statistics showed that many patients spent between three and four hours
in the hospital before going home, many of them with only a cold, the flu,
minor injuries, or other ailments that were quickly diagnosed and treated
(sometimes simply with the order of going to bed). Personnel in the emer-­
gency ward complained that capacity problems in X-ray or other test facili-­
ties caused the long lines, while the specialists claimed they were ready
and waiting for patients but sat idle for long periods of time before any
showed up.
To understand what was causing the problems, a bottleneck analysis
was performed. The result is shown in Figure 9.28, which clearly indicates
that the bottleneck was the admitting nurse. After further discussions, it
became clear that the reception station was severely understaffed, and
nurses were prevented from asking sufficient questions of the patients to
allow them to be sent to the right place from the beginning. Adding one or
two nurses here would change this and would give them the opportunity to
do the first diagnosis here.

Analysis tools are quite different in nature and in what angle they take when
viewing the data. This means that some of them lend themselves more naturally to
one type of issue, and others lend themselves to other performance shortcomings.
I suggest you consider this carefully before choosing a tool.
Capacity Capacity
available demand


Receiving nurse Waiting line Receiving doctor Waiting line Waiting line Specialist Cash desk

Receiving nurse

Blood test

Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity
available demand available demand available demand available demand available demand available demand available demand available demand

Figure 9.28  Emergency ward bottleneck analysis.

Tools for Analyzing the Performance Shortcoming  155
Tools for Generating Ideas and
Choosing among Them

he two preceding chapters described tools for two ­somewhat-­related stages
of collecting data about the performance shortcoming and analyzing the
shortcoming and its causes. Knowledge about the situation and its causes
can be used at the next stage to generate ideas about possible solutions or improve‑
ments. This chapter describes some tools that can be used to generate ideas and aid
in reaching consensus solutions:

• Brainstorming
• Brainwriting/Crawford slip method
• Six thinking hats
• Nominal group technique
• Paired comparisons

I have already mentioned that some tools and techniques can be applied elsewhere
in the improvement process than where they were presented. This is especially true
for these tools, as there will typically be many times during a project when there is
a need to come up with a broad range of ideas and choose among them.

10.1  Brainstorming

Finding solutions to the problems we are usually faced with in improvement proj‑
ects requires creativity. Most of us are far better at thinking analytically to find the
right solution. In improvement work, the purpose is to find as many solutions as
possible and then proceed with only the most promising ones. For this purpose,
brainstorming and other related techniques described in this chapter are i­nvaluable.

158  Chapter Ten

Brainstorming is perhaps the ­best-­known approach, with the exact purpose of cre‑
ating as many ideas as possible, and it has numerous benefits:

• Stimulates creativity
• Encourages joint problem solving
• Makes it possible for participants to build on one another’s ideas
• Minimizes the tendency to prematurely evaluate ideas
• Appreciates thinking that expands the traditional boundaries of the solution

There are at least two different ways to brainstorm:

• Structured brainstorming, where each participant in turn launches one idea.

This approach is more structured and ensures equal participation, but it is
less spontaneous and to some extent limits the possibility of building on one
another’s ideas.
• Unstructured brainstorming, where everyone can freely launch ideas all the
time. This approach is very spontaneous, but it is often more confusing and
can lead to one person or a few people dominating the activity.

For both ways, the procedure is as follows:

1. Clearly define the topic of the brainstorming and write it at the top of a
white board or flip chart.
2. Let the participants launch ideas according to the approach used, struc‑
tured or unstructured. Encourage everyone to launch ideas, no matter how
silly they might seem.
3. Write down every idea launched, preferably using the same wording as the
original proposition.
4. Do not allow participants to discuss, criticize, or evaluate ideas during
brainstorming. If an idea is not lucid enough to give any meaning, further
explanation can be solicited.
5. Allow one break period when the flow of ideas stagnates; it will usually
pick up again. When the ideas seem to become reformulations of previ‑
ously launched ideas or when the frequency of new ideas is decreasing for
the second time, close the process.
6. At the end, evaluate the ideas. Start by picking the obviously good ideas,
the “stars.” The rest of the ideas can be sorted into groups, either by theme
or by decreasing potential.
Tools for Generating Ideas and Choosing among Them  159

The consolidated list of ideas will now form a suitable starting point for choosing
the improvement ideas to proceed with. Also remember the established rules for

• No criticism or discussion of ideas is permitted during brainstorming.

• Laugh with crazy ideas, not at them.
• Be loose and spontaneous; there are no stupid ideas.
• Keep all ideas.
• Combine ideas. The ideas are not mine or yours; they are the group’s.


Employees of a day care who worked with a class of 30 children, aged one
to four years, received complaints from parents that their kids talked about
episodes of teasing and bullying. The employees had also noticed that
some kids made fun of others, but they did not consider it serious. Learn-­
ing that some children reacted strongly, the day care brought in an expert
on antibullying.
She suggested that the children be taught how to brainstorm ideas.
The children were enthusiastic and participated in the training. She used
examples such as games to play at birthday parties and healthful snacks
and meals. Then she asked the children to brainstorm ways to achieve a
good atmosphere in the classroom and avoid making anyone feeling sad.
The children came up with more than 40 ideas, many of them very
good; for example:

• Create buddy pairs of one older and one younger child, and let this
rotate every week
• Take turns deciding which games to play, indoors and outdoors
• No dessert at lunchtime for anyone caught teasing or bullying

Altogether, several of the ideas were implemented and seemed to create a

much better climate and play environment for all the kids.

10.2  Brainwriting/Crawford Slip Method

Brainwriting and the Crawford slip method are two methods of brainstorming in writ‑
ing. The advantage of generating ideas through writing is that it is easier to describe
more detailed and coherent ideas, which often leads to the development of equally
coherent solutions. While brainwriting is almost solely a written adaptation of brain‑
storming, the Crawford slip method is a technique that protects the anonymity of the
160  Chapter Ten

participants. This second approach is therefore often used when the group experiences
conflict that stifles creativity or when large amounts of information are expected to
surface through the idea-generating session (Swanson, 1995).
Looking at brainwriting first, it can be performed in two different ways:

• The card method, where ideas are written on small cards and circulated
among the participants for the addition of related ideas or expansion with
other elements.
• The gallery method, where ideas are written on a number of white boards or
flip charts and the participants circulate among these to add related ideas or
expand on existing ones.

The procedure for brainwriting is as follows:

1. As with brainstorming, start by clearly defining the target topic for the idea
generation. The topic can be written on either a white board or the partici‑
pants’ individual cards if the card method is used.
2. The participants then write down their ideas, either on the cards or on the
white board. Precise formulations are encouraged so as to enable under‑
standing without explanation from the owner.
3. The participants are allowed to add to others’ ideas to reap the effects from
combining ideas or further developing them.
4. At the end, the ideas are verbally discussed by the group and preferably
also sorted into classes of ideas.

The Crawford slip method is a variant of the brainwriting card method in that
the cards are not circulated for expanding on others’ ideas. Neither method is
an open evaluation of the ideas conducted, which further increases the neces‑
sity of accurate idea formulation. Sorting of the ideas is done by one person,
often aided by computer software designed for this purpose (several of which
are offered on the market). The final document summarizes all ideas and can
be openly used in the group to decide which ideas will be used in the improve‑
ment work.


A stable housed a riding school and a horse club, and private horse owners
also kept their horses there. For years there were complaints and clashes
between the different user groups. The horse club members were angry with
Tools for Generating Ideas and Choosing among Them  161

the riding school for continually occupying the riding hall and paddock, the
horse owners claimed the riding school students borrowed or stole their equip-­
ment, and everyone blamed the others for not cleaning up after themselves.
After the disappearance of an expensive saddle from the tack room
used by the horse owners, the situation was close to erupting. The stable
owner decided to do something, but he figured brainstorming would not be
productive given the tension among the groups. He passed out suggestion
cards to all users of the stable and asked them to write down as many pos-­
itive ideas as possible that could help make the place more harmonious.
After receiving some 20 idea cards from all three user groups, he sat
down to review and sort the ideas. Some had clearly not accepted his
request for positive ideas and suggested he throw people out of the stable,
close the riding school, and worse. Many ideas were genuinely construc-­
tive, though, and he ended up with quite a few suggestions he thought
could be implemented:

• Let the riding school and the riding club alternate between semes-­
ters in choosing paddock and riding hall hours
• Seek permission from a neighboring farmer to use one of his fields
as an alternative riding area
• Take the riding school students into the woods for trekking now and
• Make sure the riding school equipment is always in order, to avoid
having students look for replacement equipment elsewhere in the

Although some people left the stable, the changes implemented improved
the situation considerably.

10.3  Six Thinking Hats

Six thinking hats is a method for actively encouraging people to view a problem
and its solutions from several different perspectives. As such, it is considered
more a technique for inspiring creativity than for generating ideas. However, I
have come to view the technique as an ­idea-­generation method. Taking distinctly
different angles in a discussion seems to bring up ideas that would otherwise not
have surfaced. And deliberately having a team member take the position of devil’s
advocate ensures that suggested solutions are debated with regard to feasibility and
possible flaws.
The technique achieves this by encouraging you to recognize what type of
thinking you are using and to apply different types of thinking to the subject.
Edward de Bono (1985) originally created the method when working to improve
162  Chapter Ten

creativity and lateral thinking. He gave the six thinking hats the following colors
and respective roles:

• The white hat: cold, neutral, and objective; the person “wearing it” should
be systematic and careful in looking at the facts and figures

• The red hat: represents anger; the wearer should listen to intuition/gut feel‑
ing and his or her own emotions

• The black hat: pessimistic and negative; thinking with this hat on should
focus on why an idea will fail

• The yellow hat: optimistic, sunny, and positive; the wearer should focus on
seeing ways ideas will work and overcoming obstacles

• The green hat: represents grass, fertility, and growth; the person underneath
it should be creative in cultivating new ideas

• The blue hat: connected with the sky; the person should focus on seeing
things from a higher perspective

Although some of the connotations used in describing the hats can be construed as
negative, it is important to understand that each hat is equally important in ensur‑
ing fruitful discussions.
The steps in using the six thinking hats are the following:

1. Assign hats to the people in the discussion team, preferably one color per
person, and make sure that everybody understands that when speaking
during the session, he or she must clearly identify with the color of his or
her hat

2. The team engages in a creative discussion about the problem, and each hat
has the following responsibilities (other team members may contribute at
each step):

• The white hat presents facts about the problem

• The green hat presents ideas on how the problem could be solved

• The possible solutions are discussed, with the yellow hat focusing on
benefits and the black hat on drawbacks

• The red hat works to elicit all team members’ gut feelings about the

• The blue hat summarizes the discussion and closes the meeting
Tools for Generating Ideas and Choosing among Them  163


A movie theater with 12 theaters received many complaints from the clean-­
ing staff about the state of the theaters after shows: litter filled the rows
between the seats, soda was spilled on the seats, and the backs of the
seats were covered with dirty shoe marks. The manager assembled a team
of six people—two members of the cleaning staff, two employees from the
snack bar, a ticket salesperson, and himself. Having heard about the six
thinking hats from a friend, he wanted to try this approach.
He asked his friend, who had quite a bit of experience in using the
method, to come teach them all about it. Spending half a day listening to
this individual and discussing the approach, the team felt ready to proceed.
They met one morning before normal working hours and assigned the hats
among them. Proceeding with the discussion, they more or less managed
to stay true to their parts, although there were some slippages.
At the end, the manager, wearing the blue hat, summarized all the ideas
that had surfaced. The most promising ideas were serving soda in bottles
instead of paper cups, installing drink holders on the seats, placing better
door mats at the entrances to the theaters, and taking up an old tradition of
running ­commercial-­like messages about proper cinema behavior before
the shows.

10.4  Nominal Group Technique

As mentioned in the description of brainstorming, it’s possible that one per‑

son or a few people will dominate the activity. An unfortunate effect of this,
besides the fact that the full potential of ideas is not realized, is that those
who feel they have been overlooked will not commit to the produced conclu‑
sions later on.
The intention of nominal group technique (NGT) is to render possible a brain‑
storming session where all participants have the same vote when selecting solu‑
tions. To use NGT, the following steps are recommended:

1. As in brainwriting, the first step is written idea production, where each

person generates ideas and writes these on idea cards, one idea per card.

2. All the produced ideas are registered on a flip chart and the ideas are
briefly discussed. The purpose is to clarify the content of each idea as well
as eliminate similar ideas. At the end of this stage, each idea is assigned a
letter, starting with “A.”
164  Chapter Ten

3. The next step is again an individual activity, where participants rank

the ideas. From the complete list of ideas, each participant selects up
to five ideas and writes these on his or her ranking card. Each idea is
identified by its assigned letter from the list on the flip chart. When
ranking ideas, participants assign points to the ideas—from 5 for the
most important or best idea down to 1 for the least important or worst

4. The session leader collects the ranking cards and writes the assigned points
on the flip chart. For each idea, the points are summarized to total scores.
The idea achieving the highest total score is the group’s prioritized idea or

In the ensuing improvement work, start with the solution achieving the highest
score or the two or three most important ones.


A company with about 400 employees experienced grave delays in the

internal communication among individuals and departments. Normal
brainstorming sessions produced some good ideas, but most people did
not launch any ideas. NGT was therefore used next, and after the individu-­
ally generated ideas were collected, the list looked like this:

A. Extend the use of e-mail

B. Anyone who receives an inquiry should respond confirming they
received the question
C. Develop routines for keeping everyone informed when employees
change offices
D. Implement a system for telephone messages when traveling
E. Have weekly department meetings and meetings across depart­ments
F. Sort mail automatically
G. Rebuild the office area into an open-plan office
H. Have compulsory table rotation during lunch breaks

The ranking card of one of the participants is shown in Figure 10.1, and
the resulting flip chart list is shown in Figure 10.2. The result was that the
company extended the e-mail system by giving everyone access to it and
providing training in using it.
Tools for Generating Ideas and Choosing among Them  165

Problem: Poor internal communication

Idea Points
A 5
B 4
C 1
D 2
H 3

Figure 10.1  Individual ranking card.

Problem: Poor internal communication

Idea Points Total
A 55453254 33
B 434322 18
C 14535 18
D 2345223 21
E 121 4
F 23241 12
G 15321 12
H 35551 19

Figure 10.2  Resulting flip chart list.

10.5  Paired Comparisons

A success factor in using NGT is that those involved manage to distinguish between
the ideas and prioritize them. In real life, with many ideas to choose from, it can be
difficult to decide which idea to vote for, and the results might be determined by
coincidence. Like the other methods described in this chapter, paired comparisons
aims at prioritization and consensus reaching, but it does so through sequentially
comparing two items, instead of having to consider several items at once. Single
decisions are easier to make than decisions that require you to select from a large
number of possible solutions.
The steps in paired comparisons are the following:

1. Clearly identify the alternatives to be compared. The total number, denoted

N, should be manageable—that is, not more than eight.
2. Make a matrix with the alternatives, coded by letters, as row headings and the
pairs as column headings, indicated by letters only, to save space. The num‑
ber of pairs p is determined by the following formula: p = [N × (N – 1)]/2.
166  Chapter Ten

3. Column by column, each participant votes for one of the alternatives; the
votes are logged in the matrix.
4. After participants have voted for all pairs, add up the total number for each
pair; this should equal the number of participants.
5. Add up the number of votes for each alternative to get the row totals. The
­highest-­scoring alternative is the preferred one, according to the group.


Imagine that the company from the NGT example wanted to apply paired
comparisons instead. To avoid ending up with too large a matrix, only the
five ­highest-­ranked ideas from the NGT session are brought forward here:

A. Extend the use of e-mail

B. Anyone who receives an inquiry should respond confirming they
received the question
C. Develop routines for keeping everyone informed when employees
change offices
D. Implement a system for telephone messages when traveling
H. Have compulsory table rotation during lunch breaks

The resulting comparison matrix is shown in Figure 10.3. You might notice
that with this method, ideas A and D changed places.

A/B A/C A/D A/H B/C B/D B/H C/D C/H D/H Total
A 6 2 5 3 16
B 1 0 4 5 10
C 2 3 1 4 10
D 5 7 6 5 23
H 4 2 2 3 11
Number of votes 7 7 7 7 7 7 7 7 7 7

Figure 10.3  Paired comparisons matrix for internal communication.

You should now be ready to unleash the most exciting and powerful tools: the
improvement tools.
Tools for Creating Improvements

he categories of tools presented so far have all played important parts in
the introductory phases of the business process improvement model. The
main purpose of these tools is to contribute to improvement, even if they
cover only a small portion of the total activity leading up to this target. The tools
presented under the heading of improvement are at the outset more directly geared
toward this purpose. The following tools are presented in this chapter:

• Streamlining
• Idealizing
• Statistical process control/control chart
• Six Sigma
• Business process reengineering
• Benchmarking

11.1  Streamlining

The principle behind streamlining is to trim away excess waste and superfluous
elements in business processes. A characteristic of a streamlined process is that
it flows easily and without resistance that lowers the performance level within
the process or within the environment in which it exists. Streamlining of busi‑
ness processes is really a collective term for several smaller techniques. These
can be used completely on their own, but when put together, they produce a much
stronger streamlining effect. Some of these techniques are briefly presented in the

168  Chapter Eleven

f­ ollowing sections (there are others, but either they are quite similar to those pre‑
sented or they are aimed at objectives already covered in previous chapters). They
are presented in the order in which their use is recommended when streamlining a
business process (Harrington, 1991).
You will notice that streamlining becomes even more powerful when com‑
bined with the process model or flowchart you have made. Think of the flowchart
as a map with a traveling route penciled in. When performing the streamlining
exercise, you are identifying steep hills and deep valleys that are difficult to cross.
These obstacles are marked with different colors in the flowchart, making it easy
to see where the process flows slowly and progress is halted. Either these areas can
be eliminated or the traveling route can be retraced to avoid them.

11.1.1  Bureaucracy Elimination

The words streamlined and bureaucratic are quite opposite. Streamlined is usu‑
ally associated with a process that is efficient and flows easily; bureaucratic is
synonymous with a process that is slow and cumbersome. Bureaucracy is often an
obstacle on the way toward process thinking and the transition from departmental
to process management. A natural first step when streamlining business processes,
therefore, is to eliminate bureaucracy.
One typical effect of bureaucracy is unnecessary paperwork. Many managers
spend 40 to 50 percent of their time writing or reading ­job-­related material. Studies
have shown that approximately 60 percent of all office work in a company consists
of reviewing others’ work and storing and retrieving information—information that is
sometimes useful but very often not. The negative effects of this are many and often
impossible to measure. A critical review of all such bureaucratic tasks is therefore
important for the clear purpose of minimizing all delays, red tape, and other operations
that do not stand out as useful, as ­value-­adding, or as supporting other processes.
Here is one possible approach to eliminating bureaucracy:

1. To start, the bureaucracy must be tracked down. Bureaucracy is often iden‑

tified by asking questions like:
• Is the activity performed to inspect or approve someone else’s work?
• Does it require more than one signature?
• Are several copies made of the result?
• Are several copies stored for no apparent reason?
• Are copies sent to people who do not need or use them?
• Are people or departments involved that stifle efficiency and the quality
of the work?
Tools for Creating Improvements  169

• Is there unlimited use of the copier or access to large filing cabinets?

Studies have shown that about 90 percent of all documents stored in an
organization are never used again.
2. Activities that constitute bureaucracy should be colored blue in the flow‑
chart or process model for the process.
3. The person responsible for each blue activity presents an overview of
time and costs related to the activity as well as its usefulness. This usually
meets with some resistance, as all such activities are viewed by someone
as essential to the daily operation and ­long-­term survival of the organiza‑
tion. One study documented a situation in which a request for the purchase
of equipment exceeding a certain dollar amount had to go through five
administrative levels for approval. Of 10 requests sent for review, two con‑
sisted entirely of blank pages behind the cover page allotting space for the
signatures. Together with the other eight, these two requests passed right
through the system and were duly approved.
4. Activities that cannot be justified are eliminated.

11.1.2  Redundancy Elimination

Especially for administrative processes, it is true that many identical or similar
activities are often performed at two or more places in the process. This occurs
when different departments or different organizations in the supply chain perform
their tasks independently and without knowing what the others do. This increases
process costs and also increases the likelihood that conflicting data exist. An exam‑
ple is a case where the purchasing department solicits prices for a component. At
the same time, the design department is also soliciting prices but is quoted a dif‑
ferent price. At different places in the organization, these data are used, causing
inaccuracy and errors. Due to a lack of trust in many organizations, each person
and each department often keeps their own records over matters like absenteeism
and overtime. At the same time, this is done centrally, and the information in the
various systems often differs.
The costs this double record keeping entails and the problems caused by con‑
flicting data must be avoided. There really is no trick to this, beyond going through
each of the activities and results associated with the process and identifying and
eliminating activities, documents, and other results that are generated more than
once. In the flowchart, these activities should be colored maroon.

11.1.3 ­ Value-­Added Analysis

Value-added analysis is a central concept in the streamlining of processes. Let us
first study the concepts of value and ­value-­added. When a product or service passes
170  Chapter Eleven

through the organization and is transformed from raw material into a finished prod‑
uct or service, two things happen to its value:

• By supplying materials, labor, energy, and so on, the process incurs costs for
the organization. The added value experienced by the product is, however,
independent of these costs.
• By adding functionality, aesthetics, brand name, and so on, to the product, it
experiences an added value that makes it possible to sell it for a higher price
(hopefully) than the collective costs the process has incurred.

The challenge facing the organization is to ensure that the value of the product or
service, expressed in terms of what the market is willing to pay for it, is higher
than the costs of generating it. This way, value is actually a theoretical concept
that expresses both market value and true material value. Added value (AV) can be
expressed as follows:

AV = Va – Vb,

where Va is the value after processing and Vb is the value before processing. As
already mentioned, value in this sense is affected by a number of more subjec‑
tive factors such as functionality for the intended use, prestige, synergy with other
products or services, and so on. Furthermore, this value for the customer is inde‑
pendent of how much it costs to manufacture the product or deliver the service. If
it were possible to manufacture a Mercedes for half the current cost, the value for
the customer would still be the same.
During the processing of the product or service within the organization, it is
subject to a number of activities, which can be divided into three categories:

• Real ­value-­adding activities (RVA)—activities that from the end customer’s

point of view add value to the product. These are the typical transform‑
ing operations that create the product’s functionality and appearance or the
service’s usefulness.

• Organizational ­ value-­adding activities (OVA)—activities that from the

customer’s point of view do not add any value, but that are necessary from
the organization’s point of view. These can be production planning, mainte‑
nance, personnel administration, and so on.

• Non-value-adding activities (NVA)—activities that from both the custom‑

er’s and the organization’s point of view add no value at all. Typical exam‑
ples are waiting, storing, reworking, and so on.
Tools for Creating Improvements  171

Value-added analysis looks at each activity in the process to determine its contribution
to the added value for the end customer. The intention is to classify each activity as one
of these three types and then work to minimize OVA and eliminate NVA. The analy‑
sis is carried out as shown by the sequence of questions in Figure 11.1 (Harrington,
After each activity in the process is classified as one of these three categories,
markers are used to color them in the flowchart. RVA are marked green, OVA yel‑
low, and NVA red. This gives a good overview of the portion of activities that are
truly ­value-­adding. Such a picture can often shock the organization. Very often, less
than 30 percent of the incurred costs is spent on RVA, while less than 5 percent of
the time consumed belongs to RVA. Another way of portraying this information is
by using a ­cost-­cycle time chart, as shown in Figure 11.2 (Harrington, 1991). The
chart shows the effects that can be achieved by eliminating NVA and minimizing
OVA in order to let RVA constitute the main part of the process’s time and costs.
The core task of removing NVA and minimizing OVA is a project of its own,
and there is no consensus as to how it should be performed. Some general advice
includes the following:

• Rework can be eliminated only by removing the source of the error

• Pure movement of documents or other information can be minimized by combin‑
ing operations, moving people closer to each other, or automating the process


Yes Necessary to No
produce output?

Contributes to No Contributes to the

customer satisfaction? organization’s needs?
Yes Yes No

Real value- Organizational Non-value-

adding value-adding adding

• Order handling • Planning • Storing

• Manufacturing • Maintenance • Inspection
• Packaging • Hour registration • Rework

Activities that must Activities that do not contribute to

be performed to satisfy satisfying customer requirements and
customer requirements that should be optimized or eliminated

Figure 11.1  ­Value-­added analysis questions.

172  Chapter Eleven

• Waiting time can be minimized by combining operations, balancing the

work load, or automating processes
• Most output from NVA can be eliminated if management accepts it
• Inspection and control can be eliminated by changing policies and

The result should be an increase in the proportion of RVA, a reduction in the propor‑
tion of OVA, and a minimization of the proportion of NVA. After a typical ­value-­added
analysis, the resulting chart might look like the one in Figure 11.3 (Harrington, 1991).

11.1.4  Process Cycle Time Reduction

Critical business processes are subject to the rule that time is money. Such pro‑
cesses are usually carried out using critical resources that are often bottlenecks. At
the same time, the products or services from these processes are the ones that really
matter to the customers. Therefore, the products or services should be delivered to
them as quickly as possible. Long cycles require capacity of the critical resources
for unnecessarily lengthy periods of time, they stifle efficient delivery to the cus‑
tomer, and they incur storage costs for the company.
To determine where the crusade against long cycle times should be started,
consider activities that cause delays, activities that originally display a long cycle
time, or particularly critical processes. Typical efforts to remedy long cycle times
are the following:



1 2 3 4 5 6 7 8 9 10

Cycle time

Figure 11.2  ­Cost-­cycle time chart.

Tools for Creating Improvements  173


Costs before

Costs after

1 2 3 4 5 6 7
Cycle time after
Cycle time before

Figure 11.3  Corresponding chart after completed ­value-­added analysis.

• Perform activities in parallel instead of in sequence. Very often, most of the

steps in a business process are performed in sequence, although they could
just as well be carried out in parallel. A serial approach results in the cycle
time for the entire process being the sum of the individual steps plus trans‑
port and waiting time between steps. When using a parallel approach, the
cycle time can be reduced by as much as 80 percent while producing a better
result. A classic example is product development, where the trend for many
years has been toward concurrent engineering. Instead of the sequence of a
concept, then drawings, then bill of materials, and then processes, all activi‑
ties take place in parallel in integrated teams. The development time is dra‑
matically reduced, and the needs of all those involved are considered during
the development process.
• Change the sequence of activities. This point is closely connected to unnec‑
essary transport of documents and products. Often, documents and products
are transported between machines, departments, buildings, and so on. For
instance, a document might travel between two offices a number of times
for inspection and signing. If the sequence of some of these activities can be
changed, it might be possible to perform all processing of the object when
it comes in the first time.
• Reduce interruptions. Interruptions cause delays and increase the cycle time
for critical business processes. The production of an important order can,
for example, be stopped by an order from a far less important customer, but
174  Chapter Eleven

one that has turned into a rush order because it has been delayed. People
working inside critical business processes can be interrupted by phone calls
that could have been handled by someone else. The main principle is that
everything should be done to allow uninterrupted operation of the critical
business processes.
• Improve timing. Many activities are of such a nature that they are performed
with relatively large time intervals between each activity. For example, there
might be a report that is generated only once a week or purchasing orders
that are issued every other day. People using these reports should be aware
of such deadlines to avoid missing them. Often, the manufacturing depart‑
ment does not know that purchasing requirements that are not reported by
noon Thursday will not be purchased until the following week. Improved
timing in these processes can save many days in cycle time.

There are numerous other approaches suitable for reducing cycle times, but these
are some of the most important ones. If you combine the streamlining analyses and
consistently use color codes in the flowchart, the end diagram can become highly
revealing. A general example is shown in Figure 11.4, portraying an imaginary
process where steps representing bureaucracy, redundancy, NVA, and OVA are

Redundancy Bureaucracy NVA OVA

Figure 11.4  A flowchart with steps distinguished according to streamlining results.

Tools for Creating Improvements  175


After 25 years of continuous growth, a manufacturer of electric appli-­

ances had come to a point where many of the administrative business
processes were cumbersome. Some important processes, especially order
processing, had become so time-consuming that they inhibited further
improvements in other operative processes. The company thus decided to
streamline this process.
As the company grew, a number of different computer systems had
been installed, including an order system. However, when this system was
installed, the old routines had simply been automated or transferred to an
electronic format. These routines were designed for a much lower sales
volume and complexity. The dominant opinion was that a large bureau-­
cracy was being kept alive artificially, and that the first step was to identify
and eliminate the elements of this bureaucracy.
The company was particularly interested in procedures and mecha-­
nisms where documents and decisions needed approval from one or more
people or departments. An assessment of the system for order process-­
ing of production identified 12 instances where somebody had to approve
someone else’s work. An analysis of these 12 instances showed that they
consumed nine days altogether. At the same time, only two of these were
true reviews undertaken before approval was given. Combined with fre-­
quent customer complaints about long ­ order-­processing times, it was
obvious that something needed to be done. In the 10 instances where no
real review was performed, the operative people performing the original
work were given authority to approve their own work. This saved seven to
eight days on average and an unknown amount of money.
Furthermore, they knew that many subsystems had been connected
throughout the years. Each of them performed the same or similar tasks.
The logical step, therefore, was to perform a redundancy elimination. For
this task, a detailed flowchart portraying the process and output from it was
used. Upon closer inspection, 16 highly similar steps were found—mainly
activities where different versions of order documents were generated. If
one globally available order document was produced, 13 steps in the pro-­
cess could be removed immediately. Over a period of four months, the
order system was totally redesigned to allow information that was entered
once to be available to the entire organization. This eliminated nine days of
work and dramatically improved the quality and accuracy of the process
and its output.
Finally, a ­value-­added analysis was undertaken. The current process
contained a very small proportion of NVA. The proportion of OVA was a
little higher, but not unreasonably so. By cutting the time consumed by
individual activities related to registering internal documents, this portion
could be further reduced. Figures 11.5 and 11.6, respectively, show the
­cost-­cycle time charts for the original process and the streamlined pro-­
cess. The cycle time was reduced by about 19 days, or 64 percent, and the
costs by about $1000 for a standard order.
176  Chapter Eleven




Cycle time

1 2 3 4 5 6 7 8 9 10 11 12131415 16 17 18 19 20 21 22 23 24 25 26 27 28293031 32
29 days

Figure 11.5  ­Cost-­cycle time chart for the original process.



Cycle time
1 2 3 4 5 6 7 8 9 10 11
9 days
Figure 11.6  ­Cost-­cycle time chart for the streamlined process.
Tools for Creating Improvements  177

11.2  Idealizing

Instead of starting from the existing process and analyzing its individual steps for
potential efficiency gains, as in the ­value-­added analysis, it is possible to free yourself
from the limitations the current process imposes. This is the principle underlying ideal‑
izing. The purpose of this technique is to imagine how good the ideal process, without
any form of waste or other inhibiting elements, could be. As a tool for generating pos‑
sible solutions, this can at times be quite effective. Even if it is obvious that the ideal
process cannot be implemented in practice, it can provide insight into how it should be
implemented. The differences between the ideal process and the current situation can
thus be used as a starting point for formulating solutions and improvement projects.
Idealizing is typically done as a group exercise, to ensure that as many ideas as
possible are captured. It is important that those performing the process be part of
this group, as they are the ones who dream of the ideal process. At the same time,
people unfamiliar with practical limitations can present refreshing ideas. Thus, a
mixture should be included in the group.
There really are no specific guidelines on how to use or perform idealizing.
Flowcharts can be used in this type of exercise. After constructing a flowchart for
the ideal process, it is interesting to compare it with the chart for the current pro‑
cess. The powerful graphic representation achieved by this approach is also a good
starting point. By studying the deviations between them, it is possible to define
measures to move the process toward the ideal state, as shown in Figure 11.7.
Idealizing is closely related to both clean sheet business process reengineering
(presented in section 11.6) and the A∆T analysis, whose objective is slightly dif‑
ferent (described in section 12.1).



Current Ideal
process E process



Figure 11.7  Analysis of the current and ideal processes in idealizing.

178  Chapter Eleven


A ­ medium-­sized printing company that did a number of smaller jobs for

different customers—typically advertising brochures, catalogs, and similar
material—was troubled for a long time by errors in the finished printed
material. The company decided to use idealizing to outline the ideal pro-­
cess that would ensure discovery of errors before actual printing. Seven
people from different functional areas in the company took on this job and
had many short meetings during a period of two weeks. The resulting ideal
process, as well as the current process, is shown in Figure 11.8.
The company spent the next three months moving toward this ideal
process. Although quite a lot remained to be done before the ideal process
was reached, not in the least due to the need for investments, the amount
of errors discovered after printing was reduced from an average of 18 per
month to 1 potential error.

Ideal Current
process process

Customer Customer

Jobs in Jobs on Jobs in

digital format paper or film digital format

Test print
Copying Plate production
in color

Proofreading and Test print in Printing and

customer approval black and white finishing

Proofreading by Complete printed

Plate production material

Printing and
Test run

Complete printed Control of test run

material against test print

Figure 11.8  Idealizing in a printing company.

Tools for Creating Improvements  179

11.3  QFD

QFD was presented in section 5.6 as one approach that can be used in creating
your business process improvement road map. As I have mentioned, QFD was
originally used as a tool to facilitate ­customer-­oriented product development. Over
time, the core of the technique has proved powerful enough to warrant application
in many different areas. For our purposes, the tool works both for initial improve‑
ment planning and for more detailed process improvement analysis.
The core of QFD is that it enables you to see relationships and connections
between ends and means as well as how each alternative mean impacts the col‑
lective set of ends. This principle, performed through constructing the house of
quality, is most suitable for identifying possible improvements in a business pro‑
cess—and not only improvements, but those most likely to produce the overall
best effect for the requirements that process faces. The general house of quality
was presented earlier but is shown again in Figure 11.9. Used for business process
improvement, the house has several different fields:

• What expresses the requirement set for the business process. The most
important requirements are those that represent external customers, but also
included are requirements from external suppliers and internal customers
and suppliers, as well as other concerns internal to the organization.
• To distinguish between the importance of each of these requirements, the
field of Importance can be used to assign weight factors.
• How shows the means used to satisfy the different requirements. If one
requirement is a short operation time for a service process, the mean could,
for instance, be expanding capacity in periods around demand peaks.
• In the Relational matrix in the middle of the chart, the relations between
ends (or requirements in this case) and means are analyzed.
• In the roof of the chart, the How vs. How field, connections between the
different means can be analyzed. If another mean for the service process is
staff reduction, motivated by the company’s demands for reduced costs, the
conflict between these two can be easily visualized in the roof.
• In the same way as for products in the original use of QFD, the Why field
can be used during process improvement to perform a simple benchmarking
against other organizations’ processes.
• Finally, the How much field is used to illustrate the results of the analysis.
For each connection in the relations diagram, the importance factor for
the requirements is multiplied by the weight factor for the relation, and
the sum of these products for each requirement is placed in the How much
field. The higher the score in this field, the more requirements can be sat‑
isfied by this mean.
180  Chapter Eleven

vs. how

How Why

What Points

How much

Figure 11.9  The general house of quality, the chart used by QFD.

As for the application of QFD in improvement planning, this is best explained

using an example.


Figure 11.10 presents a house of quality for the business process of product
distribution. The improvement team working on this process had identified
the customer requirements for the process and discussed which features
of the process influenced these requirements the most. Unable to agree on
this issue, the team members eventually decided to use QFD, resulting in
the example diagram.
The requirements to the process have been split into those posed by
the customers and internal demands. Each of the requirements has been
assigned an importance factor. For example, speedy and safe delivery are
considered essential, while the internal wishes to keep overtime down and
maintain a good inventory overview are ranked quite low. The typical situa-­
tion, which is completely in line with the norm, is that the external require-­
ments count more than the internal ones.
The next step involved brainstorming a number of means or process
features of the distribution process that were assumed to have a positive
impact on the requirements. At each connection point between a demand
and a process feature, an assessment was made as to what extent the
mean impacts the requirement, which is denoted by the symbols used in
the relations matrix.
In addition, data for these requirements were collected or estimated
for two competitors of the company. This information is presented in the
benchmarking column on the ­right-­hand side. As can be seen, the com-­
pany trails with regard to safe delivery, but both competitors, particularly
Company A, scored very well. Targets based on input from the competi-­
tors’ performance levels and strategic decisions have also been defined.
Tools for Creating Improvements  181

Good communication with manufacturing

of the process

Close cooperation with customers

Computer-based ordering system
Process features

Low finished-good inventories

Sound distribution planning

Own company

Company A

Company B
Sound work planning
Dependable trucks

Local warehouses

Careful handling

Customer and internal

requirements to the process
1 2 3 4

Speedy delivery 10
Safe delivery 10

No breakage 8
Complete deliveries 9
Correct shipment papers 4
No extra costs 5
Low costs 6

No need for overtime 3

Low need for space 5
Good overview 3
Max. 150,000
Less than 1 repair
per 60,000 miles



Own company
of metrics

Company A
Company B
Target 1
32% 281 1 19% 281
1 8% 123
12% 105 3 22% 315
2 9% 141
11% 102 1 7% 102
3 13% 194
13% 115 1 8% 115

Absolute importance
3 6% 90
3 4% 67

Relative importance
Organizational difficulty
Absolute importance




adjusted for difficulty

Relative importance




adjusted for difficulty

Roof Roof Weight
Strong positive Strong 9
Positive Medium 3
Negative Weak 1
Strong negative

Figure 11.10  House of quality for process improvement.

182  Chapter Eleven

For each mean, the team evaluated whether any relevant quantifiable
targets could be defined. As the matrix reveals, a target was defined for the
number of repairs beyond ordinary maintenance for the distribution trucks,
that is, less than one per 60,000 miles. However, far from every mean has
had targets defined this way. Therefore, in the field below, an assessment on
an objective scale has been made of the company’s own performance level,
that of the competitors, and the company’s targets. For dependability of the
trucks, for instance, the company is far behind the competitors, while the
target of one repair per 60,000 miles is quite ambitious.
The correlation symbols in the relations matrix, or more correctly the
weight factors for the symbols, are multiplied by the factors that represent the
importance of each of the requirement elements. For each process feature,
these products are added up and placed in the field of absolute importance.
The numbers in this field show the impact that the improvement will have on
the collective requirements. In the field immediately below, relative importance
has been calculated, that is, how large a percentage of the total impact the
single process feature accounts for. From these calculations, we see that the
three process features that stand out are sound distribution planning, good
communication with manufacturing, and a ­computer-­based order system.
Furthermore, a factor has been estimated that indicates how difficult
it would be to implement changes in these issues in the organization. The
higher the factor, the more difficulties are expected to be encountered. The
importance values are divided by these factors, which give the correspond-­
ing importance figures, adjusted for organizational difficulty. If this aspect is
included, it seems as if sound distribution planning is the process feature that
has the highest ability to impact the requirements and the easiest implemen-­
tation. This gives the company a clear sense of what element in the distribu-­
tion process to improve. It probably also makes sense to include the issue
that achieved the ­second-­highest score.
Finally, an evaluation has been performed in the roof of the house of
quality with regard to synergy between the different means. If focusing on
the distribution planning, we see that positive effects can also be expected
from improved communication with manufacturing and a ­computer-­based
order system. Local warehouses can, on the other hand, complicate the dis-­
tribution planning task.

Even if this example does not capture all aspects of the use of QFD for improve‑
ment, it at least demonstrates how the volume of the voice of the customer can be
sufficiently turned up to be heard when creating process improvements.

11.4  Statistical Process Control/Control Chart

A useful characteristic trait of this world and the processes carried out in it is
that stabilized processes display a logical statistical behavior. This means that the
Tools for Creating Improvements  183

results from a process normally fall within certain specifiable limits. This fact can
be used for monitoring and improving an organization’s processes.

11.4.1  Definitions of Variation

Before the actual tool of statistical process control (SPC) is presented, it is neces‑
sary to define some central terms in statistics. First, it is important to understand
the difference between chronic and sporadic variation:

• Chronic variation is variation that is inherent in the process and is caused

by a long line of factors, without one single cause for the deviation. It is,
in other words, caused by something in the system; therefore, it is often
referred to as common cause variation. This variation is normal for the
process and must be expected and lived with, unless the process itself is
changed. If, for example, the time it takes to complete a mail route in a large
office building is measured, it will invariably vary due to factors such as the
pace of the mail carrier, the number of letters to be delivered, the number
and duration of breaks, the waiting time for elevators, and so on.
• Sporadic variation, also known as special cause variation, is caused by fac‑
tors that appear infrequently and is often larger than the chronic variation.
The deviation can usually be traced to one single cause and can be elimi‑
nated by removing that cause. For the mail route, this could be variation
caused by a new mail carrier, incorrectly sorted mail, or an ­ out-­of-order

The purpose of SPC is to classify variation according to these two groups and thus
facilitate further improvement of the process.

11.4.2  Basic Statistics

To understand the principles of SPC, it is absolutely essential to first understand
the basic principles of statistics, on which SPC is based. Most processes studied in
an organization can be described reasonably accurately using two variables:

• The arithmetic mean value of the process measure being used—for exam‑
ple, the diameter of a hole being drilled. The statistical term for this value is
the expected value, and the expression for it is:

1 n
X= ∑X .
n i=1 i

• The standard deviation of the process—that is, a variable that says some‑
thing about how much variation should be expected from the process. If we
want to drill a hole with a diameter of 7.9 millimeters—that is, an expected
184  Chapter Eleven

value of 7.9—the standard deviation could be 0.15 millimeters. The symbol

for the standard deviation is σx, and the formula for calculating it is:

∑(X n
− X )2
σx = i =1
( n − 1)
Actually, it is not totally correct to use σ for the standard deviation. σ refers to the
standard deviation for the distribution itself, while the standard deviation for the
sample being taken is labeled s. This connotation has been used throughout the rest
of the book.
If the observation or measurements of the process are evenly distributed around
the expected value, we can use a normal distribution to describe the process. This
is often the case for stable processes. The normal distribution in statistical matters
is a special distribution for several reasons:

• It is symmetric; that is, the likelihood of the process giving a result that
is larger than the expected value is as high as for a result smaller than the
expected value.
• If the distribution function graph is drawn, it looks like a bell; thus, another
term for this distribution is bell-shaped. Figure 11.11 shows a distribution
curve that expresses the likelihood that the process measure will assume a
certain value.
• From the graph, it can be seen that the distribution curve crosses the x axis
at two points, these points being respectively three times the standard devia‑
tion to the left of the expected value and equally far to the right of it. This
means that the probability of getting a value that is less than the expected
value minus three times the standard deviation, or a value that is larger than
the expected value plus three times the standard deviation, is practically
zero. This is a basic quality of the normal distribution, namely, that the
maximum deviation that can be experienced lies within 3σx. To be totally
correct, 68.26 percent of all values will fall within one standard deviation,
95.45 percent within two, and 99.73 percent within three. The area below
the distribution curve equals 1, which is another way of saying that all the
probability is gathered within these limits of +3σx.

Because all results from a process following this distribution are expected to fall
within the total spread of six standard deviations, data that fall outside this area are
sporadic deviations due to special causes. To monitor the process, control charts
are used that show the process limits of three standard deviations in each direction
(see Figure 11.11), as well as the continuous measurements of the process. The
construction of such charts is discussed in section 11.4.4 on page 189.
Tools for Creating Improvements  185

Lower process Probability Upper process

limit, LPL x limit, UPL x


X 1sx 2sx 3sx


Figure 11.11  Normal distribution curve.

When it comes to the quality of a process, the narrower the chronic variation,
the better the process. SPC is used to reduce this natural variation and to report
occurrences of any sporadic variation, whose cause must be found. Before this is
possible, the process must be statistically stable, that is, having a constant expected
value and variation width.
With these simple statistical principles and related rules, it is thus possible to
control a business process and discover whether the centering is changing, whether
the variation width is increasing, whether variation is seen beyond the expected
limits, but also what normal variation must be expected. If a process measure lies
far from the expected value but within three times the standard deviation, it might
be tempting to adjust it. But this introduces lopsidedness into the process, and in
the future, deviations to the other side will occur. Attempts to rectify this by adjust‑
ing the process the other way usually moves the deviation to the other side. This
evil circle that can result from adjusting processes that display only normal varia‑
tion is popularly termed tampering and can be avoided by proper use of SPC.
Far from just physical, geometric measures related to manufacturing processes
can be kept under statistic monitoring. By using a number of different control charts,
many different processes and process measures can be made subject to SPC:

• The time it takes to perform a task, physical or administrative

• The costs incurred by a process, physical or administrative
• The number of errors or defects produced by a process, physical or
• Specific geometric measures such as length, diameter, and so on
186  Chapter Eleven

By defining control limits on the basis of the expected and normal variation inher‑
ent in the process, continuous measures of the process’s characteristic features
are plotted in the chart. Depending on the results, different measures are taken.
Which measures are appropriate in which situations is discussed in detail in sec‑
tion 11.4.5. Thus, SPC presents a powerful tool for monitoring and improving
business processes.

11.4.3  Types of Control Charts

The normal distribution and how control limits can be based on three standard
deviations is the simplest type of control chart. This type is not suited for all sit‑
uations, however, so other types have been defined. First, there is a difference
between two basic types of variables:

• Variable data are variables that are based on measurement—for example,

meters, hours, volts, and so on. These variables are measured on a continu‑
ous scale and with reasonably high accuracy.
• Attribute data are variables measured by counting or classifying nonmea‑
surable characteristics (for example, acceptable/nonacceptable, brown/red,
assembled/not assembled, and so on). Attribute data are quantified by inte‑
gers or intervals of values. Variable data can be converted to attribute data
(for example, length by defining a limit for acceptable and nonacceptable).

For variable data, a pair of charts must be used: one to control the centering of the
process and one to control the variation width. For attribute data, only one chart is
required. There are two weaknesses related to the use of a chart like the one shown
in Figure 11.11:

• The values that have been measured are not necessarily normally distrib‑
uted, which renders the process limits uncertain
• Such a chart is not very sensitive, as values within the process limits can be
measured long after the process has changed its center or the variation width
has increased

These weaknesses are usually remedied by using mean values for a group of mea‑
sures instead of single values as the basis for the chart. The advantage of this is that
mean values for a group of measures, almost regardless of the original distribution
of the individual measures, approximately follow a normal distribution. Thus, the
first weakness is avoided. At the same time, a chart based on such values is far
more sensitive to changes in the centering of the process, which also counters
the second weakness. In opposition to the process limits in charts based on single
values, the limits in charts based on average values are termed control limits. The
Tools for Creating Improvements  187

expected value for a mean variable consisting of the average m groups of single
measures is then labeled:
1 m
X= ∑ X m.
m j =1

The lower control limit (LCL) and the upper control limit (UCL), respectively,
are calculated by subtracting and adding three times the standard deviation for the
mean distribution. The standard deviation for a mean variable from a group of n
individual values is calculated as follows:
sx = .
There are seven chart types that are most frequently used, classified according
to the variable being measured and the number of measurements that exist. These
are shown in Figure 11.12 and described in more detail after a brief overview
(Swanson, 1995).
For variable data, there are three types of control charts:

– –
• X and s, where the X chart plots the mean value for groups of measure‑
ments, and the s chart monitors the variation width of the average measure‑
ments. The number of measurements in each group should normally be
higher than 10.
– –
• X and R, where the X chart displays the same average measurements for
the expected value. Because the number of measurements in each group
is often lower than 10, and because the term range R, defined as the dif‑
ference between the largest and smallest value in the group, is easier to
understand than s, an R chart is often used to monitor the variation width.

For this reason, the chart pair X and s will not be described further.
• x and R, where the x chart is based on the type of individual measure‑
ments first described. This type of chart does not quickly detect changes
in the process. The reason for occasionally using this chart is that the
measurement costs are lower. At the same time, it may be the only alter‑
native if the variable changes very quickly, which makes it impossible
to make group measurements—for example, during a measurement of
temperature. The variation width is controlled using an R chart, but
in this case, R means the difference between the current and previous

Before the charts of attribute data are presented, it is important to understand the
difference between defective and defects. Defective items are scrapped because
they do not satisfy the requirements, and defects are units that have some kind of
188  Chapter Eleven

Chart type

n > 10 X and s

n < 10 X and R

n=1 x and R
charts n is constant p or np
n is variable p
n is constant c or u

n is variable u

Figure 11.12  Types of control charts and when to use them.

weakness but can be accepted because the specifications allow some degree of
weakness. For attribute data, there are four chart types:

• np, which is used to monitor the number of defective items in a sample of

constant size. The process measure used could be the number of defective
cars manufactured in a day or the number of errors detected during order
processing in a day.
• p, which is used for the number of defective items in samples of either con‑
stant or varying size.
• c, which is used to monitor the number of defects occurring in samples of
constant size. Examples are the number of errors in a ­five-­page document or
the number of defects found on a circuit board.
• u, a more general variant of the c chart, which can also be used for samples
of varying size.

Notice that the np chart plots values that cannot exceed the sample size, because
only a portion of the items can be defective. Because the c and u charts are used for
the number of defects, this number might very well exceed the sample size. The
differences between defective and defects can be further illustrated by an example.
When the surfaces of 10 items were inspected, 5 items were found to have a total
of 11 inconsistencies, while the other 5 were perfect. An np chart would plot five
defectives if the specifications did not allow such inconsistencies. A c chart would
plot 11, as this was the total number of errors, but would not say anything about
how many would have to be scrapped.
Tools for Creating Improvements  189

11.4.4  Constructing Control Charts

In this section, a general procedure for using SPC will be introduced. Afterward,
specific instructions will be given for calculating control limits for the different
types of charts and for how the charts are used.
Generally, when using SPC, the steps are:

1. Determine which type of data will be collected and controlled—variable

or attribute data—and the expected number of samples.
2. Following the guidelines of Figure 11.12, select a suitable type of control
3. Collect the data required to calculate the control limits.
4. Construct the control chart by letting the y axis represent the variable being
measured. The x axis represents the process and is divided into measurement
segments (time, subgroups, measurement number, and so on). Mark suit‑
able intervals on the axes that make it easy to plot the measured values.
5. Calculate the lower and upper control limits and, where required, the mean
and range of the data set. If one of the points falls outside the control lim‑
its, this point is removed and new limits are calculated. If further points
fall outside, the process is not statistically stable and the cause must be
found and eliminated before the use of SPC can be continued.
6. Draw the control limits in the chart along with the existing measurements.
7. Continue plotting values as they are collected from the running process.
The meaning of the points and the necessary actions are interpreted fol‑
lowing the guidelines given in the next section.

While this is a manual procedure for using SPC, many PC programs designed for
quality improvement contain modules for constructing control charts. The rest of this
section provides detailed guidelines for constructing the individual chart types.

X and R Charts

As mentioned, the X and R charts are a pair of charts in which the first focuses on
the centering of the process and the other controls the variation width. To construct
such a chart pair, the following must be done:

1. Collect the data; typically, more than 125 measurements of recent date are
2. Divide the data points into logical subgroups, where at least 25 subgroups
are required to calculate the control limits. Subgroups should consist of
190  Chapter Eleven

data points that logically belong together; that is, they are measured under
the same conditions, close to each other in time. Usually, there will be two
to five data points in each subgroup, which defines the sample size n.

3. For each subgroup, calculate the mean value X and the range R (that is, the
difference between the largest and smallest value in the subgroup). For the
mean value, add one decimal place to what is used in the measurements

4. Calculate the mean value for the entire data set X by either averaging all
the individual measurements or averaging the calculated means for the sub‑
groups. Here, add two decimal places to what is used in the original measure‑

ments. This value will represent the center line of the X chart.

5. Calculate the average range R by finding the mean of all the range values
calculated in step 3. This value represents the center line of the R chart.

6. Calculate the control limits for both charts. To simplify the calculations,
calculation factors have been developed that correspond to a deviation of +/–
three standard deviations for sample sizes from 1 to 10 (see Table 11.1). For

larger sample sizes, an X and s chart should be used. Notice that for all sub‑
groups of six or fewer measurements, no LCL for the R chart is calculated.

For the X chart, the control limits are calculated as follows, using two
more decimal places than in the data points:

Table 11.1  Calculation factors.

Number of data Factors for the Factors for the
points in the Factors for the LCL of the R UCL of the R

subgroup n X chart, A2 chart, D3 chart, D4
 2 1.880 — 3.267
 3 1.023 — 2.575
 4 0.729 — 2.282
 5 0.577 — 2.115
 6 0.483 — 2.004
 7 0.419 0.076 1.924
 8 0.373 0.136 1.864
 9 0.337 0.184 1.816
10 0.308 0.223 1.777
Tools for Creating Improvements  191

LCL X = X − A2 R ,
UCL X = X + A2 R.

For the R chart, the corresponding formulae for the control limits are as fol‑
lows, using one more decimal place than in the individual measurements:

LCL R = D3 R ,
UCL R = D4 R.

7. Draw the axes, adding the correct variable notations. Divide the axes into
suitable intervals and indicate any other relevant information (for example,
when the measurement was performed, by whom, or sample size). Draw
the center line and control limits, using a solid line for the center line and
dotted lines for the control limits.
8. Plot the data points as new measurements are made. Interpretation and
actions are explained in the next section.

Examples of these types of charts are shown in Figures 11.13 and 11.14.

x and R Charts

The x and R charts are a pair of charts used for the centering and the variation, respec‑
tively. As previously mentioned, this is a less sensitive chart than the preceding one,
but one that is useful where measurement is very expensive or difficult. The construc‑

tion follows almost the same steps as the X and R charts, using only slightly different
calculation formulae for the control limits. The steps are as follows:


11.00 UCL = 10.78


X = 6.12


1.00 LCL = 1.46

n=6 Subgroups

Figure 11.13  Example of an X chart.
192  Chapter Eleven

UCL = 11.3


R = 5.7



Figure 11.14  Example of an R chart.

1. Collect the data; typically, more than 25 measurements of recent date are

2. Calculate the mean value for the entire data set X . This value will represent
the center line of the x chart.
3. Calculate the ranges R for the data set, where R is defined as the numerical
distance between one data point and the preceding one. Thus, for the first
data point, no R is calculated, resulting in the number of R values being
one less than the total number of measurements.

4. Calculate the average range R for the entire data set by finding the mean
of all the range values calculated in step 3. This value represents the center
line of the R chart.
5. Calculate the control limits for both charts. To simplify the calculations,
calculation factors have been developed that correspond to a deviation of
+/– three standard deviations. For the x chart, this factor is 2.66, which gives
the following calculation formulae for the control limits of this chart:

LCL x = X − 2.66 R ,
UCL x = X + 2.66 R.

For the R chart, we know that the sample size is always two, as the range
is the distance between two measurements. Thus, we can use the factor n =
2 from Table 11.1. For the UCL, we see that the factor is 3.267, while we
remember that for sample sizes of six or fewer, there is not a lower limit.
The expression for the UCL is:

UCL R = 3.267 R.
Tools for Creating Improvements  193

6. Draw the axes, adding the correct variable notations. Divide the axes into
suitable intervals and indicate any other relevant information (for example,
when the measurement was performed, by whom, or sample size). Fur‑
thermore, draw the center line and control limits, using a solid line for the
center line and dotted lines for the control limits.
7. Plot the data points as new measurements are made. Interpretation and
actions are explained in the next section.

The R chart will look exactly like the example in Figure 11.14, and the x chart will
be very similar to the example in Figure 11.13, but with a different variable name
for the y axis.

p Chart
Unlike the charts described thus far, which are applied to variable data, the charts
for attribute data are not used in pairs. Each chart for attribute data covers both the
centering and the variation width. The procedure to construct a p chart, where p
represents the portion of defective units, is as follows:

1. Collect the data set to be used for the construction of the chart and divide
the measurements into subgroups sorted, for example, by date or series.
Generally, there should be more than 25 subgroups of a sample size n,
with 50 or more in each subgroup, and the average number of defectives
np should be larger than three or four. The data set should come from a
process that was recently run.
2. For each subgroup, the portion of defectives p is calculated simply by
dividing the total number of defectives by the sample size n and multiply‑
ing by 100 to produce a percentage figure.
3. The center line of the chart is the average defective portion for the entire
data set p–, which is arrived at by dividing the total number of defectives
by the total number of units inspected.
4. The control limits are calculated using the following formulae:

LCL p = p − 3
p 1− p ),

UCL p = p + 3
p 1− p ).
If the numerical value for the LCL is negative, no lower limit can be cal‑
culated, so it is left out of the chart. Also notice that as the sample size n
changes with time, the control limits also change. This means that for each
subgroup of a different sample size, new control limits must be calculated.
194  Chapter Eleven

As this is very cumbersome, a simplification has been devised that gives

accurate results. If all subgroups have a sample size that lies within +/– 20
percent of the average sample size, then in the preceding formulae, the n of
the denominators can be replaced by the average sample size n–. For sub‑
groups of sample sizes outside this 20 percent limit, separate limits must
be calculated.
5. Draw the axes, adding the correct variable notations. Divide the axes into
suitable intervals and indicate any other relevant information (for example,
when the measurement was performed, by whom, or sample size). Draw
the center line and control limits, using a solid line for the center line and
dotted lines for the control limits.
6. Plot the data points as new measurements are made. Interpretation and
actions are explained in the next section.

All four chart types for attribute data look like the preceding examples but use
a different variable for the center line. Thus, no separate examples have been

np Chart
Whereas the p chart can be used for samples of varying size, the np chart can be
used only for samples of constant size. Otherwise, the procedure is quite similar to
the one used for the p chart:

1. Collect the data set to be used for the construction of the chart and divide
the measurements into subgroups. Generally, there should be more than 25
subgroups of a constant sample size n of 50 or more, with an average num‑
ber of defectives np that is larger than three or four. The data set should
come from a process that was recently run.
2. The center line of the chart is the average defective portion per subgroup
np, which is arrived at by dividing the total number of defective units by
the number of subgroups.
3. The control limits are calculated using the following formulae:

LCL np = np − 3 np 1 − p , )

UCL np = np + 3 np 1 − p . )
If the numerical value for the LCL is negative, no lower limit can be cal‑
culated and it is therefore left out of the chart.
4. Draw the axes, adding the correct variable notations. Divide the axes into
suitable intervals and indicate any other relevant information (for example,
Tools for Creating Improvements  195

when the measurement was performed, by whom, or sample size). Draw

the center line and control limits, using a solid line for the center line and
dotted lines for the control limits.
5. Plot the data points as new measurements are made. Interpretation and
actions are explained in the next section.

u Chart

As previously mentioned, the main difference between p/np charts and u/c charts
is that the former are used for defective units and the latter are used for the number
of defects. Defects do not necessarily mean that the unit is defective. The u chart
is used for constant and varying sample sizes and is constructed according to the
following steps:

1. Collect the data points; typically, there should be more than 25 data pairs
consisting of the number of units inspected n and the number of defects
found c. The data set should be recent, and the number of inspected units
should be more than 50 with an average number of defects per subgroup
of more than one to three.
2. For each subgroup, the portion of defects u is calculated (that is, the num‑
ber of defects c is divided by the sample size n). Because the sample size
varies, this must be done for each subgroup.
3. The center line of the chart is the average defect portion u–. This is found
by dividing the total number of defects in the entire data set by the total
number of inspected units.
4. The control limits are calculated using the following formulae:

LCLu = u − 3 ,

UCLu = u + 3 .
If the numerical value for the LCL is negative, no lower limit can be cal‑
culated, and it is left out of the chart. Notice that as the sample size n
changes with time, the control limits also change. This means that for each
subgroup of a different sample size, new control limits must be calculated.
A similar simplification to the one used for the p chart can be used in this
case. If all subgroups have a sample size that lies within +/– 20 percent of
the average sample size, the n of the denominators can be replaced by the
average sample size n– in the preceding formulae. For subgroups of sample
sizes outside this 20 percent limit, separate limits must be calculated.
196  Chapter Eleven

5. Draw the axes, adding the correct variable notations. Divide the axes into
suitable intervals and indicate any other relevant information (for example,
when the measurement was performed, by whom, or sample size). Draw
the center line and control limits, using a solid line for the center line and
dotted lines for the control limits.
6. Plot the data points as new measurements are made. Interpretation and
actions are explained in the next section.

c Chart
For the c chart, the chart for portion of defects in samples of constant size, the
procedure is quite similar to that used for the u chart:

1. Collect the data points and divide them into subgroups. Typically, there
should be more than 25 subgroups of more than 50 recent measurements,
with an average number of defects per subgroup of more than two to
2. The center line of the chart is the average number of defects c–. This is
found by dividing the total number of defects in the entire data set by the
number of subgroups.
3. The control limits are calculated using the following formulae:
LCL c = c − 3 c ,

UCL c = c + 3 c .
If the numerical value for the LCL is negative, no lower limit can be cal‑
culated and it is therefore left out of the chart.
4. Draw the axes, adding the correct variable notations. Divide the axes into
suitable intervals and indicate any other relevant information (for example,
when the measurement was performed, by whom, or sample size). Fur‑
thermore, draw the center line and control limits, using a solid line for the
center line and dotted lines for the control limits.
5. Plot the data points as new measurements are made. Interpretation and
actions are explained in the next section.

11.4.5  Interpreting the Control Charts

It is important to be able to select the right chart type for a given situation and
to construct that chart. The chart is useful only when we are able to interpret the
meaning of the data points and understand what actions should be taken.
Most of the variation seen in a control chart is chronic variation, which is a
natural part of the process. This variation requires no adjustments or tampering
Tools for Creating Improvements  197

with the process to keep it under control. However, it should be a constant objec‑
tive to improve the process, that is, improve the expected value and/or reduce the
natural variation width. If the process for issuing an important document is stabi‑
lized at an expected value of 3 hours and 20 minutes and a standard deviation of 30
minutes, SPC can very well be used for registering that after some time, it has been
improved to an average of 2 hours and a standard deviation of 15 minutes.
Only a small portion of variation can be attributed to special causes—what we
called sporadic variation, which requires special attention or action. Results that
fall outside the control limits obviously constitute such variation, as do patterns of
variation within the limits, which can signal the existence of special causes. In the
following list, some examples of control charts displaying different types of varia‑
tion and the meaning of these are presented:

1. The natural progression of a process under control that displays only varia‑
tion that can be statistically expected is shown in the previous examples
(see, for instance, Figure 11.13). Unless there is a wish to improve such a
process, there is no need to do anything about it.
2. Figure 11.15 shows a control chart in which one point has fallen outside
the UCL. It is statistically very improbable that this should happen. With
limits based on three times the standard deviation, only two or three of
1000 will fall outside. There is, however, sufficient probability that one
isolated point, a wild card, can exist without there being any special cause
for it. It is not alarming as long as no more points appear outside the limits
in a short time.
3. If several points fall outside the limits, as shown in Figure 11.16, they
undoubtedly are the result of special causes. A search for the causes must
then be started in order to eliminate them, thus preventing a recurrence.
4. A process is changing or unstable if two of three consecutive points on the
same side of the center line are more than two standard ­deviations away


Center line


Figure 11.15  Control chart with one point outside the control limits.
198  Chapter Eleven


Center line


Figure 11.16  Chart with several points outside the control limits.

from it, or if four of five consecutive points on the same side of the center
line are more than one standard deviation away from it. One example is
shown in Figure 11.17. Such a change in level indicates a special cause—
for example, a new operator has taken over, a new material has been intro‑
duced, or a new tool has been put to use. If the cause is found, new control
limits must be calculated under the new conditions before the monitoring
is continued. Incidentally, if this is due to a change in operators, this is a
good opportunity for improving the process. The operator producing the
best results can teach the others what he or she is doing—a form of bench‑
marking between operators.
5. Another sign indicating that the centering of the process has changed is if
relatively many points fall on the same side of the center line. Initially, we
would expect an equal number of points on either side of the center line.
If series of seven, 10 of 11, 12 of 14, or 14 of 17 fall on the same side, this
indicates systematic variation. The causes of the change must be found
and eliminated. An example is shown in Figure 11.18.
6. Furthermore, we expect that the measurements would alternately increase and
decrease from the previous measurement. A trend where six or more measure‑


2 standard deviations

1 standard deviation
Center line


Figure 11.17  Process with change in level.

Tools for Creating Improvements  199


Center line


Figure 11.18  A series of 12 or 14 points on the same side of the center line.

ments show a constant increase or decrease, as shown in Figure 11.19, is a

signal that the process is changing. Again, the causes must be found and
7. While alternating measurements should be expected on both sides of
the center line, it is statistically improbable that 14 or more points
should systematically alternate from one side to the other. This is
shown in Figure 11.20, and it is probably caused by tampering with
the process, that is, continually changing the centering as a result of
individual measurements.
8. The last alternative when it comes to the appearance of the control chart is
shown in Figure 11.21. Here, it is possible to see clear cycles of the measure‑
ments, which is probably caused by rotation of operators, a process span‑
ning several shifts, temperature changes throughout the day, and so on.
When using SPC, we must combine a basic understanding of the process being
monitored with knowledge about the appearance of the control charts. In this way,
SPC is a powerful tool that can be used both for inspecting the process and for
improving it. Signals from the chart that indicate something is wrong must be
followed up quickly. It is quite common in companies using SPC that when the


Center line


Figure 11.19  Process with trend.

200  Chapter Eleven


Center line


Figure 11.20  Process probably involving tampering.


Center line


Figure 11.21  Chart displaying cycles.

operator is not immediately able to determine special causes, a ­ cross-­functional

team is established to continue working on the case. Such groups were discussed
in section 6.4.

11.4.6  Process Capability

Because it is connected to SPC, it is logical to cover the concept of process capa‑
bility here. This concept is used to determine whether a process, given its natural
variation, is capable of satisfying the specifications it is expected to produce within.
A process can be under statistical control and still not be able to produce units or
results that satisfy the specifications. When calculating the process capability, an
index is used to compare the specifications with the control limits.
Specifications normally define a nominal value, the desired result, plus or
minus a value that determines how large a deviation from the desired value can be
accepted without negatively impacting the performance of the process. Normally,
­two-­sided tolerance limits are specified (both an upper and lower limit, UTL and
LTL), but for some processes, only one of the limits is specified.
The less complex of the two capability indexes, Cp compares the difference
between the tolerance limits with the natural variation of the process. If the dis‑
Tools for Creating Improvements  201

tance between the tolerance limits exceeds six times the standard deviation, the
process can stay within the specification, provided it has been centered on the
nominal value. The index is calculated as follows:
Cp = .

The standard deviation σ for the process is rarely known and must therefore
be estimated. This is done by calculating the average range, that is, the difference
between the highest and lowest value within a subgroup, for all subgroups. This is

usually R. An estimate of the standard deviation is calculated using the following
σ= .

The factor d2 has already been calculated and can be found in Table 11.2, listed
according to sample size n.
A general rule for this simple index is that it should be greater than 1.2. This
ensures that the process variation is sufficiently small compared with the specifi‑
cations. The problem with this index, however, is that it does not capture the fact
that the process can be unevenly centered compared with the nominal value. It is of
little help if the variation is small compared with the tolerance limits, if the center
of it is located close to one of the limits.
Therefore, a less sensitive index has been introduced, Cpk. This index covers
the variation of the process as well as the centering of it. It actually consists of two

Table 11.2  Factors for estimating the standard deviation.

Sample size n d2
 2 1.128
 3 1.693
 4 2.059
 5 2.326
 6 2.534
 7 2.704
 8 2.847
 9 2.970
10 3.078
202  Chapter Eleven

indexes: one lower, Cpl, and one upper, Cpu. The smallest of these is used as Cpk.
They are calculated as follows:

C pl = ,

C pu = .

X in this case means the average value for the process. Values below 1.0 mean
that a 100 percent inspection must be performed. If any changes in the process’s
centering can be accepted, Cpk must be greater than 1.33, and preferably 2.0 or
even larger. The ideal is that the index is over 5. This index can be used to deter‑
mine whether the existing process is at all able to produce what is required, to set
realistic tolerance limits suited to the process, and to negotiate targets for suppliers’
process capability.


A manufacturer of various types of special cables experienced high quan-­

tities of cables with defects. In fact, defects were so common that pro-­
ducing excess quantities to accommodate for this had become the norm.
The defects happened in spite of the stringent inspection along the entire
manufacturing line. Operators were responsible for undertaking visual con-­
trol of the cable they were working on.
To investigate what caused the defects and possibly correct them, the
company wanted to try SPC. One day’s production was viewed as a sam-­
ple. Since a cable with defects in practice is always defective and has to be
scrapped, a chart type covering defectives was selected. A p chart based
on data from the previous month was first constructed. The center line was
calculated at an average defective portion of 11.5 percent, which gave the
chart shown in Figure 11.22.
As the chart shows, the process was actually under control, with an
average portion of defectives of 11.5 percent. This was shocking news
that revealed how large a portion of defectives the company generated.
To learn why the process was performing so poorly, the data points were
divided into other subgroups. Instead of viewing the data from one day’s
production, data representing a week’s production from each individ-­
ual operator were collected. A new p chart, shown in Figure 11.23, was
This chart quite clearly reveals that while the other operators gen-­
erated few defectives, operators 5 and 12 were at the other end of the
scale. After these two operators were studied more closely, it was found
that both had such poor eyesight that they did not discover the many
Tools for Creating Improvements  203


p = 11.5



1.00 Working days

1 5 10 15 20

Figure 11.22  p chart for one day’s production.


p = 11.5


1.00 Operator number

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Figure 11.23  Control chart divided into operators.

smaller defects that caused defective units. The company covered the
cost of glasses for them, which quickly gave results. The average portion
of defectives was continually reduced and after three months was down
to 1.85 percent. This number was still too high, but it was significantly
better than the starting point. After this, the differences between the
operators were so small that common charts could be used for the entire
production of a day.
204  Chapter Eleven

11.5  Six Sigma

Having included SPC as an improvement tool, it is virtually impossible not to men‑

tion Six Sigma. Some will undoubtedly argue that Six Sigma should precede SPC,
and I will not argue about that, but the fact is that they are highly related. Histori‑
cally, SPC is an antecedent of Six Sigma (Quinn, in Bertels, 2003), but whether
it is correct to include Six Sigma as a tool is debatable. As you will see for both
business process reengineering and benchmarking later in this chapter, Six Sigma
is more of a meta-tool. More specific tools and techniques are included in Six
Sigma, and the concept itself outlines more of a model for how these are used in a
concerted effort. In addition, Six Sigma is as much a leadership model and a state
of mind as it is an improvement approach.
The reason that it is related to SPC should, however, be obvious. Six Sigma
takes its name (indirectly) from the +/– 3 standard deviations explained pre‑
viously and shown in Figure 11.11. It is generally accepted that Motorola
pioneered Six Sigma from the mid-1980s onward, out of a realization that mea‑
suring defect rates as parts per hundred was far from good enough in the emerg‑
ing global competition from Europe, Japan, Korea, and China. Staying within
the +/– three sigma limits, 0.135 percent of the output is still unacceptable. To
win that game, defects need to be measured in parts per million (ppm), which is
how current purchasing agreements in many industries—for example, automo‑
tive and electronics—define quality levels. By reducing the variation (sigma)
until the distance between the mean value and the specification limits is six
sigma, much better performance is achieved. Even though the specification lim‑
its are not moved, since the value of sigma is smaller, the new average defect
rate is only 3.4 per million.
One of the first companies to win the newly introduced Malcolm Baldrige
National Quality Award in 1988, Motorola shared its methods broadly, making its
Six Sigma Quality program accessible to other companies. The number of com‑
panies using Six Sigma, along with the number of books and consultants offering
advice, has dramatically grown since then. For our purposes, I especially find the
overall Six Sigma process and how Six Sigma work is organized in companies to
be of the highest relevance. For more detailed treatments of these and other aspects
of Six Sigma, consult the following sources: Meisel, Babb, Marsh, and Schlicht‑
ing, 2007; Przekop, 2006; and De Feo and Barnard, 2004.

11.5.1  The Six Sigma Process

The basic improvement process outlined in the Six Sigma concept is the DMAIC
process: define, measure, analyze, improve, and control. You will certainly recog‑
nize facets of the plan, do, check, act cycle in this process, but a distinguishing fac‑
tor setting the DMAIC process apart is how systematic and rigorous it is. Another
Tools for Creating Improvements  205

key factor is that Six Sigma projects are selected by the company’s management in
order to maximize payoff. This is completely in line with the overall methodology
of this book, advocating the development of an overall improvement road map to
ensure coherency among initiatives.
The five phases of the DMAIC process focus on the following issues:

• Define: The purpose is to scope the project by defining its focus, objec‑
tives, and desired benefits. Important outputs include a project charter, an
understanding of the business process that causes the problems addressed,
and insight into the stakeholders of the process. In this phase, many of the
tools discussed earlier in the book are applied, including process mapping
(where Six Sigma often terms a process map a SIPOC [suppliers, inputs,
process steps, outputs, and customers] map), stakeholder analysis (some‑
times including Kano model requirements), and VOC (voice of the cus‑
tomer, a variant of QFD).
• Measure: This can be compared to the phase for collecting data about the
performance shortcoming presented in this book, with the purpose of nar‑
rowing down the range of potential causes of problems and understand‑
ing the capabilities of the current process. Achievements to be attained by
the end of the measure phase include metrics for the process, that perfor‑
mance data have been collected and displayed to understand the variation
of the process, and a calculated process capability. Typical tools used are
check sheets for data collection; data presentation and analysis tools such
as Pareto charts, histograms, run charts, and so on; and formulas for process
• Analyze: The purpose here is to better understand the problem: identifying
possible causes, organizing these causes, and validating which are the root
causes. Comparing again with the structure of this book, this is quite similar
to the performance shortcoming analysis stage. Completing this phase of
the Six Sigma project normally means producing a ­cause-­and-effect under‑
standing, creating a more detailed process map than the SIPOC chart, and
verifying which causes have the most impact on the process. Tools for this
stage include detailed process mapping, ­ value-­added analysis, the ­ cause-
­and-effect chart, the scatter chart, and more advanced hypothesis testing.
• Improve: As the name clearly indicates, this phase of the Six Sigma proj‑
ect aims to implement improvements that remedy the identified problems,
mirroring the phase of creating improvements in our overall improvement
model. Outcomes from the improve phase are a ­ cost-­benefit analysis for
the suggested improvement, a risk analysis for the implementation of the
solution, an implementation plan, often a pilot implementation to verify the
206  Chapter Eleven

solution, and finally an implemented solution to the problems addressed by

the project. Typical tools applied are different idea-generation techniques,
­cost-­benefit analysis, stakeholder analysis, force field analysis, and project
management techniques.
• Control: The purpose of this phase is to ensure that the process does not
revert to its old state and to make sure that the solution actually works.
Completing this final phase means documenting the results of the improve‑
ment, creating a plan for sustaining the improvement, devising a training
plan to ensure minimal variation in the new process, and documenting the
entire project. This normally relies on tools such as control charts, process
capability calculations, and techniques for standardization.

From this presentation, it should be quite obvious why many, including me, are
reluctant to label Six Sigma a tool. I should also mention that the DMAIC process
is not the only overall Six Sigma process model that exists. You might be familiar
with the term design for X, where the X can be replaced with words such as manu‑
facturing, logistics, ­eco-­friendliness, and so on. Another replacement for X could
be minimal variation, which would then translate into design for Six Sigma. For
developing new products or for completely overhauling processes, there is in fact a
technique called Design for Six Sigma (DFSS). Under this heading, an alternative
to DMAIC has been developed: DMADV (define, measure, analyze, design, and
review). DMAIC and DMADV are quite similar, so I will not present the details of
this alternative process model.
Common to both is that each phase is followed by a review of the outcomes.
These are called toll gate reviews and are designed to ensure that one phase has
produced the required platform for the next before the team is allowed to move on.
Such reviews are held with people external to the project team. In Six Sigma, these
people have special titles, which are described in the next section.

11.5.2  Six Sigma in the Organization

To ensure that Six Sigma is carried out according to the rigor prescribed, an organi‑
zational structure has been designed to surround the individual Six Sigma projects.
The main elements and roles in this structure are:

• Careful project selection, as was already briefly mentioned, to ensure that

the best projects are selected and that the projects target critical stakeholders
and match the organization’s goals.
• Black Belts and Master Black Belts, people who have undergone careful
training in the Six Sigma methodology and are thereby qualified as project
managers. They typically spend all their time running Six Sigma projects,
but the position is not permanent. A selection of those certified as Black
Tools for Creating Improvements  207

Belts can be selected for the position of Master Black Belt, a highly quali‑
fied Six Sigma consultant. Master Black Belts teach Six Sigma inside the
organization and work as consultants on projects.
• Green Belts, employees who have received some training in Six Sigma and
work part-time on projects (typically 20–25 percent of their time). They can
even lead their own Green Belt projects, which are typically inside a single
• Other team members, who are employees without Black or Green Belt cer‑
tification. There are rarely enough Black or Green Belts to fill all the posi‑
tions on the Six Sigma project teams, so other people are enrolled, either
because they know the process well or because of personal skills.
• Project sponsors, senior or middle managers who act as Champions or
sponsors of individual projects. The sponsor, among other responsibili‑
ties, ensures that the team has the required resources, helps set the strategic
direction of the project, and participates in the project reviews.
• Six Sigma steering committee, the primary driver of Six Sigma deploy‑
ment in the organization. Larger companies can have more than one steering
committee. The committee responsibilities include project and team selec‑
tion and preparing the organization for Six Sigma.

There are other roles in the overall Six Sigma scheme, but these are the most
important. If you are contemplating initiating a Six Sigma program in your organi‑
zation, I suggest you view this brief Six Sigma description as an introduction and
consult more extensive sources before proceeding.
In terms of an example of Six Sigma, you have already seen that the overall
Six Sigma process is quite similar to the overall improvement model advocated by
this book. A full Six Sigma example would more or less mean covering most of the
tools presented thus far. Therefore, I have not included one here, but instead refer
you to the two more comprehensive cases presented in Chapters 13 and 14. Let us
therefore move on to the concept of business process reengineering.

11.6  Business Process Reengineering

Business process reengineering (BPR) blossomed during the 1980s and parts of
the 1990s. In all fairness, the approach has been heavily criticized—among other
reasons, for recommending reorganization as a quick fix for complex process prob‑
lems and for advocating downsizing and a ­cost-­reduction focus over seeing growth
opportunities. Long after its heyday and with most of the strong criticism behind
it, BPR has a core message that still seems relevant: business processes can be
tweaked and slightly adjusted, or they can be completely rebuilt. Although BPR,
208  Chapter Eleven

like many other approaches, does not represent a unified method, it contains prin‑
ciples that support both ways.
As have other popular tools, BPR has been defined in a number of different
ways. According to Hammer and Champy (1993, 32):

Reengineering is fundamental rethinking and radical redesign of busi‑

ness processes to achieve dramatic improvements in critical contemporary
measures of performance, such as cost, quality, service, and speed.

This definition focuses on the objectives of BPR, that is, the breakthrough
improvements. This element is also present in a definition by Peppard and Row‑
land (1995, 20):

BPR is an improvement philosophy. It aims to achieve step improvements

in performance by redesigning the processes through which an organi‑
zation operates, maximizing their ­ value-­added content and minimizing
everything else. This approach can be applied at an individual process
level or to the whole organization.

If you compare these two definitions with some of the other tools available,
many differences are apparent, but so are some similarities. What primarily
separates BPR from the tools presented so far is the strong emphasis on the
breakthroughs or radical improvements. BPR is not the tool to select if the
objective is to achieve a 10 percent improvement. Rather, it is a strategic tool
that can be useful when the organization wishes to achieve dramatic improve‑
ments in its performance level. At the same time, this seems very close to the
streamlining presented in section 11.1, that is, increasing the proportion of
­value-­adding activities in the processes. The fact is, BPR is a collection of
more or less related tools. Unlike streamlining, they have been put together
to form a structured process that, when coherently applied, can generate the
desired breakthrough improvement.
It is not the intention of this book to dwell on definitions of BPR. It should,
however, be mentioned that there are two principally different approaches when
it comes to how BPR is carried out. These two approaches are linked to the role
existing business processes play when applying BPR:

• Should the existing processes form the basis for the new, redesigned
• Should the existing processes be changed in the improvement phase, or
should new processes be designed to replace them?
• Should the organization simply start the reengineering with a clean sheet?
Tools for Creating Improvements  209

So far in this book, much emphasis has been put on understanding our existing
business processes. Ignoring these processes is very risky, as it means ignoring the
knowledge and experience that have accumulated over a long period of time. In
addition, there is a risk of repeating the same mistakes. It is also a fact that very
few organizations have succeeded in implementing completely new processes. On
the other hand, it avoids the danger of becoming buried in too many details of the
existing processes and therefore allowing the limitations incorporated in them to
impact the new process and thus limit improvement. As for most matters in this
area, there are no correct answers. Between extremes, there is an endless number
of degrees to which the existing processes influence the BPR effort. The important
issue is to strike a suitable balance between how things are done today and how
they should be done in a perfect world.
From the two extremes, two types of BPR can be defined:

• Systematic reengineering, where the current processes are understood,

documented, and analyzed in order to systematically create new and better
• Clean sheet reengineering, where the current processes are scrapped and
new processes are created from scratch through fundamental rethinking

Systematic reengineering can in many cases be synonymous with streamlining.

Because starting with a completely clean sheet can rarely be done in practice,
when referring to BPR throughout this book, something more moderate is intended.
As opposed to streamlining, BPR is not only about incremental improvements in
existing processes. BPR must seek to combine the best from existing processes
with new ideas about how the ideal process should look. This is quite close to the
technique of idealizing, whose objective is to create an image of the ideal process.
Thus, a true application of BPR will be a structured process where these two ele‑
ments, idealizing and streamlining, are combined.

11.6.1  A Process for Conducting BPR

There are many alternative processes that describe a BPR project. Kubeck (1995)
inspired the process presented in this book. As with most other improvement pro‑
cesses, this process can be divided into parts: planning, solutions generation, and
implementation. Since by definition the solutions generated in BPR can radically
deviate from the existing processes, the middle stage has been divided into two
phases: one for the actual generation of change proposals and one to evaluate how
these can be implemented. This gives a process spanning four phases, as depicted
in Figure 11.24.
Before each of these phases is described in more detail, a brief summary of the
process follows:
210  Chapter Eleven

Strategic External performance

plans requirements

1. Planning

BPR project

2. Reengineering


3. Transformation

Approved solutions

4. Implementation

Reengineered process

Figure 11.24  The process for conducting a BPR project.

1. Planning, where the focus of the BPR project is selected, a team to carry
out the project is formed, and, if possible, objectives for the project are
2. Reengineering, where, more or less based on the existing process, a set of
techniques is employed to reengineer the process to a level that will result
in significant improvements
3. Transformation, where it is determined how the reengineered process can
be implemented with regard to the existing process, the need for invest‑
ments, training, and so on
4. Implementation, where the solutions generated and approved during the
two preceding phases are implemented and the process changed

11.6.2  Planning
There are four central tasks in the planning phase:

• Select the process to be improved through BPR and consider its scope
• Assess the possibilities for achieving improvements and establish targets
Tools for Creating Improvements  211

• Establish a project team to perform the work in the project

• Produce a project plan for the BPR project

If this were a book dealing exclusively with BPR, the planning phase would be
described over several pages. However, the contents of this phase are highly related
to what has already been covered in this book, so most of the tasks have already
been discussed and performed in the overall improvement process. The first step,
selecting the process to be improved, is based on the results from the improvement
planning and the improvement road map.
The next step is an introductory assessment of whether BPR really is a suitable
tool for improving this process. It is necessary to consider which improvement
targets are realistic and desirable. This will aid in securing the necessary resources,
both time and money, that are allotted for the project.
The next task of the planning phase is to establish a project team. Most of the
tools presented in this book are meant for use in teams or groups of some form.
Some general guidelines for forming such groups were outlined in Chapter 6. The
last step in this phase is to produce a project plan. The planning phase of the BPR
project actually fits any preparation phase regardless of what tool is being used.
This description is not unique to BPR, but it can be used in any improvement proj‑
ect. The ­BPR-­specific content appears in the next phase, the reengineering phase.

11.6.3  Reengineering

The main purpose of the reengineering phase is, as the name implies, to reengineer
the process to be improved. Before and after this main task, there is preparation
and tidying up to do; thus, the steps of this phase are:

• Document the existing process

• Reengineer the process
• Develop recommendations for improvement

Whether choosing one or the other extreme (systematic reengineering or clean

sheet), there is a need to know how the process is currently carried out. The
first step of this phase is therefore to document the current process, usually by
using a flowchart. If the book’s approach to improvement work has been fol‑
lowed, this job has already been done. This simplifies the work in this phase
and makes it possible to start with the reengineering task. If much time has
passed since the process was documented or if those participating in the BPR
team are not familiar with this documentation, it might be necessary to update
or repeat it.
212  Chapter Eleven

Turning to the reengineering of the business process itself, the procedure

depends on the extent to which it is based on the existing process. Therefore, the
two procedures are described separately.

Systematic Reengineering
As has been mentioned, systematic reengineering resembles the streamlining
described in section 11.1. In BPR literature, the ESIA rules are often referred to.
They are rules and techniques designed to accomplish four core steps in the sys‑
tematic reengineering of an existing process:

• Elimination
• Simplification
• Integration
• Automation

Areas where each of these can be applied are listed in Table 11.3 (Peppard and
Rowland, 1995).
The ESIA rules should be applied in this order. The first step is to eliminate all
activities that do not add value (see section 11.1). In organizations where process
thinking is novel, a number of activities usually turn out to be ­non-­value-­adding
and thus eligible for elimination without any negative consequences. Toyota’s rule

Table 11.3  Focus areas for systematic reengineering.

Eliminate Simplify Integrate Automate
Excess Forms Jobs Dirty work
Waiting time Procedures Groups Difficult
Transport Communication Customers Dangerous
Processing Technology Suppliers Boring
Storage Problem areas Data capturing
Defects/Errors Flow Data transfer
Duplication Processes Data analysis
Tools for Creating Improvements  213

of thumb is that in many of the traditional manufacturing processes, 85 percent

of the employees will be performing nonproductive work at any point in time
(Griffiths, 1993, 20):

• 5 percent can be observed not working

• 25 percent are waiting for something
• 30 percent are manufacturing something that increases inventory levels but
does not add value
• 25 percent are working, but in accordance with less efficient standards or

It is obvious that dramatic improvements can be achieved by attacking the areas

listed under the “eliminate” heading in Table 11.3 and, of course, truly managing
to eliminate them.
When as many of these unnecessary tasks as possible have been eliminated,
the next step is to simplify as many of those remaining as possible. Areas that are
often overly complex are where the elements listed under “simplify” in Table 11.3
are generally involved.
Thereafter, the remaining tasks are to integrate and further improve the flow of
deliveries between the suppliers, the organization, and the customers. This can be
accomplished at several levels. First, several smaller jobs can often be integrated
into one single job. Such a job extension has several positive effects: it enriches
the content of the individual employee’s work while at the same time eliminating
interfacing between areas of responsibility. All such interfaces, no matter how well
managed, create waiting time and potential quality problems. Second, different
specialists can be integrated into groups that are responsible for more jobs. The
closeness and autonomy created result in many of the same positive effects as the
integration of several jobs into one. The third level is to attempt to integrate the
organization with customers and suppliers. Organizations that achieve a particu‑
larly close relationship to their suppliers and customers usually experience more
stable conditions and display better opportunities for common improvement. A
typical application is to integrate customers and suppliers of core components into
the product development processes.
The final approach is automation. Information technology, computeriza‑
tion, and robotic equipment are support aids that can be quite powerful and
give very good results if the processes being automated from the outset are
sound. Automating problem processes can, in many cases, make things worse;
that is, automation can enable errors to be made more rapidly. Therefore, it is
essential to apply the principle of automation last, after elimination, simplifi‑
cation, and integration have been successfully applied. Processes suitable for
automation are listed in Table 11.3.
214  Chapter Eleven

Automation usually fits the 80/20 rule. While many automation projects strive
for 100 percent functionality in the automated system, a general rule is that 80 per‑
cent functionality can be achieved at 20 percent of the cost. Making systems that
take into account every possible exception and special case can be very expensive.
If a system that can handle most cases, with occasional manual intervention, is
acceptable, both time and money can be saved.

Clean Sheet
It is difficult to give general advice about how to proceed with the clean sheet
approach, that is, starting with a clean sheet and drawing an entirely new process.
There are few, if any, specific techniques for this objective; we depend on creativ‑
ity, imagination, knowledge about available technology and human resources, and
so on. Each BPR project utilizing this approach is unique. However, there are some
basic questions that need answering:

• Which basic needs do we want to fulfill, and for whom? By answering this
question, we might realize that there are services or products that are pref‑
erable to those currently provided. A manufacturer of explosives asked this
question, and the answer was the need for a hole in a rock, not the explo‑
sives per se. The customers would prefer to have the hole supplied without
using the explosives. Therefore, the company started offering its services
as explosive experts in addition to marketing the explosives, delivering fin‑
ished holes at the points indicated by the customers.
• Why do we try to fulfill these needs? Does it fit in with the overall strategy
of the organization?
• Where must these needs be fulfilled? At home? In a separate service area? In
the company’s facilities? The answer to this question can form the basis for
dramatic changes in the processes. This happened in the restaurant industry,
which discovered that not all customers wanted to eat at the restaurant, but
would rather eat the restaurant meals at home. Thus, express delivery of hot
food to private homes was created.
• When must these needs be fulfilled? At what time of day and by which
deadlines? Even if this is not a pure BPR project, many service institutions
such as stores and banks have adapted their operating hours to meet the
customers’ needs. This entailed many changes in their processes to accom‑
modate afternoon or evening needs.
• How will the needs be fulfilled? What processes will be required? Who
must perform these processes, and which technologies does this require?

After answering these basic questions, we should start thinking about practical
issues regarding how to satisfy these needs. That way, these basic questions are not
Tools for Creating Improvements  215

clouded by practical limitations. Just from looking at these questions and the chal‑
lenge the clean sheet approach poses, it is obvious that the creativity of the people
involved is crucial. Following are some further questions to inspire the necessary
creativity and imagination:

• If you were to create a competitor of your own organization, how would it

look to have the highest possible impact?
• How should the ideal process look?
• If we could build the entire organization from scratch, how would the orga‑
nization and this specific process look?

These questions are summarized in Figure 11.25. Crazy ideas stemming from
brainstorming not limited by practical considerations are combined with the
answers to the preceding questions, ideas from benchmarking of other pro‑
cesses (described in section 11.7), available human resources, and technological
The last issue in the reengineering phase, no matter which approach is used, is
to summarize the changes needed to implement the new process. By comparing on
paper the newly designed process to the old process, we can decide which changes
are required. This list of recommended changes will form the basis for the work in
the transformation phase.

11.6.4  Transformation
The purpose of the transformation phase is to build the foundation for an effective
and successful implementation of the new process. The main tasks will be:

• Evaluating the changes required to implement the new process

• Planning the need for investments, training, purchases, and so on

Needs fulfillment: Customer

requirements, demand patterns, when,
where, practical limitations

Process design
crazy ideas

Benchmarking: Destroying sacred

cows, what is actually possible,

Figure 11.25  Process design using the clean sheet approach.

216  Chapter Eleven

• Creating a favorable climate for change

• Planning the implementation

The reengineering phase created a number of recommendations for change without

critically assessing their feasibility. The first step of the transformation phase is to
apply a more sober perspective on things so as to evaluate whether all activities
are possible to implement. The purpose is, of course, not to be so critical that most
of the recommendations are rejected, but rather to ensure that infeasible changes
or changes in opposition to the organization’s strategy are removed. For example,
a recommendation that a company change its operating hours from 9 a.m.–5 p.m.
to 2 p.m.–11 p.m. might be so radically against the human resource policy that it
should be rejected.
The next issue is to map the need for the changes and plan the necessary invest‑
ments, training, new staff, and so on. Many of the recommended changes will
require these kinds of efforts, and the needs must be known before planning the
implementation itself. If not, we risk, for instance, introducing the new process to
operators who have not been sufficiently trained, as this requirement was never
Before the actual implementation starts, it is very important that the right cli‑
mate for change be created. Even the most ­ well-­planned improvement projects
can be sabotaged if those who are impacted by the changes do not support them.
Creating a favorable climate is a science of its own, one that is not covered by
this book. It’s always favorable to have those who will later have to live with the
changes take part in planning them. This means that including on the project team
people already deeply involved in the process will create a good foundation for
a painless implementation. Another issue is to always be open and honest about
what is in the pipeline and provide a continuous flow of information to the affected
parties. A department facing extensive change is like an area of low pressure in the
atmosphere: it will suck up information of any sort until it is filled up. If this infor‑
mation is not constructive information from the management, it will be rumors
and gossip that can be harmful to the situation. Force field analysis, as treated in
section 12.3, is a suitable tool for creating the desired climate.
The last step before the implementation phase is planning. The need for this
plan is similar to the need for a plan for the BPR project itself, and it must contain
approximately the same elements, in addition to the required investments, training,
and so on.

11.6.5  Implementation
Like the planning phase, the implementation phase of the BPR project, where
changes are effected, is quite general. Regardless of which improvement project is
discussed, the last activity is always implementation of improvements. Although
Tools for Creating Improvements  217

the changes resulting from a BPR project are often more comprehensive than when
other tools are used, this phase in the BPR project is not in any way unique. At the
same time, there is no doubt that implementation can be the most difficult phase.
Many improvement projects that have performed excellently to this point fail in
The implementation phase will not be described in detail here, but it will
be described more generally in Chapter 12. The steps of this phase are as

• Set targets for the improvements

• Create a favorable climate for change
• Carry out the implementation plan
• Monitor the progress of the implementation and handle any deviations

Many talk about the two tools of BPR and benchmarking simultaneously, and they
are to some extent related. This is treated in the next section.


As an example of the use of BPR, we will look at a manufacturer of hydrau-­

lic hoisting equipment for industrial applications. This company produced
some standard systems, accounting for about 60 percent of its sales. At
the same time, the company offered ­tailor-­made solutions for all potential
purposes. Its reputation in the market was that it could do almost anything.
Thus, for almost half of the sales, some degree of redesign of standard
systems or development of new products and engineering of new solutions
was required.
The company was one of the few in the market offering ­tailor-­made sys-­
tems, but lately, several competitors had moved into this segment. Some
competitors could offer shorter delivery times, even when the delivery
involved product development, which endangered the competitiveness of
our company. It was therefore decided to undertake a BPR project with the
objective of reducing the delivery times for such systems by 50 percent.

The process to be improved was identified, and a project team was
established that consisted of two people from each of the following

• Sales
• Manufacturing
218  Chapter Eleven

• Product development/projects (consisting of hydraulic and mechan-­

ical design)
• Procurement

The team decided to make a project plan and expected a ­six-­month dura-­
tion. The clean sheet approach was selected with consideration to the cur-­
rent situation. Thus, the intention was to combine the best from the existing
process with elements of the ideal process.

Following completion of these introductory tasks, the next step was to
document the current process, as shown in Figure 11.26.
A common denominator for the entire process was sequential develop-­
ment. First, the hydraulic department designed the core system functional-­
ity. Next, the mechanical department designed the mechanical system to
fit the hydraulic system. Then the manufacturing department had to find a
way to manufacture this system, while the procurement department was
left to quarrel with the suppliers about components and processes for
these. Often, the final product was different from what the customer had
specified. On the positive side, the product development department used
a very rational process to make drawings, system charts, ­bill-­of-materials,
and so on.
The core task of the BPR project, which was viewed with great expecta-­
tions, was to describe the ideal process. This task was carefully planned by:

• Involving people known for their creativity and enthusiasm at work

• Reserving two days per week for four weeks exclusively for this
• Creating in advance a climate of trust and cooperation within the
group through two social gatherings
• Ensuring the availability of all necessary aids for the job

Without describing the group’s work in detail, the result, shown in Figure
11.27, was an outline of an ideal process that all agreed represented a dra-­
matic improvement over the current situation.
In this process, parallelism was emphasized using ideas from concur-­
rent engineering and cooperation between internal departments and exter-­
nal elements such as customers and suppliers. The project group found
this ideal process to be so powerful that it wanted to implement as much
of it as possible. The reengineering phase was therefore concluded by sys-­
tematically summarizing all changes required to reach this target.

As the transformation phase was started by employing a more sober atti-­
tude, talks with the customers and suppliers revealed that while most of
The project department
The sales department assesses the need for Modification
Customer request Type of project? project
specifies the system redesign or product

Product development

Hydraulic modification Mechanical Generation of process

by the hydraulic modification by the plans by the production Production planning
department mechanical department preparation department

Negotiations with
Delivery and possibly
suppliers and Manufacturing Delivered system

Development of Mechanical design Generation of process

Hydraulic design by
system functionality by by the mechanical plans by the production Production planning
the hydraulic department
the project department department preparation department

Negotiations with
suppliers, possible
Delivery and possibly
development of new Manufacturing Delivered system
components, and

Figure 11.26  The current process for developing a tailor-­made system.

Tools for Creating Improvements  219
Establishment of an
Preparatory integrated team of the
discussions between customer and one person Modification
Customer request from the project,
Type of project? project
the customer and the
sales department manufacturing, and
220  Chapter Eleven

procurement departments

Product development

Parallel development of Joint production

hydraulic and mechanical planning between the
system and process plans manufacturing and
by the integrated team procurement departments

Joint production planning Agreement on the supply

Extension of the Parallel development of
between the manufacturing of components between
integrated team with a hydraulic and mechanical
and procurement the procurement
representative from the system and process plans
departments and the department and the
involved suppliers by the integrated team
suppliers suppliers

Delivery and possible

Manufacturing Delivered system

Figure 11.27  The ideal process.

Tools for Creating Improvements  221

them were interested in cooperation, three central suppliers did not want
this solution. Thus, the involvement of suppliers in the integrated team was
scrapped. No other obstacles were found against organizing such proj-­
ects using integrated teams. No investments were required, but training
was necessary in the areas of cooperation, group processes, and effective
Because central people from the affected departments had partici-­
pated in outlining the new process, the climate for change was favorable
from the outset. To reinforce this climate, two ­one-­day information semi-­
nars were planned where the project group, in cooperation with external
resource people within concurrent engineering, would present the project
and its advantages. An implementation plan was also produced during this

Implementing the new process took nine months longer than expected.
The results, however, were well worth the effort. The average time from
when the first customer request was received to when a complete system
had been developed, manufactured, and delivered was reduced from 4.5
months to less than two months. In addition, the climate in and between
departments was significantly improved.

11.7  Benchmarking

Benchmarking, which has been mentioned numerous times in the book, is a versa‑
tile method whose core element is comparison. It can be used for many different
purposes: comparing your performance with that of competitors to identify per‑
formance gaps, comparing regional service offices’ customer satisfaction perfor‑
mance to determine bonuses, or comparing universities to present rankings of the
best schools in different areas. Benchmarking and the key principle of comparison
can be utilized at many different stages in the improvement work. However, in
this chapter, I include it as an improvement tool. Before the actual process for
conducting benchmarking is introduced, some definitions and other introductory
issues related to benchmarking will be presented. The bulk of the description of
benchmarking is based on Andersen and Pettersen (1996).

11.7.1  Definitions of Benchmarking

The original meaning of the noun benchmarking is a predefined position, used
as a reference point for taking measures against. This word has migrated into the
business world, where it has come to mean a measured ­best-­in-class achievement
recognized as the standard of excellence for that business process.
222  Chapter Eleven

As a verb, benchmarking can be somewhat philosophically defined as


Benchmarking is the practice of being humble enough to admit that some‑

one else is better at something, and being wise enough to learn how to
match them and even surpass them at it. (APQC, 1993)

This definition captures the essence of benchmarking, namely, learning from oth‑
ers. The term benchmark probably comes from geographic surveying, before the
GPS era, where points in the terrain were given with reference to a fixed point,
often a tall peak or some other easily recognizable and definable point. At that
point, two stone slates were raised on end and spaced somewhat apart. A third slate
was laid on top of these, thus forming a stone bench—hence the word benchmark.
Because of its meaning of comparison against a reference point or the optimal, it
was put to use in the business world, was extended, and can be defined as shown
in Figure 11.28.
Benchmarking, like BPR, emerged as a globally known tool during the 1980s.
Like many other new tools, benchmarking was in fashion for a few years and was
used and abused for all kinds of purposes. Fortunately, the tendency to view bench‑
marking as the answer to any issue has passed, and like BPR, benchmarking is
now a tool just like many others, with its strengths and weaknesses. There has also
been an evolution over the years in what we view as the strengths of benchmark‑
ing. Many used to see benchmarking as a method for comparing only key figures,
often financial key figures, for the purpose of ranking the organization in relation to
competitors or the industry average. This might have been the main application of
benchmarking in the past, but today it is a far more powerful tool that is much more
widely applicable. The core of the current interpretation of benchmarking is:

Comparing business
A structured
processes, not just Learn from others
performance measures

“Benchmarking is the process of continuously measuring and

comparing one’s business processes against comparable
processes in leading organizations to obtain information that will
help the organization identify and implement improvements.”

External Change, not

focus evaluation

Figure 11.28  Operational definition of benchmarking.

Tools for Creating Improvements  223

• Measurement of your own and the benchmarking partners’ performance

level, both for comparison and for registering improvements
• Comparison of performance levels, processes, practices, and so on
• Learning from the benchmarking partners to introduce improvements in
your own organization
• Improvement, which is the ultimate objective of any benchmarking study

There are four main reasons for advocating the use of benchmarking in improve‑
ment work:

1. Benchmarking helps the organization understand and develop a critical

attitude toward its business processes
2. Benchmarking encourages an open attitude toward seeking and sharing
information and thereby is an active learning process that motivates change
and improvement in the organization
3. Through benchmarking, the organization can find new sources of improve‑
ment and new ways of doing things outside its own environment
4. Through benchmarking, reference points are established for performance
measurement of business processes

Thus, benchmarking is about comparing your organization with other organiza‑

tions. Different types of benchmarking can be defined depending on who you
compare against and what you compare. First, you can compare on three different
levels (at least) in terms of who you use as benchmarking partners:

• Internal benchmarking—comparison against the best within the same orga‑

nization or corporation
• Competitor benchmarking—comparison against the best direct competitors
• Generic benchmarking—comparison against the best, regardless of indus‑
try or market

The farther down the list we go in these types of benchmarking, the farther away
from our own little world we move. This is illustrated in Figure 11.29. Because
gaining access to competitors for benchmarking can be difficult, turning to other
industries might be wise. At the same time, the chances are often higher that an
organization in a different sector has developed approaches and business processes
that are different from yours.
224  Chapter Eleven

Other industries


Figure 11.29  Benchmarking out of the box.

On the other hand, identifying relevant benchmarking partners from other

industries often requires significant creative leaps. The key is to think of someone
with a similar challenge who is in a different setting. When successfully made,
such novel links can cause breakthroughs for a whole industry. Some examples are
shown in Table 11.4.
Depending on what you compare, there are again at least three levels of bench‑
marking (the differences are illustrated in Figure 11.30):

• Performance benchmarking—comparison of pure key figures or other perfor‑

mance measures. In sports, a high jumper doing performance benchmarking
could learn where to set his targets from what the best in his age group achieved
nationally, but he would learn nothing about how to make that height.
• Process benchmarking—goes beyond performance measures by comparing
how business processes are performed as well as how well they are per‑
formed. For the high jumper, this could answer questions like which jumping
technique to apply, which equipment to use, which training methods to fol‑
low, what dietary issues to be aware of, and so on, to clear this height.

Table 11.4  Breakthroughs through benchmarking.

Problem Compared with
Long admittance times in hospitals Hotel reception areas
Too lengthy setup of machines Formula 1 pit crews
Planning the delivery of fresh Pizza delivery
Unstructured maintenance of power Maintenance of aircraft engines
Difficult to manufacture shell cases Manufacturing of lipstick tubes
with the right cylindrical shape and
smooth surface
Tools for Creating Improvements  225


Figure 11.30  Three types of benchmarking based on what is being compared.

• Strategic benchmarking—comparison of strategic decisions and disposi‑

tions at a higher level. This is a less frequently used variant of benchmark‑
ing. In the sports example, this could perhaps teach the high jumper which
competitions to enter to promote his career most effectively.

Each type from these two categories—that is, who and what are being bench‑
marked—can, in theory, be combined into a benchmarking study with a given focus.
In practice, not all combinations are equally suitable, as is shown in Figure 11.31.
This is not explained in detail, but different studies have shown that the best results
are generally achieved by a combination of process and generic benchmarking,
using partners from other industries.
The different types of benchmarking can be used in combination to some extent. A
typical benchmarking study is outlined in Figure 11.32. As a start, performance bench‑
marking can be used to give an indication of where the organization stands and what

Internal Competitor Generic

benchmarking benchmarking benchmarking




Relevance/value: Low Medium High

Figure 11.31  Recommended combinations of types of benchmarking.

226  Chapter Eleven

High-level comparison:
Performance • Where are we?
benchmarking • How good are we in central areas?
• Who is good?
• Who can we learn something from?

partners List of

Detailed comparison
of processes:
• What do the best do?
• What can we learn from them?

Figure 11.32  Progress of a typical benchmarking study.

should be improved, as well as to find possible benchmarking partners. Next, process

benchmarking is used to improve through observing what the best do.
The last definition issue relates to how benchmarking can be conducted with
respect to the relationship to the benchmarking partners. The most common mod‑
els are shown in Figure 11.33. The most used model is where one organization
initiates a benchmarking study with several other organizations as benchmarking
partners. A ­one-­on-one comparison is made against each of these, often without
letting the other partners know that more than one organization is being used and
who the others are. It has become common practice for several organizations to join
together to conduct a benchmarking study where comparisons are made among all
the organizations involved. This usually gives better value for all partners involved.
The third model, which is perhaps somewhat ­old-­fashioned, is where a third party,


in a group

comparison through
a third party

Figure 11.33  Models for organizing a benchmarking study.

Tools for Creating Improvements  227

usually a consultant, performs the data collection and analysis and then presents
the results to the organizations that contributed their data. As each organization
cannot observe the best practices in use, it is often difficult to make any changes as
a result of such studies. The model is best suited for performance benchmarking.

11.7.2  Conducting a Benchmarking Study

Benchmarking is usually conducted in separate projects, normally called bench‑
marking studies, whose individual objective is to improve one of the organization’s
business processes. In the larger setting of business process improvement, a bench‑
marking study typically forms one activity in the improvement stage. A full bench‑
marking study includes the activities necessary for the organization to:

• Study and understand its own process

• Find benchmarking partners
• Study the benchmarking partners’ processes
• Analyze the differences between its own and the benchmarking partners’
• Implement improvements based on what was learned from the benchmark‑
ing partners

You will recognize some of these tasks as belonging to other phases of the improve‑
ment process. Typically, the first phase of studying your own process will have been
conducted in earlier stages of the improvement project. Improvement implementa‑
tion also naturally falls within the next stage of the larger improvement process.
There are a number of models that describe the different steps that constitute a
benchmarking study; they differ in what to include and what to leave to other tools
to handle. One such model is the benchmarking wheel, portrayed in Figure 11.34.
Ordinarily, a benchmarking study lasts between six weeks and eight months,
depending on the extent of the process being benchmarked, the number of bench‑
marking partners used, how data are collected from them, and so on. Often, imple‑
menting the improvements themselves, the main content of the adaptation phase,
can take longer. For the other phases, approximately 50 percent of the time is spent
on planning, 20 percent on observing the benchmarking partners, and 30 percent
on analyzing collected data. A brief description of each of the phases follows.

The planning phase contains four basic activities:

1. Select the process to be benchmarked

2. Establish a benchmarking team
228  Chapter Eleven

5 1
Adapt: Choose “best Plan: Critical success
practice,” adapt to factors: select a process for
the company’s conditions, benchmarking, document the
and implement changes process, and develop
performance measures

Analyze: Identify gaps in 2
performance and find Search: Find bench-
the root causes for the marking partners
performance gaps

Observe: Understand
and document the
partners’ process, both
performance and practice

Figure 11.34  The benchmarking process based on the benchmarking wheel.

3. Understand and document the process to be benchmarked

4. Establish performance measures for the process

These steps apply only if benchmarking is used as an isolated tool. Because this
book encourages a more coherent process where improvement needs are identified
at an early stage, we can proceed directly to the next step of the planning phase.
Thus, no description is given on how to select a process for benchmarking. The
­benchmarking-­specific content appears in the search phase.

From the benchmarking wheel, it might seem that the different phases should be
performed sequentially, but this is not the case. If the entire planning phase is com‑
pleted before the search for benchmarking partners is started, the progress of the
study will usually display a significant drop, as searching for partners usually takes
some time. It is recommended that the task of searching for benchmarking partners
be started in parallel with the planning, often quite early in the planning phase. The
contents of the search phase are as follows:
Tools for Creating Improvements  229

1. Compile a list of criteria that an ideal benchmarking partner should

2. Search for potential benchmarking partners
3. Compare the candidates and select one or more partners
4. Establish contact with the selected partners and gain acceptance for their
participation in the study

The first of these steps is actually a ­search-­technical task. No matter what is being
searched for (for example, books in a library or partners for benchmarking), it
is preferable to limit the search parameters. Starting a search among all possible
organizations and considering them potential partners can result in an extremely
complicated search. It is wise to consider the qualities the ideal benchmarking
partner should possess, and consciously search for these. Typical issues included
in such criteria are:

• Geographic location
• Size
• Technology used and markets served
• Industry
• Structure and organization

This list could be endless. However, it is important not to limit the search too
much, either, as we might exclude partners that could be suitable, if not perfect,
for the study. To find potential partners, a number of sources can be utilized. The
organization’s own network can usually provide useful information pertaining to
customers, suppliers, and others with whom there is cooperation. Different experts
within the field can often be of help. Various industry associations possess knowl‑
edge about players in the field. Media coverage can give hints, as can all public
information. By using carefully selected search keywords, vast amounts of infor‑
mation can be found on the Internet. In fact, only the imagination sets boundaries
for the search, although this has its pros and cons. Since the information available
is so vast, it is easy to become paralyzed, not knowing quite where to start. Many
benchmarking study participants conclude that this phase is indeed difficult and
should not be underestimated.
It is also worth considering what performance level we want the bench‑
marking partners to be at. The pyramid in Figure 11.35 depicts the number of
potential benchmarking partners available within each performance level. For
world class, there may be at most only one candidate for each business process,
230  Chapter Eleven

while the number increases rapidly with decreasing ambition level. Thus, it can
be much easier to find a moderately good benchmarking partner than holding
out for someone of ­world-­class ability. A stepwise climb is more reasonable, as
it might be close to impossible to go directly to the business processes found
among the very best. Note that there might be many organizations using radi‑
cally different processes but not achieving any better performance levels. Do
not mistake these different processes for better processes and fall into the trap
of adopting them.
After a number of potential benchmarking partners have been identified,
compare and assess them before making a final selection. Often during this
phase, a form of performance benchmarking will be carried out to determine
how good the different contenders are. Generally, more than one partner is rec‑
ommended, the usual being between three and five. The costs increase with each
additional partner, but they are outweighed by the advantages and the potential
for finding best practices.
The last activity in the search phase is to establish contact with the desired
partners for the benchmarking study. There is no sure way to design such a query
or to present it to ensure acceptance from the potential partners. An important ele‑
ment, however, is that you are willing to enter into a ­give-­and-take agreement. A
willingness to open your organization and share the benchmarking information
increases the likelihood of a positive response.

Best in
world class

Best practice

Improvement compared
to current performance

Equal to or lower than
current performance

Figure 11.35  Ambition pyramid for benchmarking partners.

Tools for Creating Improvements  231

In the observation phase, the benchmarking partners’ processes are studied and
documented, just as was done for our own process in the planning phase. The
observation phase covers three steps:

1. Assess the information needs and information sources

2. Select a method and tool for collecting data and information
3. Perform data collection and debriefing

In benchmarking, we often say that information is sought at three levels, as shown in

Figure 11.36. It is important to know the differences between these levels. The perfor‑
mance level indicates how well the organization performs a business process. To learn
something and become better, it is necessary to go beyond the performance level and
look at how the process is performed: the practices. These practices are specific process
elements that can be implemented in our own situation to achieve improvements.
There is no guarantee that a process will work in your organization. Therefore,
enablers must be captured when studying the process of the benchmarking part‑
ners. Enablers are conditions surrounding the process that render the partner able
to perform the process in a particular way. These can be IT systems used, contracts
with suppliers, competence levels of employees, and so on. If these conditions are
not captured, time and money will be spent attempting to implement a new process
that cannot work due to lacking enablers.
A number of methods are available for collecting information and data about
benchmarking partners. We separate between methods (the means utilized to get in
touch with the partner) and tools (the specific techniques used to collect data). The
matrix of Figure 11.37 summarizes some of the most common tools and methods,
and the Xs indicate relevant combinations. Each of these methods and tools is actu‑
ally a field of its own.
The last step of the observation phase is debriefing, which happens after data have
been collected from a benchmarking partner. After being away from the office for a


Practice Performance


Figure 11.36  Three levels of benchmarking information.

232  Chapter Eleven

Tools Direct
Methods Questionnaire Interview observation
Telephone X X
Partner visit X X X

Figure 11.37  Methods and tools for the observation phase.

while because of this task, you will often have numerous urgent matters to deal with
upon your return. If these matters are attended to before you process the collected infor‑
mation, many impressions and ideas may be lost. It is therefore important to set aside
time immediately after data collection for a debriefing within the team in order to:

• Collect and sort information and data

• Transcribe interview notes
• Take down mental impressions and other ideas not yet captured on paper

After the information and data about the benchmarking partners’ processes have
been collected, the next step is to analyze the data and identify improvement sug‑
gestions. The analysis phase consists of five steps:

1. Sort the collected information and data

2. Quality control the collected information and data
3. Normalize the data
4. Identify gaps in performance levels
5. Identify causes of the gaps

The first two are really preparatory tasks that are included to ensure that all the data
are available and accurate. Before starting the actual analysis of the data, it can
be helpful to normalize the data set. A common argument against benchmarking
is that what is being compared is unique, so no comparison is even close to being
relevant. Most things can, however, be compared if a normalization is carried out
beforehand. Normalizing means adjusting for the conditions that are truly different.
It often consists of recalculating the data to average values or ratios that eliminate
aspects like size, market conditions, different legislation, cost levels, and so on.
One central element of the analysis phase is a gap analysis. This entails iden‑
tifying gaps as well as determining the causes of these gaps. The gap itself is of
little help when it comes to identifying improvement actions. The purpose of the
Tools for Creating Improvements  233

gap is to prove that one partner does something that makes him or her able to per‑
form better than another. The presence of a gap is thus more of a signal that there
is something worth examining more closely.
There are several techniques available for finding the causes of the gap. A rela‑
tively simple approach is to compare the flowcharts for the processes directly. This can
provide much information about the differences that contribute to higher performance.
Other tools that have already been described but that are suitable for this purpose are
­cause-­and-effect charts, relations diagrams, and five whys analysis. Irrespective of the
approach employed, the goal of this phase is a list of the conditions believed to con‑
tribute to the superior performance level of the benchmarking partners.

As with the planning phase, the adaptation or implementation phase is not really
benchmarking specific. Implementation in general is treated in Chapter 12; thus, the
description here will be very brief. The adaptation phase consists of four steps:

1. Describe the ideal process and summarize improvement actions based on it

2. Set targets for the improvements
3. Develop an implementation plan, carry out the plan, and monitor the
4. Write a final report from the benchmarking study

After the benchmarking study has been completed, a final report should be produced
that covers the entire study and contains lessons learned and recommendations for
future studies. This report will constitute documentation of the study to all parties
involved and interested. Furthermore, it should describe improvement actions that this
project did not implement. The document should be sent to the benchmarking partners
in return for their participation. Finally, it should contain some record of all bench‑
marking partners involved and that could become involved in ensuing studies.


As an example of a successful application of benchmarking, a description

of a study performed by the telecommunications company Pacific Bell is
included. Pacific Bell is one of several companies offering phone services
to the residents of California. It was separated from the AT&T monopoly
when the market was deregulated. One means of survival in an increas-­
ingly competitive environment has been continuous measurement of cus-­
tomer satisfaction. After using this method since the early 1980s, Pacific
Bell started a benchmarking study to look at and improve the system. The
234  Chapter Eleven

basis for the study was a general concern about whether the measure-­
ments were accurate and whether they were used sensibly. The bench-­
marking study is described here, phase by phase.

Planning of the Study

First, a benchmarking team was established, consisting of:

• The manager for the measurement department

• Four employees from the measurement department, responsible for
the areas of collection, analysis, and application of customer satis-­
faction data
• The central benchmarking manager of the company, who acted as a
facilitator of the benchmarking methodology

The first task of the team was to document the current process for measuring
customer satisfaction. A very thorough job was done in this respect, result-­
ing in a 17-page description of all processes related to this area. Besides this
task, the team members participated in different conferences, seminars, and
other events that focused on the topic of customer satisfaction measurement.
The result was that the team was knowledgeable about its own processes as
well as the field in general by the conclusion of the planning phase.

Search for Benchmarking Partners

To find the best possible benchmarking partners, Pacific Bell defined a list of
criteria for the organization and its customer service. For each element on this
list, a value was assigned indicating its importance for a potential partner. In
order of decreasing importance, the criteria related to the company itself were:

• Profitability the last five years, as this was seen as an expression of

the degree of customer satisfaction
• Multiple market segments, to find partners operating in a similar
situation to Pacific Bell’s
• Service industry, as customer satisfaction measurement in the ser-­
vice industries was considered quite different from what was being
done in manufacturing organizations
• Long-term customer relationships, as opposed to companies depen-­
dent on ­one-­time sales, as ­long-­term relationships would enable a
continuous measurement of customer satisfaction
• Based in California, as this geographical area has a unique blend of
• Technology driven, to resemble Pacific Bell as much as possible
• Changed regulation conditions, to see how such conditions impacted
customer satisfaction and the measurement of it
Tools for Creating Improvements  235

The criteria defined regarding the customer service of potential partners


• A leader in customer satisfaction

• Active use of feedback from customers for process improvement
• A quantitative and systematic system for measuring customer
• Using several different instruments for customer satisfaction data

To identify potential benchmarking partners, one ­full-­time person was hired

to conduct a very thorough search of literature to find companies match-­
ing the criteria. This resulted in a list of about 20 companies that seemed to
match the profile. These 20 potential partners were divided among the team
members to conduct an even more detailed assessment of each. From these
evaluations, eight organizations were agreed on as objects for further study.

Information Collection
The selected partners belonged to industries such as banking, telecommu-­
nications, insurance, and public relations. To collect information, a ques-­
tionnaire was produced that consisted of two parts, one quantitative and
one qualitative. The quantitative part focused on figures for the number of
employees who performed different tasks, costs, the number of customers
followed up, response rates for measurements, and so on. The qualitative
part was far more extensive and focused on how customer satisfaction
measurement was performed, how it was applied, who measured it, what
specific performance measures were used, and so on.
Due to limited means for the benchmarking study, the team found it
impossible to undertake the data collection through partner visits. Training
was therefore given in interview techniques, and the questionnaire was
answered by phone and fax. First, the companies agreeing to participate in
the study received a document containing Pacific Bell’s own answers to all
the questions in the questionnaire. Next, the quantitative part was filled in
by the partner and faxed back to Pacific Bell. After examining the answers,
the team handled the qualitative part over the phone.

Data Analysis
To conduct the analysis in a sensible way—that is, by spending sufficient time
on it—a number of ­full-­day meetings were agreed to. First, the quantitative data
were compiled in a matrix and analyzed, which was a rather easy task. The
analysis of the qualitative answers was far more complex, because all answers
had to be discussed in plenary to create a full understanding of the partners’
processes. This phase led to a list of recommendations for Pacific Bell:

• Establish an organization to handle all customer responses and use

it to improve the products and services
236  Chapter Eleven

• Terminate the generation of customer satisfaction data on the level

below managers, to avoid the employees’ fear of repercussions
• Stop basing the payment for ­lower-­level managers on customer sat-­
isfaction data
• Develop internal process indicators linked to customer requirements
• Expand the scope of customer satisfaction measurement, but
reduce the measurement frequency to once every three months
• Survey both customers who had recently been surveyed and those
who had not, while also trying to reach the customers who rarely
give any feedback
• Use customer satisfaction data at a strategic level
• Eliminate the frustration of employees at lower levels in the organi-­
zation who are held responsible for measures they can only partially

Adaptation and Improvement

Within a month after having presented the recommendations, management
agreed to reduce the extent of the questionnaire used for measurement.
This led to savings of about $1 million, and thus immediately saved far
more than the total costs for the benchmarking study.
More importantly, actions were taken to understand in detail the custom-­
ers’ true requirements for the services delivered by Pacific Bell, as exemplified
by the benchmarking partners. Furthermore, pilot implementations were under-­
taken of a totally new system for customer satisfaction measurement, where
the responses were used only at a high organizational level, and the other
employees had process level measures developed. A number of less exten-­
sive changes were introduced regarding measurement of customer satisfac-­
tion, related to specific processes for customer service. For instance, a system
was established where immediately after the installation or repair is completed,
the customer is notified—a practice that was not common before.

Achieved Results
The costs for the study reached $70,000. The reduction in measurement
extent and the introduction of a new measurement system for customer
satisfaction are expected to give savings in the area of $5 million. Even
more important than the savings is the fact that a number of processes
were dramatically improved as a result of the study, which is expected to
result in large gains in terms of increased customer satisfaction. In addition,
many of the changes improved the work situations of many employees.

Having concluded this phase of the improvement project, you should be ready
for the final leg: implementing the improvements.
Tools for Implementing

ll the previous categories of tools have to some extent been aimed at creat‑
ing solutions or improvement actions to solve different problems. If these
improvements are not implemented in the organization, the entire effort
is wasted. Excellent improvement suggestions do no good lying in a drawer or on
a shelf; they must be brought to life. This very difficult task consists of several

• Sorting and prioritizing among the improvement proposals

• Organizing the implementation
• Setting targets for the improvements
• Developing an implementation plan
• Creating acceptance for the required changes and a favorable climate for the
• Carrying out the implementation itself

This chapter partly addresses some general questions related to these tasks and
partly describes some specific tools. The tools presented are:

• A∆T analysis
• Tree diagram and process decision program chart
• Force field analysis

The remainder of this chapter contains a blend of descriptions of the more general
tasks in the implementation phase and the more common tools. The first decision to
be made in the implementation phase of an improvement project is usually what to

238  Chapter Twelve

implement. Often, the work of the preceding phases will have generated a number
of possible changes. Rarely will the resources required to implement all suggestions
be available. It is therefore a natural first step to prioritize the improvement propos‑
als and start with those expected to produce the greatest effects or that for some
other reason are preferred. Criteria that can be used for sorting the proposals are:

• The investment needs for introducing a new method or process

• The training needs for a new process
• Time limitations, both in the form of deadlines for the project and organi‑
zational restrictions with regard to the time available for performing the
implementation task
• The organization’s motivation level (that is, burned out or still enthusiastic)

A little more systematically, a ­four-­field matrix can be designed, using the follow‑
ing issues as determining factors:

• Expected improvement effects

• The “implementability” of the improvements, that is, how easy or difficult
they will be to introduce in the organization

This will give a matrix like the one shown in Figure 12.1. Improvements that fall
within the ­low-­hanging-fruit area are obviously the preferred ones to start with.
Such a sorting will give a prioritized list of improvement actions in the order
they should be implemented. There are several options for deciding how the imple‑
mentation should be organized:

• By the original improvement team, where the same team that has carried
out the project so far also undertakes implementing the improvements. The
advantage of this approach is that the team knows the project and what the
solutions entail.

Expected ease of implementation

High Low
Expected improvement

High effect, difficult

High to implement;
second choice

Low effect, easy

Low to implement; Wasted effort
second choice

Figure 12.1  Matrix for assessing implementation order.

Tools for Implementing Improvements  239

• By a specific implementation team, where a new team is formed that con‑

sists of the necessary and suitable people to assume responsibility for the
implementation itself. Even if this team does not know the work so far, it is
often wise to include people in the implementation who were not involved
in the development of improvement proposals. Thus, this approach can be
sensible in some cases.
• In the line organization, where the functionally responsible people assume
responsibility for implementing changes by using the resources of the ordinary
organization. This is perhaps the most common model, where we ensure that
those who will later perform the process also take part in implementing it.

After the model for organizing the implementation is selected, the next logical step is to
set targets for the improvements: which performance level we want to achieve after the
implementation has been completed. For this purpose, A∆T analysis is a helpful tool.

12.1  A∆T Analysis

A∆T is a tool that is quite closely related to both idealizing and ­value-­added analy‑
sis. Whereas both of these tools are used to create improvement proposals, the main
purpose of A∆T analysis is to set ambitious targets for the improvement work. The
method is based on the assumption that it is always possible to find two durations,
accumulated costs, total numbers of defects, or other accumulated performance
measures for a given process:

• “A” stands for actual: the actual time, costs, and so forth, currently related
to performing the process to be improved
• “T” stands for theoretical: the theoretically fastest time, lowest cost, and so
forth, that can be achieved when performing the process

With A∆T defined this way, we see that the theoretical value is closely related to the
ideal process in idealizing. Furthermore, if we are considering only time or cost, the
theoretical value can often be found simply by subtracting OVA and NVA, identified
through a ­value-­added analysis. The analysis can be used in conjunction with these
two tools. When performing an A∆T analysis, these two values can be used in two
different ways. First of all, we can calculate the ratio between the A and T values:

∆= .
This ratio, ∆, expresses the improvement potential in eliminating all unnecessary
activities and performing the process as efficiently as possible. The higher the ratio,
the higher the potential. This ratio can also be used as an expression of how much
there is to gain by approaching the ideal process.
240  Chapter Twelve

For setting improvement targets, there is no need to calculate ∆. Instead, the target
can be based on the T value. Whether the target should be set equal to the T value or
somewhat less ambitious to take into account any practical limitations is an assessment
that must be done in each specific situation. The steps of the analysis are as follows:

1. Start the analysis from the flowchart for the current process.
2. In the flowchart, add figures for time, cost, number of defects, and so on,
for each activity.
3. Critically evaluate each activity to determine whether it adds value. If it
does not, determine whether it can be eliminated. Activities that can be
eliminated are marked with a color or by some other suitable means.
4. Summarize the A values and the T values, where the T values are all the
nonmarked activities, and calculate the ratio A/T.
5. Set the improvement target at or close to the T value.


A company that had reached the implementation phase of an improvement

project had some problems in determining an improvement target for a
reduced process duration. A suitable tool for this task was the A∆T analysis, as
the company had produced a flowchart for the process and partly performed
a ­ value-­added analysis. The flowchart with attached activity duration times
and activities eligible for elimination marked is shown in Figure 12.2. As can be
seen, the ratio between A and T is 2.05, that is, a potential for approximately
halving the duration time of the process. To be a little more realistic, the target
was reduced somewhat compared with the ideal, down to 12 days.

3.5 2.0 1.0 3.5
G 4.5

= can be eliminated
A 21.5 0.5
= = 2.05 K H 2.0
T 10.5

Figure 12.2  Flowchart for the A∆T analysis.

Tools for Implementing Improvements  241

When it comes to improvement targets, or targets in general, they should be:

• Ambitious enough to require some effort to be reached. Targets that are too
easily attainable do not pose any challenges and rarely inspire great motiva‑
tion. The result can therefore be a lower improvement gradient than what
could have been the case had the targets been more ambitious.
• Realistic, so as not to deter. Targets that are too ambitious can cause the pen‑
dulum to swing too far to the other side, leading to frustration and reduced
• Operative, so they are easy to comprehend and follow up, to monitor whether
we are approaching the target.


It has been claimed that when Ronald Reagan won his first presidential
election, he gave all federal offices instructions to increase productivity by
5 percent within one year. After a year had passed, the results started com-­
ing in to the administration office, and most of them were between 4.8 and
5.2 percent productivity increases. The target of 5 percent was probably
not sufficiently ambitious, and a target of 10 percent would probably have
resulted in figures between 9.8 and 10.2 percent. Because the target was
set too low, it did not inspire and motivate sufficiently to reap the full effects
possible. If, on the other hand, the target had been set at 50 percent, the
results still might have been in the vicinity of 5 percent, as this would have
been asking too much, creating frustration.

12.2  Tree Diagram and Process Decision Program Chart

Any project should have a project plan, as has already been emphasized. This is
also true for improvement projects, especially in the implementation phase of
such a project, which can often be viewed as a project of its own. Implementing
improvements can often take longer than the total duration of the project thus far.
It is therefore essential to have a project plan to guide this activity. Generally, such
a plan should cover the following elements:

• Activities—those tasks that need to be carried out to implement the improve‑

ment proposals
• Sequence—the order in which the activities must be carried out
• Organization and responsibility—an indication of who is responsible for
each activity, both carrying it out and monitoring its progress
242  Chapter Twelve

• Schedule—a more detailed plan for when the activities should be carried out,
including milestones for central results expected throughout the project
• Costs—estimates for the costs involved in the implementation

There are several techniques for project planning that involve different levels of
scope and complexity. One ­easy-­to-use tool, suitable for breaking down larger tasks
into activities of manageable size, is a tree diagram. This diagram can be combined
with more complicated calculation methods for the project, for example, PERT
(Program Evaluation and Review Technique) or CPM (Critical Path Method). The
approach for creating a tree diagram is as follows:

1. Generate a list of activities that must be performed to implement the

improvement proposals
2. On sticky notes, write down each activity in the form of a verb followed by
a noun
3. Arrange the activities in logical subgroups that must be performed in
4. Arrange the subgroups in an overall sequence to illustrate the entire plan
in the tree diagram

A typical diagram will look like the one shown in Figure 12.3. The result will
typically be a hierarchy of activities. The main activities, read from left to right
in the diagram, represent the main tasks of the implementation. These main tasks
will often be artificial activities (that is, names of a group of subactivities). Each of
these main activities will therefore include a number of subactivities below them,
to be carried out in the order they are presented from left to right. For each activity
in the diagram, information about deadlines, responsibility, costs, and so on, can
be attached.

Implementation objective

Main activity Main activity Main activity

3/15 6/15 10/1
1 2 3

Subactivity Subactivity Subactivity Subactivity Subactivity Subactivity Subactivity Subactivity

1.1 1.2 1.3 2.1 2.2 2.3 3.1 3.2
1/1 2/15 3/15

Figure 12.3  A principal tree diagram.

Tools for Implementing Improvements  243


A library decided to introduce a new ­computer-­based registration system.

To plan this task, the employees designed the tree diagram in Figure 12.4.
For each activity, the completion date was attached.

Introduce computer-based registration system

Buy the Train the Put the system

3/15 6/15 10/1
system staff into use

Invite Evaluate Negotiate Buy training Run Test the Transfer Start the
bids bids contract package courses skills old data system
1/1 2/15 3/15 4/1 6/1 6/15 9/1 10/1

Figure 12.4  Tree diagram for introducing a computer system.

The plan described in the preceding example is not very detailed and does not
consider any unforeseen events during the implementation. This is typical in project
plans of this type: they do not say what to do if something goes wrong. To include
this element in the planning, we can use a process decision program chart (PDPC),
which is really an extension of the tree diagram. This is a planning tool for making
detailed implementation plans that include all possible negative events and prob‑
lems that could occur along the way. Predicting such problems before they occur
makes it possible to address them. This enables some form of preemptive problem
solving, which is inexpensive compared with starting to think about solutions only
after the problem has occurred. The PDPC is most often used when a large and com‑
plex task is to be carried out for the first time, where the costs associated with failure
are exceedingly high, and where finishing by the deadline is critical.
The approach for using PDPC is as follows:

1. Generate a tree diagram for the implementation task or use one that has
already been designed. At least to start with, it is wise to use a tree diagram
that is not too complicated, as this can require undue time for analyzing
possible problems for small, unimportant activities. An appropriate com‑
plexity contains the main activities with one level below them, as in the
example shown in Figure 12.3.
2. For each element at the lowest level of the tree diagram, ask questions
like, what potential problems could occur during this activity? and what
244  Chapter Twelve

could go wrong here? For these questions, brainstorm a list of answers for
each potential problem area. When no more answers surface, examine the
list to eliminate problems that are unlikely or that are expected to have no
significant consequences. Each element should include an assessment of
the consequences in terms of time, cost, and quality.
3. Add the remaining potential problems, those that are considered sig‑
nificant, to the diagram as ­what-­if elements below the lowest level of
activities. Use a different color or shape to separate these elements
from the activities.
4. For each ­what-­if element, brainstorm possible countermeasures that can
be undertaken if the problem occurs. These countermeasures should
consist of reserve activities and indications of duration and cost. Place
all countermeasures in the diagram, which is being transformed from
a tree diagram to a PDPC. Place the countermeasures under the ­what-
­if elements and link them to the potential problems they solve. Again,
use a different color to separate them from the ­what-­if elements and the
5. Finally, evaluate each countermeasure with regard to ease of implementa‑
tion, practicality, effectiveness, and so on. Mark difficult or ineffective
ones with an X, and mark those you expect to be effective with an O.

A general PDPC is shown in Figure 12.5. Designing such a chart forces us to

anticipate all possible problems that could occur during implementation and,

Implementation objective

Main activity Main activity Main activity

1 2 3

Subactivity Subactivity Subactivity

1.1 2.1 3.1

What if What if What if What if What if What if

1.1.1 1.1.2 2.1.1 2.1.2 2.1.3 3.1.1

Counter- Counter- Counter- Counter- Counter- Counter-

measure A measure B measure C measure D measure E measure F


Figure 12.5  A general PDPC.

Tools for Implementing Improvements  245

importantly, develop countermeasures. This might change the original plan in order
to avoid potential problems as well as prepare countermeasures in case problems do


The library that designed a tree diagram for introducing a new ­computer-­based
registration system took its study even further and created a PDPC for the imple-­
mentation plan. The chart is shown in Figure 12.6.

Introduce computer-based registration system

Buy the Train the Put the system

system staff into use

Invite Evaluate Negotiate Buy training Run Test the Transfer Start the
bids bids contract package courses skills old data system

No All bids too No packages No free Hard to learn Wrong System

bidders expensive available time how to use data format crash

Cover Make own Have tailor- Use More Adapt Trial

entire U.S. system made overtime training the data operation


Figure 12.6  PDPC for the library.

12.3  Force Field Analysis

As mentioned at the beginning of this chapter, an important subtask in the implementa‑

tion phase is to create acceptance for the suggested changes and a favorable climate for
their implementation. This is a significant task that involves disciplines such as psychol‑
ogy and human resources management. A change process can be viewed as a formula
involving the following elements:

• Q = quality of the change approach

• A = acceptance of the change among those involved
• E = effectiveness of the change process
246  Chapter Twelve

The formula is:

E = Q × A.

Studies have shown that all successful change processes have high values for both
Q and A, while most failed change projects also have high Q. The shortcomings
are related to the acceptance of the changes. According to Miller (2002), there are
often six layers of resistance:

• Disagreement that there is a problem at all

• Thinking the problem is outside our control

• Argumentation that the suggested solution cannot solve the problem

• Argumentation that the suggested solution will also cause negative effects

• Creation of barriers against the implementation

• Creation of doubt about whether others will cooperate for the solution

A general piece of advice is that the more information given to those who will be
affected by the changes, the less resistance will be met. Thus, communication with
everyone affected and others who might represent obstacles to an effective imple‑
mentation is essential. This includes:

• Top management, which has the authority to decide on implementation and

allot the necessary resources for it

• Everyone involved in the process to be changed, as it is essential to motivate

them for the change

• Everyone delivering input into the process or receiving output from it, as
they could also be affected by the changes

• Other gatekeepers or people who can impact the implementation and its
progress, usually people with financial authority

When it comes to creating a positive climate for the ensuing changes, force field
analysis is a helpful tool that can contribute to creating an overview of the situa‑
tion and possible actions to improve it (Andersen and Pettersen, 1996). Force field
analysis is based on the assumption that any situation is a result of forces for and
Tools for Implementing Improvements  247

against the current state that are in equilibrium. An increase or decrease in the
strength of some of the forces will induce change—a fact that can be used to cre‑
ate positive changes. A similar method is often referred to as benefits and barriers
analysis. The approach is almost identical to force field analysis, with the excep‑
tions that forces for the change are called benefits and forces against are called
barriers. This method does not involve a diagram.
The procedure for using force field analysis is as follows:

1. Clearly define the change desired. This is information that can usually
be taken directly from the implementation plan and its improvement

2. Brainstorm all possible forces in the organization that could be expected to

work for or against the change.

3. Assess the strength of each of the forces and place them in a force field
diagram. The length of each arrow in the diagram expresses the strength of
the force it represents.

4. Consider actions that could increase the forces for the change and reduce
those against it, especially the stronger forces.

With these actions, the power balance surrounding the change could smooth the
implementation. A force field diagram is shown in Figure 12.7.


Forces for Forces against

Figure 12.7  Force field diagram.

248  Chapter Twelve


After identifying potential problems facing the introduction of a new

computer system for registering transactions, as well as countermea-­
sures for these, the library undertook a force field analysis to understand
the climate for this major change. Important forces for the implementa-­
tion included:

• The new system would dramatically improve the process of check-­

ing books in and out of the library, making it easier on employees
• The system would automatically generate reminders for overdue
books, thus replacing a currently manual job
• Searches for titles would be quicker and easier
• If the system worked as anticipated, it would dramatically improve
customer service, which currently had a somewhat tarnished

Several forces working against the change were also identified:

• A general skepticism toward computer tools among many of the

employees, whose average age was 42
• Fear of computer problems that could not be solved as easily as
with manual systems
• Fear of becoming redundant due to reduced manpower needs

Put into a diagram with the forces’ strength represented by the arrow
length, the resulting picture of the situation is as shown in Figure 12.8.

Introducing a computer-based registration system

Forces for Forces against

Simplified transactions Computer skeptics

Automatic reminders Computer problems

Quicker search Redundancy fear

Better service

Figure 12.8  Force field analysis for a new library computer system.
Tools for Implementing Improvements  249

For each of the forces, several measures were identified that could
point the library in the right direction: visits to other libraries where com-­
puter systems had already been introduced with good results, general
information and training in computer use, and so on.

Finally, some general advice about effective implementation:

• To ensure full support for the changes, involve everyone responsible for
results from the process being improved
• Try to elicit inspiration from those involved in the project
• Follow a clearly communicated plan
• Keep the affected employees informed about progress and achieved results
• Emphasize the importance of patience—changes do not happen overnight

Perhaps the greatest challenge in this phase of the improvement project is to main‑
tain the team’s intensity to ensure that the implementation is carried out. Therefore,
continuous monitoring of progress is necessary with regard to:

• Time spent compared with the planned duration

• Resources consumed compared with the implementation budget
• The quality of the results
Sample University—Improving
Student Satisfaction

nfortunately, it is not easy to find good working examples of the toolbox
in a coherent improvement process. As with the literature in this field,
there are a number of good case descriptions that portray the use of one
or more isolated tools, but very few that illustrate the interplay between the tools.
Therefore, this chapter contains a description of an organization, Sample Univer‑
sity, and its systematic use of a coherent system for process improvement. The
description is based on a real case, but parts of the story have been embellished to
demonstrate the use of different tools in a larger context.

13.1  Description of the Organization

Sample University is a small university, with about 3000 students at the bachelor,
master, and PhD levels. Courses of study include social sciences, engineering sci‑
ences, and management subjects. Sample University is divided into six faculties,
each containing three to six departments. The university has about 150 professor‑
ships, close to 400 other faculty positions, and a technical/administrative staff of
about 600 employees.
The number of applicants to Sample University has been quite stable, although it
has displayed a small increase in the past year. There is, however, an increasing pro‑
portion of students who choose shorter studies or who quit or transfer before complet‑
ing their degrees. Sample University receives some funding from the state as well as
some support from industry and private donations, but the bulk of its income comes
from tuition fees. The university is not among the ­highest-­ranked schools in the United
States within its areas of study, but it holds a good position above the average.
The immediate challenge was to drastically increase student satisfaction and
to improve the university’s ranking in order to increase the ­long-­term number of
applicants. This chapter describes how Sample University handled this challenge.

252  Chapter Thirteen

13.2  Development of Performance Priorities

For years, Sample University had based its operations on a performance measure‑
ment that included:

• National ranking performed by different agencies that was based on many

aspects of the universities

• Development in the body of applicants

• Study completion percentage

• Separate course evaluations

From these performance measures, it was difficult to prioritize the areas or pro‑
cesses of desired improvement. At the same time, there was a wish to start with a
pilot project because of the limited ­short-­term resources for this work and because
this type of project would help individuals become familiar with the improvement
process. To help prioritize the resources for the pilot project, a better set of perfor‑
mance measures was needed. Ideally, a comprehensive documentation of important
processes would have been undertaken, but the pilot project and the preparations
for it did not allow for this. The university’s management team, consisting of core
academic and administrative personnel, decided therefore to conduct a measure‑
ment of the customer satisfaction of a selection of students. The management team
wanted to combine this with input from companies that had hired graduates of
Sample University over the last two years.
A questionnaire was designed and distributed to about 200 students from dif‑
ferent departments. Concurrently, telephone interviews with both alumni and man‑
agers from approximately 20 companies in the area were conducted. There were
eight main areas of evaluation:

1. Usefulness of the academic content of the studies

2. The faculty’s ability to convey the academic message

3. Routines for admissions and other administrative student policies

4. Housing and leisure facilities on campus

5. Pedagogical tools used in teaching

6. Tuition fees and other study costs

7. Relevance of the education for later jobs

8. The students’ abilities to cooperate in ensuing work situations

Sample University—Improving Student Satisfaction  253

Even though some variation was seen at the start among the faculties, average
evaluation scores were calculated for the whole of Sample University. Because
the competition among universities is fierce, the management wanted to form an
impression of the situation compared with other colleges. Many universities col‑
lected and published corresponding data in study catalogs and similar material.
From such sources, Sample University collected data from other universities in the
state as well as from national competitors. The performance levels were compared
using a spider chart, among others, as shown in Figure 13.1.
In some areas, the results were far from encouraging. The areas displaying the
largest gaps were:

• Ability to convey the academic message

• Pedagogical tools
• Administrative routines
• Students’ ability to cooperate in later work positions

The results did not indicate what areas should be focused on, so Sample University
used a performance matrix to prioritize the areas with a gap. The performance
matrix is shown in Figure 13.2.

National A
State B National B
Ability to Ability to
cooperate convey message

State A
State C

Educational Administrative
relevance routines

Sample University

Tuition fees Housing

and costs and leisure


Figure 13.1  Spider chart for external comparison.

254  Chapter Thirteen

9 6

4 7

Current performance
3 2

1 5 9


Figure 13.2  Performance matrix for establishing priorities.

Altogether, these analyses identified two related areas, ability to convey

the academic message and pedagogical tools, as the areas to focus on. It was
assumed that these were somehow connected to students’ ability to cooper‑
ate after completed studies. Focus was therefore put on the core process of
teaching. The Department of Physics was selected as a pilot area. This was a
field that could be difficult to teach, and this department had one of the low‑
est scores at the university and had experienced decreasing applications over
the last few years. Results developed here would later be applied to the rest of
Sample University.

13.3  Understanding the Current Processes and Performance


After the Department of Physics was selected as the pilot area, a working group in
the department was established to carry out the project. Together with the central
pedagogical coordinator at the university, the working group consisted of:

• Two professors from different academic fields

• One research assistant
• One person responsible for student laboratory work
• Two students at different levels of study

To enable the group to perform the work, substitute teachers were provided for
the teaching personnel, and the participating students received an hourly wage.
Sample University—Improving Student Satisfaction  255

Introductory discussions concerning which processes could be defined within the

department’s teaching activities led to viewing one course from start to finish as a
process. As the different courses in the department were organized almost identi‑
cally, the group decided to start with only one course, assuming that any improve‑
ments achieved could be applied to the others.
The first major task was to agree on how the process was currently conducted.
Closer inspection revealed some minor differences between courses, so the work‑
ing group chose to document something close to a standard process. Through ple‑
nary discussions over a period of three to four weeks that focused on the level of
detail to use, the group arrived at the flowchart shown in Figure 13.3.


the students


for lectures





and grading

Report on
course grade

Figure 13.3  Flowchart for the course process.

256  Chapter Thirteen

Before the start of the semester, a brief course planning was performed
where lecture topics were defined and other practical details were settled. At
the beginning of the semester, the students taking the course were greeted
with an introductory gathering where the course plan and curriculum were
presented. The course was primarily composed of weekly cycles of theory lec‑
tures followed by a compulsory calculation exercise. The lectures were based
on slide shows and blackboard teaching in relatively large classes of between
100 and 150 students. The calculation exercises were completed by the stu‑
dents on their own and then handed in for evaluation. Students were required
to complete a certain number of exercises before they were allowed to take the
final exam.
Toward the end of the semester, a summarizing lecture was arranged, where
the main contents of the curriculum were presented. The courses concluded
with a written exam that was subsequently graded, and the grade was reported
to the students. If a student failed the exam, the student had to retake it, up to
three times.
To enable measuring the improvement project’s results, the university man‑
agement measured the current performance level on the basis of its statistics on
student grades. On a scale from 1 to 5, where 1 was weak and 5 was strong, the
results for the department were as follows:

• Usefulness of academic content: 3.4

• Ability to communicate the academic message: 2.3

• Pedagogical tools: 1.7

• Relevance for later jobs: 3.1

• Ability to cooperate: 1.9

Some central statistical information follows:

• On a grading scale from 1.0 (worst) to 4.0 (best), the average grade in the
department was 2.43

• The average grade varied somewhat between the student levels: the aver‑
age for ­first-­year students was 2.28, while the average of the ­highest-­level
students was 3.11

• Within six months of graduation, 64 percent of the students were


• After completing master’s degrees, 7 percent of the students proceeded to

doctoral studies
Sample University—Improving Student Satisfaction  257

13.4  Analysis of the Performance Shortcoming

To develop an overview of the pedagogical arrangement and the inability to get the
message across to the students, the working group used a fishbone chart. Causes
were sought for lacking ability to communicate the academic content. Through
brainstorming and subsequent discussion and sorting of causes, the working group
developed the diagram in Figure 13.4.
Some of the main causes of the problems were the following:

• The teaching was too theory focused, with little emphasis on application
and industrial use
• Coherency was missing in the studies and between individual courses
• Learning was based on ­old-­fashioned principles (for example, one professor
per 120 students, pure calculation exercises, and “learn and forget” exams)

To better understand the interplay between the causes and their effects on the
problem, the working group constructed a matrix diagram to analyze the rela‑
tionships between them. As there were really only two dimensions to the vari‑
ables—causes and problems—an L-matrix was used, but with the inclusion of
a roof matrix to analyze connections between the different causes. The result is
shown in Figure 13.5.

Academic Study
content arrangement

Too theory Lacking demon- Missing course coherency

based stration of
Little industry
focus No industry
Too static contact
Not updated about
Low progression
new research Lacking
between courses
Material too Individual Only ability
theory based learning lectures
textbooks Calculation exercises
Only books too theory based
Old-fashioned No group
No industry lab equipment work Exams are
standards “learn and forget”
Teaching Teaching
material methods

Figure 13.4  Fishbone chart for lacking communication ability.

258  Chapter Thirteen

Not updated on research

Lacking study coherency

Old teaching material

Lacking applications
Individual learning

Too theory based

weak positive

Only lectures
Poor exam
strong positive

Academic content
Ability to convey
Pedagogical tools
Relevance for later job
Ability to cooperate
Grade levels

Figure 13.5  Matrix diagram for Sample University.

Completing these analyses gave the working group good insight into the
underlying causes for the disappointing performance levels. Thus, it was con‑
cluded that the problem analysis was complete and the search for solutions was

13.5  Generating Ideas and Improvement Proposals

The working group envisioned using several improvement tools in this project.
From the outset, many group members had ideas about what could be done dif‑
ferently and better; however, few had been clearly formulated and they were far
from implementation. This richness of potential ideas led to the assumption that
idealizing could be a suitable tool. Using it, the working group could brainstorm
the ideal process for academic communication based on the group’s latent ideas
and then use the results as a basis for improvement.
The idealizing approach was selected, and elements of the ideal process were
generated through brainwriting. Since the steps of the process were quite indepen‑
dent of each other, ideas could be generated for new ways of doing things at the
individual steps. After a ­two-­day gathering where ideas were written and subse‑
quently sorted, grouped, and reformulated, the group ended up with the following
improvement suggestions:
Sample University—Improving Student Satisfaction  259

• Start each course with a ­three-­hour motivational gathering where, in coop‑

eration between the lecturer and the students, a common understanding of
the course objectives and content would be created
• Communicate theory through colloquial sessions between the lecturer and
a maximum of 15 students
• Teach practical applications of theory by solving practical group exercise
problems that are presented orally, and subsequently discuss solutions
• Use one exam consisting of one individual theory part, oral or written, and
one practical group work part with common grading
• Encourage students to give feedback on arrangement and content through‑
out the course

To supplement these proposals with even more suggestions, the working group
wanted to perform a process benchmarking of a few other schools. Due to the lim‑
ited time and money for such an activity, this was reduced to a comparison against
one other university. The state where Sample University is located presents an
award each year to the higher education institution that has been noticed for excep‑
tional pedagogic merits. This year’s award was given to a small university special‑
izing in industrial management that is located in a neighboring town. A request
to this college was favorably answered, and Sample University was welcomed to
discuss its teaching methods. A ­half-­day visit with presentations and discussion
around a typical course gave the following additional improvement suggestions:

• Within each department, a complete evaluation of the course composition

was conducted every third year. During this evaluation, the following issues
were focused on:
— All courses should fit together as a whole

— The contents of the individual courses should not overlap too much
— New research at the college itself and elsewhere should be included in
the courses

• In addition to colloquial teaching, and in line with what Sample University

had already discussed, this college also used a system where the students, in
groups and in turn, had to recite the most important aspects of last week’s
chapter to the rest of the group.
• In courses where this was possible, at the start of the semester, students
would visit an industrial company or a research institution, where applica‑
tions of the course’s theory were demonstrated.
260  Chapter Thirteen

• Throughout the course, students worked on solutions to the problems

observed during the excursion. At the end of the course, another visit to
the company/research institution was arranged where the host organization
presented its own solutions to these problems.
• Before the exam, a ­four-­hour gathering was arranged to discuss the curricu‑
lum. The exam, however, was organized according to the suggestion from
Sample University about a group exam.

Combining Sample University’s own suggestions produced through brainwriting/

idealizing with the benchmarking outcomes created an extensive list of improve‑
ment proposals.

13.6  Implementation of Improvements

The preceding phase, with the resulting set of improvement proposals, was concluded
in early October. Objectives for the ­short-­term implementation at the Department of
Physics were defined as planning, preparing, and arranging for a new educational
system for the spring semester in three of the department’s six courses.
The individual lecturers responsible for the courses, together with the working
group, were responsible for this task. Because these lecturers did not teach any fall
semester courses, they could devote the next three months to the implementation.
To plan the work, a very simple implementation plan outlining some milestones
was devised. Over the next three months, most of the following tasks were under‑
taken in all three courses:

• Identifying new and better textbooks, as well as other multimedia

• Outlining a program for colloquial teaching with smaller classes and stu‑
dent recitation
• Planning industry and/or research excursions during the semester
• Producing new group exercises demonstrating practical applications of the
theory of the course
• Developing a new exam format based on group solving
• Developing a system for active feedback throughout the semester from the

In addition, a gathering was planned for the middle of October, where all the lec‑
turers would debate and coordinate the coherency of the set of courses offered by
the department.
Sample University—Improving Student Satisfaction  261

Some resistance to the upcoming adjustments was expected from the faculty
group, the administration, and the students. In an effort to ease the implementation
and the changes it would entail, force field analysis was used. Following a brain‑
storming of forces for and against, the force field diagram shown in Figure 13.6
was constructed.
Some countermeasures to these forces were identified as follows:

• New information about the advantages of the new arrangement was distrib‑
uted to the students
• Motivational gatherings where a pedagogical expert would describe experi‑
ences from other teaching institutions were organized for the lecturers
• The rules were changed to facilitate common grading

Prior to the course commencement in January, practically all of these changes were
implemented. The only thing missing was the preparation necessary for a common
exam, but there was plenty of time to do this over the course of the semester.

Changing the teaching arrangement

Forces for Forces against

Fear of unknown
Better study value

More rewarding study Increased workload for

situation the lecturers

Formal problems connected

Better grades
to common grading

Need for additional small

Easier to get a job
colloquial rooms

Increased number
of applicants

Better university

Figure 13.6  Force field analysis for the implementation.

262  Chapter Thirteen

After the spring semester courses were completed and in the months follow‑
ing, new measurements of the performance levels were conducted. The average
results for the three courses were as follows:

• Academic content usefulness: 3.9

• Communication ability: 4.5
• Pedagogical tools: 4.4
• Relevance for later jobs (few measures): 4.2
• Cooperation ability (few measures): 3.8
• Average grades: 3.28

As a result of the unconditional success, the department management decided to

do the same in all the other courses. At the same time, the university management
wanted the working group to make preparations for and support the same changes
in all other departments at Sample University.
Green Carpet Seed—Revamping
the Business Model

he Green Carpet Seed case study, to some extent, is also based on a ­real-­life
improvement case. However, as was the case with the Sample University
study, this project used only some of the tools presented in the book. There‑
fore, I have again taken the liberty of embellishing the story here and there, both to
make it fit the overall improvement model of this book and to better illustrate the
use of some tools and techniques. The case is from a company that produces seed
for grass lawns and that found itself in trouble when sales continually declined
for three years. Facing either extinction or a radical rethinking of its business, the
company embarked on an ambitious improvement project that redefined its entire
business model.

14.1  Description of the Organization

Green Carpet Seed is a fairly small player in the seed industry, employing
about 150 people. Established in 1952, the company is some 50 years old and
remains family owned. It is based in a small town but is only 45 minutes from
a large city. Its production facilities and administration offices are located in
the same spot where the company was founded by the grandfather of the cur‑
rent CEO. Green Carpet Seed purchases seeds from local farmers, processes
and treats them, mixes different types of seeds into combination blends, pack‑
ages the seeds, and sells them through garden centers, plant stores, and other
The grass seed industry has radically changed since the founding of the
company. Where there used to be small seed suppliers all over the country,
there are now just a few left. This is a result of acquisitions of smaller units
and the closing of many small production sites in favor of larger and more
industrialized plants. Green Carpet Seed has avoided a takeover by responding

264  Chapter Fourteen

to tougher competition through rationalization efforts and automating parts of

the production.
For three years, however, sales have steadily declined, dropping 35 per‑
cent since their peak. Although the company has always enjoyed a comfort‑
able financial position, funds are eroding and there have been serious talks
about whether it will make it through another year of falling sales. The CEO
and his small executive group decided something drastic had to be done and
launched an improvement project to look into how Green Carpet Seed could
turn the situation around. A task force was established, bringing on board the
sales manager, a biologist from R&D, a production worker, and a student from
the local community college who was working on her master’s thesis and had
proved extremely creative. They were charged with a mandate to propose ways
to regain the lost sales. The project has now been completed, with a completely
new business model as the result.

14.2  Development of Performance Priorities

Having followed a strategy of remaining independent and not trying to grow

by buying competitors, Green Carpet Seed had for many years been some‑
what cut off from discussions about the industry structure and its development.
People had eyes and ears and saw the trends, but felt a little out of touch with
the undercurrents of the sector. To reacquaint themselves with the latest think‑
ing and to gain a better insight into the competitive position of the company
itself, the team wanted to conduct an analysis of the competitive situation in
the industry.
Together with a consultant who had worked with many players in various
stages of the seed supply chain, the team spent a whole day discussing competitive
forces. A long list of issues came up, some of them representing very scary sce‑
narios. Trying to eliminate items of less importance and focus on forces that would
have a strong impact on the company, they ended up with a handful of issues they
needed to be acutely aware of throughout the project:

• The consolidation of seed suppliers was continuing, with a few companies

growing quite large and still looking for smaller units to buy, either for their
capacity and customers or to close them down as competitors.
• Larger competitors courted local seed suppliers, and many had signed
supply contracts with these competitors as well. In some cases, they were
forced to sign exclusive contracts or preferred customer clauses that made it
difficult to deliver to Green Carpet Seed.
Green Carpet Seed—Revamping the Business Model  265

• Some of the larger garden center chains developed their own brands for
products such as seeds, flower bulbs, fertilizer, and even garden equipment.
They still needed to cooperate with seed producers, but they negotiated very
low prices in return for stable and larger volumes, and they would not allow
the suppliers to market their own brands.
• With home owners busier than ever, a general trend the past few years had
been the search for ­easy-­to-maintain gardens. Some did away with lawns
altogether in favor of patios and graveled areas; others kept their lawns but
did no upkeep, meaning they reseeded less frequently.
• Substitute products had existed for quite a while, but they were now being
developed into stronger contenders to seeded lawns. ­Roll-­out lawn was easy
to install and had proved more resistant to wear than before, and many home
owners chose to replace the lawn every few years with new rolls instead of
using seeds. Artificial grass was also becoming more common in private
gardens and could further threaten seed sales.

With all these threats looming on the horizon, the team felt anything but optimis‑
tic. Some of the team members had more than 20 years’ experience in the seed
business, but none of them were really able to tell what would dictate the win‑
ners in the future. At one stage, they argued that lower prices would be the key
to increasing sales, but this was followed by a conviction that customers did not
mind the higher prices as long as the seeds gave them beautiful lawns. Realizing
they were not able to debate their way to an answer, the sales manager suggested
they try “calculating” an answer. In sales, they used various methods to determine
the importance of sales channels and regions, and he figured a similar approach
could be used here.
Looking into this further, he realized that what he was looking for was the tool
called criteria testing. The team members came up with a handful of critical suc‑
cess factors (CSFs) they believed determined the sales. They weighted them and
listed business processes they believed influenced the CSFs. Carefully discuss‑
ing how each business process had an impact on each of the CSFs, they assigned
impact factors and ended up with the matrix shown in Figure 14.1.
To their surprise, a process they had seriously considered leaving out of the
analysis came out on top: dealer management and support. During the analysis,
they had made a point of not looking at the matrix itself, but assigning impact
factors one by one to avoid letting their expectations of an answer dictate their
assessments. When they finally compiled the numbers, this was a real ­eye-­opener.
Checking the analysis again, they conceded that it made sense; the process truly
did influence most of the success factors. With this result, the question was, what
could be done about this process to gain a competitive edge?
266  Chapter Fourteen

Convenience of use
High-quality seeds

Solid packaging

Brand name

Total score
Low prices
Business processes Weight 3 1 2 1 3

Seed sourcing and supplier management 9 2 2 1 3 17

Seed processing 9 2 2 1 3 17
Product and process R&D 6 2 2 1 6 17

Seed packaging 3 1 6 1 6 17
Marketing and brand name promotion 3 1 2 3 3 12

Dealer management and support 6 3 2 3 6 20

Distribution 6 2 4 1 3 16

Figure 14.1  Criteria testing matrix for Green Carpet Seed.

14.3  Understanding the Current Processes and Performance


The sales manager knew very well how the company worked with its distribution
channels, but the rest of the team had virtually no insight into this part of the busi‑
ness. To educate them, the sales manager enlisted the help of a couple of sales‑
people in constructing a process model. Spanning many different departments and
units in the total supply chain, a ­cross-­functional flowchart was the obvious choice.
Somewhat simplified, the process is depicted in Figure 14.2.
The sales manager was quite pleased with the flowchart, especially since
one of the salespeople mentioned that the neighborhood kids who mowed
lawns were affected by the trouble the seeds from Green Carpet Seed caused
in terms of good grass growth. They added this “player” to the chart and pre‑
sented it to the rest of the team. Having looked at it, the others asked what the
difference was between this process and distribution. The first reflex of the
sales manager was to dismiss the question as silly: they were two completely
different things. Looking at the chart again, however, he realized the question
was warranted.
Thinking they were modeling the dealer management and support process,
they had indeed drawn the sales and distribution process. Mulling this over, the
sales department people had to acknowledge that they didn’t really have much of
a dealer support process. They accepted their orders well enough and sometimes
went to see the dealers, generating more orders. However, once the orders were
Green Carpet Seed—Revamping the Business Model  267

Dealer management and support

receiving Shipment

incoming order




Receipt of

Purchase Receipt of
decision seeds

of seeds

Lawn mowing

Figure 14.2  ­Cross-­functional flowchart for the dealer management and support process.

secured, everything was handed over to someone else to deliver the seeds to the
stores. There was very little cooperation with the sales outlets about things like
promotions, campaigns, meeting customers, and so on.
Another realization came when the student on the team asked questions about
the category of the kids who mowed lawns, added just for fun. Why do garden own‑
ers pay kids to mow their lawn? And if they hire mowers, do they also hire ­gardeners?
268  Chapter Fourteen

If the garden owners don’t tend to their lawns themselves, how do they know if and
when they need seeding? And if they don’t do it, who does? The kids mowing
the lawns? The gardeners? She seemed to have an idea behind these questions,
and although the sales manager didn’t quite see where she was headed, something
stirred at the back of his mind. He knew that there were many successful companies
in the city that provided only one service: tending to plants in office buildings, train
stations, and other public spaces. Sometimes the owners of the building or facility
owned the plants; in other cases, these companies owned the plants and leased them
to customers, in addition to providing care services for them.
Both the sales manager and the student seemed to arrive at the same basic idea
at almost the same time: was it possible to add a service element to the sale of seeds?
As with the plant service companies, this could be a ­revenue-­generating service of its
own and at the same time help sell more of the company’s seeds. The rest of the team
was quite enthusiastic about the idea but had many objections as well. To move ahead
and decide whether this could be a solution, they debated the idea thoroughly.

14.4  Generating Improvement Ideas

Although the team members believed they might be on to something with this
added service idea, they felt a need to view it from several angles. The biologist
had learned the technique of the six thinking hats at her university and suggested
they try the approach. Getting the others to agree, she quickly explained the prin‑
ciples to them and they met one morning, armed with plenty of coffee, to run a
session. Two more people joined them: one of the salespeople who had helped out
with the flowchart and the CEO himself.
They shared the hats among them, even putting on real hats, and got ready to
discuss how they could add a service element to the seed sales. The session ran for
almost four hours, stopping only for a short break. When they finished and the stu‑
dent, wearing the blue hat, summarized their debate, they had several very creative
ideas on the table:

• Closer cooperation with dealers. Working more actively with dealers, they
could train them in helping customers select the right seeds and determine
the correct time to use them. They could also train them in arranging cam‑
paigns in their stores. Launching an annual lawn week, they could have
stands in the stores, supported by advertisements in different media, to
increase general lawn awareness. They could offer expert assistance to cus‑
tomers who came to the stores during that week, and sales would probably
increase during that time, as well as the rest of the year, due to the knowl‑
edge gained by the customers.
Green Carpet Seed—Revamping the Business Model  269

• A lawn consulting service could be offered to customers. An expert would

study the lawn, or the area where the customer wanted to establish a lawn,
and give advice on which type of seed to use, when to sow, how to treat the
lawn, and so on.

• Lawn mowing and lawn tending services could be developed. Working with
both students and other ­low-­cost employees, the company could offer lawn
mowing, airing, reseeding, and other care services.

• They could work with seed dealers to develop an automatic seed delivery
process. The thinking was that a computer system could keep track of all
the customers, where they lived, when the lawn was established, which
seeds were used, soil conditions, weather conditions during the year, and
so on. Using this information, the computer would calculate when reseed‑
ing was beneficial and would either remind the customers to come and buy
seeds or deliver seeds directly to them. This service could be combined with
tending services, where a lawn care person would sow the seeds and water

• Going beyond the lawn, it would be possible to set up an entire garden

care service system, where cooperation with gardeners, lawn tenders,
snow removers, and so on, would provide a complete range of services to

Several of these ideas could be combined, and common to all was that they would
directly or indirectly lead to increased sales, with many of the services also making
money on their own.

14.5  Developing an Improved Process

The team members, including the CEO, were extremely enthusiastic about their
many ideas. Realizing that few, if any, of these approaches had been tried before in
the industry, they felt they were on the brink of a true breakthrough. They understood
that most of the ideas, if implemented, could be rather quickly copied, but they hoped
the ­first-­mover advantage would keep them ahead of the competition for a while.
Although they had many more or less radical ideas, they still didn’t quite see
how such new services should be organized and designed. The next step was to
attempt designing the services and the processes required to deliver them in some
more detail. Understanding how much change this would require, they quickly
adopted the term reengineering and set about drawing up the new processes.
Reading up on BPR, they had no doubt that they would need to use clean sheet
270  Chapter Fourteen

They brought in a person from one of the dealers they knew well and the
founder of one of the ­better-­known plant service companies in the nearby city to
strengthen the team. To guide the discussion, they went through many of the rec‑
ommended questions to prompt their creativity further:

• Which needs of the customers can we fill with a new service?

• With which organization can we offer services that fill these needs? Through
the stores, directly from the company, or through an independent service

• Where and when can the service be delivered?

• How will the needs be fulfilled? What business processes, equipment, per‑
sonnel, and so on, will be required?

• Can we learn something from the plant service industry and how it launched
its services?

In the end, they decided to proceed with two main ideas. The first was the annual
lawn week event. This idea was fairly simple to develop further, and it was quite
similar to campaigns run by other industries—for example, the annual “Go bike!”
event that local sporting goods stores arranged every spring. A dedicated team
from sales and marketing was given the task of pursuing this idea.
The other idea was an integrated lawn service, which was a combination of
several of the ideas they had developed during the six thinking hats session. Hav‑
ing discussed it for several weeks, including calculating cost and income, they
concluded that such a service would probably be difficult to run at much of a profit.
They had therefore agreed that it should be priced so low that it would attract a
maximum number of customers. This way, it would be most effective in fulfilling
its main purpose: generating higher sales of seeds, which would continue to be the
main source of revenue.
Without going into details, the service was intended to be organized and
designed as follows:

• Green Carpet Seed would set up a new company to operate the service. It
had considered a joint venture with one of the large garden center chains,
but feared “marrying” one would cause other dealers to switch to other
• The new company would deliver a ­three-­tier integrated lawn service pack‑
age consisting of:
— The aforementioned lawn consulting service. This would ensure that new
customers would be continually recruited into the program.
Green Carpet Seed—Revamping the Business Model  271

— The lawn tending service, employing a network of students from local

high schools and colleges and overseen by a team of gardeners.
— An automatic seed delivery service monitored by computer, as described
• This computer system would very much be at the core of the ­full-­service
package, managing the delivery of the various services.
• Any fertilizer and equipment required would be sourced through local deal‑
ers, but Green Carpet Seed would be used for seed deliveries.

The final proposal for the establishment of this new integrated service package was
naturally much more extensive than this and outlined a number of details. Com‑
prehensive financial calculations supported the service design to show its feasibil‑
ity and impact on seed sales. Having delivered the proposal to the CEO, the team
temporarily disbanded to await the decision of the executive team.

14.6  Implementation of the New Service

The executive team failed to decide, being squarely split down the middle, and the
proposal was sent to the board. Because Green Carpet Seed was a ­family-­owned
company, the board tended to be conservative and was skeptical of the idea. Speak‑
ing warmly for the idea and warning that bankruptcy was the likely outcome if the
plan was not implemented, the CEO managed to secure the board’s approval in the
The improvement team was asked to resume its work and to start planning the
establishment of the company. It searched for a manager of the company, recruited
one from one of the plant service companies, and proceeded with the planning. In
this phase, several versions of force field diagrams were developed, addressing dif‑
ferent stakeholders that would be affected by the plans. These diagrams gave the
team many important insights into how competitors, dealers, and future employees
could react to this move.
Next, the team developed an overall implementation plan in the form of a
tree diagram. This was combined with the outcome of the force field analyses
to construct a PDPC for the process. A simplified version of this is shown in
Figure 14.3.
The new company was launched a little more than a year after the project was
started. At this point, red flags were waving, signaling the imminent death of the
mother company. Only by convincing the bank that the new service would revital‑
ize sales was Green Carpet Seed allowed to go on for another few months.
Right from the beginning, customers flocked in to sign up for the service pack‑
age. Impressed with the responsibility the company would take for their gardens,
Launch new integrated lawn service
272  Chapter Fourteen

Establish new Develop new Install computer

company services system

Develop lawn Develop lawn Develop seed Specify system Select vendor and
Legal paperwork Recruit manager Recruit employees Request tenders
consulting tending delivery requirements purchase system

Finding enough Organizing the

Finding good Working more Organizing the Low in-house IT Few vendors Sufficient training
temporary temporary
name closely with dealers dealer interface expertise available for users
employees employees

Work with local Work with local Create Web

Hire marketing Appoint liaison Establish clear Search for global Develop training
employment schools and appointment Hire consultant
bureau person contracts vendors program
O X agency O colleges X system X O O X O

Figure 14.3  PDPC for Green Carpet Seed.

Green Carpet Seed—Revamping the Business Model  273

they were happy to pay the modest fee asked. After six months of operation, the
new company had a client list of close to 2000 lawn owners. The operating margin
for the new service company was only 1.2 percent, barely breaking even. Seed
sales for the parent company, on the other hand, had increased by 16 percent,
almost catching up with half of the lost sales.
Template Package

his section contains different templates or forms for tools that can be used in
your improvement process. I considered developing templates that could be
used on your computer, but decided that many of these tools and techniques
do not easily lend themselves to such electronic templates. Scaling, the number of
fields, how things are drawn, and so on, make it very difficult to make templates
that could accompany the book on a CD. Furthermore, software is available for
many of the tools that are widely used. Even in programs such as Excel and Visio
in the Microsoft Office suite, many of these tools appear as standard options.
Thus, the templates included here are those that are suitable for presentation
on paper. You can copy them directly from the book, preferably in an enlarged
format, and write on them, or they can simply function as input for creating your
own templates. Use of each template is explained in the chapter where the specific
tool is described.

276  Appendix

SWOT Analysis

Strengths Weaknesses

Opportunities Threats

Trend Analysis

Time Today
Template Package  277

Spider Chart






Performance Matrix

Current performance

1 5 9

278  Appendix

Criteria Testing

CSF 1 2 3 4 5 Total
Processes Weight score
Template Package  279

Quality Function Deployment (QFD)

Direction of improvement




1 2 3 4 5



Absolute importance

Relative importance

Organizational difficulty

Absolute importance
corrected for difficulty
Relative importance
corrected for difficulty

Roof Roof Weight

Strong positive Strong 9
Positive Medium 3
Negative Weak 1
Strong negative
280  Appendix

Relationship Mapping

Own company

Flow of goods Information


Cross-Functional Flowchart
Template Package  281

Check Sheet

Total number of
Problem Period 1 Period 2 Period 3 per problem

Total number of
problems per period

Pareto Chart



282  Appendix

Cause-and-Effect Chart

Employees Methods

Environment Equipment

Five Whys Analysis





Template Package  283

Scatter Chart



284  Appendix

Relations Diagram

in out

in out in out

in out in out

in out

Is–Is Not Analysis

shortcoming Is Is not Distinction

What occurs?
What objects
are affected?

Where does
the problem

When does
the problem

Extent of

Who is
A/B A/C A/D A/E A/F B/C B/D B/E B/F C/D C/E C/F D/E D/F E/F Total

Paired Comparisons

Number of
Template Package  285
286  Appendix

Control Chart


Center line


Force Field Analysis

Forces for Forces against

Ackerman, R. B., R. J. Coleman, E. Leger, and J. C. MacDorman. 1987. Process Quality
Management & Improvement Guidelines. Indianapolis, IN: Publication Center, AT&T
Akao, Yoji, ed. 1990. Quality Function Deployment: Integrating Customer Requirements
into Product Design. Cambridge, MA: Productivity Press.
American Productivity & Quality Center (APQC). 1993. Basics of Benchmarking (Course
Material). Houston, TX: APQC.
Andersen, Bjørn, and Tom Fagerhaug. 2001. Performance Measurement Explained:
Designing and Implementing Your ­ State-­of-the-Art System. Milwaukee, WI: ASQ
Quality Press.
Andersen, Bjørn, and Bjørnar Henriksen. 2004. Performance Measurement System Suc‑
cess Depends on Design Alignment with the Core Purpose of Its Implementation.
Proceedings from PMA 2004—Public and Private Conference, Performance Measure‑
ment Association, Edinburgh, Scotland, July 28–30.
Andersen, Bjørn, and ­Per-­Gaute Pettersen. 1996. The Benchmarking Handbook: ­Step-­by-
Step Instructions. London: Chapman & Hall.
Aune, Asbjørn. 1985. Kvalitetssirkler: Problemløsningsgrupper for personlig vekst, kvalitet
og produktivitet [Quality Circles: Problem Solving Teams for Personal Growth, Qual‑
ity, and Productivity]. Oslo, Norway: Universitetsforlaget.
———. 1993. Kvalitetsstyrte bedrifter [Quality-Managed Companies]. Oslo, Norway: Ad
Balm, Gerald J. 1992. Benchmarking: A Practitioner’s Guide for Becoming and Staying
Best of the Best. Schaumburg, IL: QPMA Press.
Bendell, Tony, Louise Boulter, and John Kelly. 1993. Benchmarking for Competitive
Advantage. London: Pitman Publishing.
Bertels, Thomas, ed. 2003. Rath & Strong’s Six Sigma Leadership Handbook. Hoboken,
NJ: John Wiley & Sons.
Bogan, Christopher E., and Michael J. English. 1994. Benchmarking for Best Practices:
Winning through Innovative Adaptation. New York: ­McGraw-­Hill.
Bossert, James L. 1991. Quality Function Deployment: A Practitioner’s Approach. Mil‑
waukee, WI: ASQC Quality Press.
Boxwell, Robert J. 1994. Benchmarking for Competitive Advantage. New York:
Bredrup, Harald. 1995. Performance Measurement in a Changing Competitive Industrial
Environment: Breaking the Financial Paradigm. PhD thesis, Norwegian Institute of
Technology, Trondheim, Norway.

288  Bibliography

Burr, J. T. 1990. “The Tools of Quality: Part I: Going with the Flow(chart).” ASQC Quality
Progress, June, pp. 64–67.
Camp, Robert C. 1989. Benchmarking: The Search for Industry Best Practices That Lead
to Superior Performance. Milwaukee, WI: ASQC Quality Press.
———. 1994. Business Process Benchmarking: Finding and Implementing Best Practices.
Milwaukee, WI: ASQC Quality Press.
———, ed. 1998. Global Cases in Benchmarking: Best Practices from Organizations
around the World. Milwaukee, WI: ASQ Quality Press.
Codling, Sylvia. 1992. Best Practice Benchmarking: The Management Guide to Successful
Implementation. Dunstable, England: Industrial Newsletters.
Cohen, Lou. 1995. Quality Function Deployment: How to Make QFD Work for You. Cam‑
bridge, MA: ­Addison-­Wesley.
Collins, Jim. 2001. Good to Great: Why Some Companies Make the Leap and Others Don’t.
New York: HarperBusiness.
Conti, Tito. 1997. Organizational ­Self-­Assessment. London: Chapman & Hall.
Davenport, T. H. 1993. Process Innovation: ­ Re-­engineering Work through Information
Technology. Boston: Harvard Business School Press.
Day, Ronald G. 1993. Quality Function Deployment: Linking a Company with Its Custom‑
ers. Milwaukee, WI: ASQC Quality Press.
De Bono, Edward. 1985. Six Thinking Hats. Boston: Little, Brown, and Company.
De Feo, Joseph A., and William W. Barnard. 2004. Juran Institute’s Six Sigma Break‑
through and Beyond. New York: ­McGraw-­Hill.
Deming, W. E. 1986. Out of the Crisis: Quality, Productivity and Competitive Position.
Cambridge, MA: Cambridge University Press.
Dertouzos, M. L., R. K. Lester, R. M. Solow, and the MIT Commission on Industrial Pro‑
ductivity. 1989. Made in America: Regaining the Productive Edge. Cambridge, MA:
MIT Press.
D’Herbemont, O., and B. Cesar. 1998. Managing Sensitive Projects: Lateral Approach.
New York: Routledge.
Eastman Kodak Company. 1990. Quality Leadership Process Guidebook. Rochester, NY:
Eastman Kodak Company.
ENAPS. 1997. Deliverable F3-4: A Set of Refined and Agreed Performance Indicators
Defined by Business Processes. Galway, Ireland: ENAPS.
Ericsson Quality Institute. 1993. Business Process Management. Gothenburg, Sweden:
Fellers, G. 1992. The Deming Vision: SPC/TQM for Administrators. Milwaukee, WI:
ASQC Quality Press.
Galloway, Dianne. 1994. Mapping Work Processes. Milwaukee, WI: ASQC Quality Press.
Gitlow, Howard, Alan Oppenheim, and Rosa Oppenheim. 1995. Quality Management:
Tools and Methods for Improvement. Burr Ridge, IL: Irwin.
Goldratt, Eliyahu M. 1990. What Is This Thing Called Theory of Constraints and How
Should It Be Implemented? Great Barrington, MA: North River Press.
Goldratt, Eliyahu M., and Jeff Cox. 1992. The Goal: A Process of Ongoing Improvement.
2nd ed. Great Barrington, MA: North River Press.
Bibliography  289

Govers, C.P.M. 1994. What and How of Quality Function Deployment (QFD). Proceedings
from the Eighth International Working Seminar on Production Economics. Igls/Inns‑
bruck, Austria, February 21–25.
Griffiths, John. 1993. “Driving Out the Old Regime.” Financial Times, August, p. 20.
Hammer, M., and J. Champy. 1993. Re-engineering the Corporation: A Manifesto for Busi‑
ness Revolution. New York: Harper Business.
Harrington, H. James. 1991. Business Process Improvement: The Breakthrough Strategy
for Total Quality, Productivity, and Competitiveness. New York: ­McGraw-­Hill.
Hayes, R. H., and S. C. Wheelwright. 1984. Restoring Our Competitive Edge: Competing
through Manufacturing. Boston: John Wiley & Sons.
Hronec, Steven M. 1993. Vital Signs: Using Quality, Time, and Cost Performance Mea‑
surements to Chart Your Company’s Future. New York: Amacom.
Hunt, Daniel. 1996. Process Mapping: How to Reengineer Your Business Processes. New
York: Wiley.
Hutchins, David. 1985. Quality Circles Handbook. New York: Nichols.
Imai, M. 1986. KAIZEN: The Key to Japan’s Competitive Success. New York: Random
House Business Division.
Johnson, H. T., and R. S. Kaplan. 1987. Relevance Lost: The Rise and Fall of Management
Accounting. Boston: Harvard Business School Press.
Kano, N., N. Seraku, F. Takahashi, and S. Tsuji. 1984. “Attractive Quality and ­Must-­Be
Quality.” The Journal of the Japanese Society for Quality Control, April, pp. 39–48.
Kaplan, R. S., ed. 1990. Measures for Manufacturing Excellence. Boston: Harvard Busi‑
ness School Press.
Kaplan, R. S., and D. P. Norton. 1996. The Balanced Scorecard. Boston: Harvard Business
School Press.
———. 2004. Strategy Maps: Converting Intangible Assets into Tangible Outcomes. Bos‑
ton: Harvard Business School Press.
Kubeck, Lynn C. 1995. Techniques for Business Process Redesign: Tying It All Together.
New York: John Wiley & Sons.
Kume, H. 1992. Statistical Methods for Quality Improvement. Tokyo, Japan: Association
for Overseas Technical Scholarships (AOTS).
Lawlor, Alan. 1985. Productivity Improvement Manual. Aldershot, England: Gower
Lynch, R. L., and K. C. Cross. 1991. Measure Up! Yardstick for Continuous Improvement.
Cambridge, MA: Blackwell Business.
McCormack, Kevin. 2001. “Business Process Orientation: Do You Have It?” ASQ Quality
Progress, January, pp. 51–57.
McNair, Carol J., and Kathleen H. J. Leibfried. 1992. Benchmarking: A Tool for Continu‑
ous Improvement. New York: Harper Business.
Meisel, Robert M., Steven J. Babb, Steven F. Marsh, and James P. Schlichting. 2007. The
Executive Guide to Understanding and Implementing Lean Six Sigma. Milwaukee,
WI: ASQ Quality Press.
Miller, K. 2002. The Change Agent’s Guide to Radical Improvement. Milwaukee, WI: ASQ
Quality Press.
Mintzberg, H. 1994. The Rise and Fall of Strategic Planning. New York: The Free Press.
290  Bibliography

Mizuno, Shigeru, ed. 1988. Management for Quality Improvement: The 7 New QC Tools.
Cambridge, MA: Productivity Press.
Peppard, Joe. 1998. “Benchmarking Business Process: A Framework and Classification
Scheme,” in Proceedings of the ESPRIT—COPERNICUS Symposium Distributed
Enterprise, Intelligent Automation, and Industrial Benchmarking, Wroclaw, Poland.
Peppard, Joe, and Philip Rowland. 1995. The Essence of Business Process ­Re-­engineering.
Hemel Hempstead, England: Prentice Hall.
Porter, M. E. 1980. Competitive Strategy: Techniques for Analyzing Industries and Com‑
petitors. New York: Free Press.
———. 1985. Competitive Advantage: Creating and Sustaining Superior Performance.
New York: Free Press.
———. 1990. Competitive Advantage of Nations. London: The MacMillan Press.
Przekop, Penelope. 2006. Six Sigma for Business Excellence: A Manager’s Guide to Super‑
vising Six Sigma Projects. Milwaukee, WI: ASQ Quality Press.
Robson, G. D. 1991. Continuous Process Improvement: Simplifying Work Flow Systems.
New York: The Free Press.
Robson, Mike. 1988. Quality Circles: A Practical Guide. Aldershot, England: Gower.
Rolstadås, Asbjørn, ed. 1995. Performance Management: A Business Process Benchmark‑
ing Approach. London: Chapman & Hall.
Savage, Grant T., Timothy W. Nix, Charlton J. Whitehead, and John D. Blair. 1991. “Strate‑
gies for Assessing and Managing Organizational Stakeholders.” Academy of Manage‑
ment Executive 5 (2): 61–75.
Scholtes, Peter R., Brian L. Joiner, and Barbara J. Streibel. 2003. The Team Handbook, 3rd
ed. Madison, WI: Oriel Incorporated.
Sink, D. S., and T. C. Tuttle. 1989. Planning and Measurement in Your Organization of the
Future. Norcross, VA: Industrial Engineering and Management Press.
Spendolini, Michael J. 1992. The Benchmarking Book. New York: Amacom.
Spitzer, Dean R. 2007. Transforming Performance Measurement: Rethinking the Way We
Measure and Drive Success. New York: Amacom.
Stalk, George Jr., and Thomas M. Hout. 1990. Competing against Time: How ­Time-­Based
Competition Is Reshaping Global Markets. New York: The Free Press.
Straker, David. 1995. A Toolbook for Quality Improvement and Problem Solving. London:
Swanson, Roger C. 1995. The Quality Improvement Handbook: Team Guide to Tools and
Techniques. London: Kogan Page.
Syan, Chanan S., and Unny Menon, eds. 1994. Concurrent Engineering: Concepts, Imple‑
mentation and Practice. London: Chapman & Hall.
Watson, Gregory H. 1992. The Benchmarking Workbook: Adapting Best Practices for Per‑
formance Improvement. Cambridge, MA: Productivity Press.
———. 1993. Strategic Benchmarking: How to Rate Your Company’s Performance against
the World’s Best. New York: John Wiley & Sons.
Wikipedia. 2006. Stakeholder.
Zairi, M., and Paul Leonard. 1994. Practical Benchmarking: A Complete Guide. London:
Chapman & Hall.

Page numbers followed by f or t refer to figures observation, 231–232

or tables, respectively. planning, 227–228
search, 228–230
A process of, 225–226, 226f
ADT analysis, 239–241 benchmarking wheel, 227, 228f
defined, 239 Black Belts, 206–207
example of, 240–241 bottleneck analysis, 151–155
procedures for, 240 example of, 153–154, 155f
arithmetic mean value, 183 identifying bottlenecks in, 153
attribute data bottleneck resources, 151–152, 152f
chart types for BPR. See business process reengineering
c, 188 (BPR)
np, 188 brainstorming, 157–159
p, 188 approaches to, 158
u, 188 benefits of, 158
automation, 213–214 for clean sheets, 215, 215f
80-20 rule, 214 example of, 159
Automotive Industry Action Group (AIAG), 35 for identifying stakeholders, 11
procedure for, 158–159
B structured, 158
bar charts. See histograms unstructured, 158
bell-shaped curves, 184, 185f brainwriting, 159–160
benchmarking, 221–236 methods of, 160
benefits of, 223 procedure for, 160
combinations of types of, 225, 225f bureaucracy, eliminating, 168–169
competitive, 223 business processes. See also modeling, of
defined, 221–226, 222f business processes; processes
generic, 223 classifying, 34–36
internal, 223 defined, 32–34
strategic, 225 with department boundaries, 30, 30f
benchmarking partners, 229–230, 230f departments and, 27–32
benchmarking studies derived from strategy and stakeholders, 37, 37f
conducting, 227–233 documenting, 39
examples of, 233–236 generic, 29f
models for, 226–227, 226f hierarchy of models for, 40–43, 41–42f
phases of horizontal, 28–32
analysis, 232–233 identifying organizational, 37–39
implementation, 233 modeling, 38–43

292  Index

business process improvement. See also constructing, 189–196

improvement project planning; tools c charts, 196
defining organizational strategy for, np charts, 194–195
17–18 p charts, 193–194
framework for, 4–7, 5f u charts, 195–196
model of, 108–109 x and R charts, 191–193
organizational modes for, 95–100 X bar and R charts, 189–191, 191f, 192f
performance measurement in, 65–67 interpreting, 196–200
planning projects for, 75–76 template, 286
skills for, 93–94 types of, 186–188
business process reengineering (BPR), control limits, 186–187
207–221 CPM (Critical Path Method), 242
approaches to, 208–209 Crawford slip method, 159–161
case study of, 269–271 criteria testing, 83–85
defined, 208 case study of, 265, 266f
example of, 217–221 example of, 85
implementation phase, 216–217 form for, 84f
process for, 209–217, 210f procedure for using, 84
implementation, 210 template, 278
planning, 210–211 critical incident technique, 123–125
reengineering, 210, 211–215 example of, 124–125
transformation, 210, 215–216 process of, 124
types of, 209 critical success factor (CSF)
case study of, 265
C defined, 83
card method, for brainwriting, 160 cross-functional flowcharts, 51–54
case studies case study of, 266, 267f
Green Carpet Seed, 263–273 constructing, 52–53
Sample University, 251–262 example of, 52f, 53–54
cause-and-effect charts, 128–131 size of, 103
example of, 131 swim lanes in, 51, 52f
fishbone charts, 128–130 template, 280
process charts, 131 cross-functional teams, 102–103
template, 282 defined, 103
c charts, for attribute data, 188 work rules for, 103
constructing, 196 CSF. See critical success factor (CSF)
change process, formula for, 245–246 C-shaped matrix diagrams, 146
check sheets, 117–119, 136f cycle times, reducing, 172–174
example of, 118–119, 118f
procedure for, 117–118 D
template, 281 data
chronic variation, 183, 185 attribute, 186–188
clean sheet reengineering, 209, 214–215 variable, 186–188
cluster sampling, 114 de Bono, Edward, 161–162
competitive force analysis, 20–21 defective vs. defects, 187–188
example of, 21–22 Deming, W. Edwards, 6
constraints, theory of, 151, 152 Deming wheel, 6–7, 7f
continuous improvement, need for, 4 department boundaries, 30, 30f
control charts. See also statistical process departments, business processes and, 27–32
control (SPC) development processes, 36
for attribute data DMAIC (define, measure, analyze, improve,
c, 188 and control) process, of Six Sigma,
np, 188 204–206
p, 188 documentation
u, 188 of business processes, 39
Index  293

80–20 rule, 125–126 idealizing, 177–178, 177f
automation and, 214 example of, 178
e-learning approaches, 94 IDEF0, 48–51, 50f
emergent strategies, 18 implementation phase, of BPR, 210, 216–217
ENAPS (European Network or Advanced improvement culture, stimulating, 103–105
Performance Studies) project, 36, 36f improvement project planning
ESIA (elimination, simplification, integration overview of, 75–76
automation) rules, 212–213 tools for
European Foundation for Quality Management criteria testing, 83–85
(EFQM), 35 performance matrix, 81–83
quality function deployment (QFD),
F 85–92
feedback spider charts, 78–81
from customers, 9 trend analysis, 76–78
from employees, 9 improvement projects
financial indicators, 71–72, 72t preconditions for, 94
fishbone charts, 128–130, 128f, 130f, training for, 94
257f improvements
producing, 129 maintenance for, 3, 3f
five whys analysis, 132–133, 132f necessity for, 1–2
example of, 133 stakeholder analysis for, 10
procedure for, 132 stakeholders expectations of, 9–17
template, 282 tools for
flowcharts, 255f benchmarking, 221–236
cross-functional, 51–54 business process reengineering,
designing, 48 207–221
divided into process areas, 54–55 idealizing, 177–178
example of, 47f, 48, 49f quality function deployment, 179–182
paper/pencil or computers for constructing, Six Sigma, 204–207
62–63 statistical process control/control charts,
several-leveled, 55–58 182–203
with statistics, 58–62, 61f streamlining, 167–176
symbols for, 46–47 tools for implementing, 237–239
traditional, 46–51 ADT analysis, 239–241
force field analysis, 245–249, 261, 261f force field analysis, 245–249
diagram for, 247f tree diagrams, 241–243
example of, 248–249 improvement teams, 254–255
procedure for using, 247 composing, 98
template, 286 example of, 100
requirements for selecting members to,
G 99–100
gallery method, for brainwriting, 160 roles of, 98–99
gap analysis, 232–233 indicators, for measurement
Green Belts, 207 financial vs. nonfinancial, 71–72
Green Carpet Seed, case study of, hard vs. soft, 70–71
263–273 leading vs. lagging, 72–74
internal benchmarking, 223
H competitive, 223
hard indicators, 70–71, 70t generic, 223
histograms, 136–142, 136f performance, 224
constructing, 136–138 process, 224
patterns indicating problems in, International Benchmarking Clearinghouse
139–142 (IBC), 35
template, 283 interviews, 115–117
294  Index

is–is not analysis, 148–151 P

example of, 150–151 paired comparisons, 165–166
steps for undertaking, 14 example of, 166
template, 283 procedure for, 165–166
template, 285
Pareto, Vilfredo, 125
Kano model, 12–14, 13f, 16 Pareto charts, 125–127, 126f
example of, 127
steps for using, 126–127
lagging indicators, 72–74 template, 281
leading indicators, 72–74 p charts, for attribute data, 188
L-shaped matrix diagrams, 146, 146f constructing, 193–194
PDPC. See process decision program chart
mailboxes, for feedback, 9 performance benchmarking, 224
Malcolm Baldrige National Quality Award, 204 performance matrix, 254f
Master Black Belts, 206–207 defined, 81
matrix diagrams, 145–148, 258f example of, 82–83
C-shaped, 146 quadrants of, 81, 82f
example of, 148 template, 277
L-shaped, 146, 146f using, 82
relations symbols for, 147, 147f performance measurement
roof-shaped, 146, 146f in business process improvement, 65–67
steps for using, 147 implementing, 67–68, 69f
T-shaped, 146, 146f indicators for, 70–74
X-shaped, 146, 146f financial vs. nonfinancial, 71–72, 72t
Y-shaped, 146, 146f hard vs. soft, 70–71, 70t
matrix organizations, 98 leading vs. lagging, 72–74
modeling, of business processes, 38–43 PERT (Program Evaluation and Review
approaches for Technique), 242
cross-functional flowcharts, 51–54 planning phase, of BPR, 210–211
flowcharts divided into process areas, 54–55 Porter’s value chain, 35
flowcharts with statistics, 58–62, 61f primary processes, 35–36, 35f
relationship mapping, 43–46 problem concentration diagrams, 119–122
several-leveled flowcharts, 55–58 example of, 120–122
traditional flowcharts, 46–51 steps for using, 120
Motorola, 204 process areas, flowcharts divided into, 54–55
process benchmarking, 224
process capability, 200–202
nominal group technique (NGT), 163–165 process cycle times, reducing, 172–174
example of, 164–165 process decision program chart (PDPC),
procedure for, 163–164 243–245, 244f
non-bottleneck resources, 151 approach for using, 243–244
nonfinancial indicators, 71–72 case study of, 271, 272f
non-value-adding activities (NVA), 170–172 example of, 245
normal distribution curves, 184, 185f process departments, 97
normalizing, defined, 232 processes. See also business processes
np charts, for attribute data, 188 development, 36
constructing, 194–195 primary, 35–36, 35f
support, 35, 35f
process ownership, 95
optimized production technology (OPT)
theory, 151 Q
organizational value-adding activities (OVA), QFD. See quality function deployment
170–172 (QFD)
organizations qualitative relations diagrams, 142–143,
stimulating improvement culture of, 103–105 142f
Index  295

quality circles, 9, 100–102 Six Sigma, 204–207

defined, 101 DMAIC process for, 204–206
objectives of, 101 organizational structure for, 206–207
organization of, 101, 102f six thinking hats, 161–163
shortcomings of, in the West, 101–102 case study of, 268–269
quality function deployment (QFD), 85–92, colors for, 162
179–182 example of, 163
example of, 90–92, 180–181 procedure for, 162
process for conducting, 86–88 skills, for business process improvement, 93–94
structure of, 86, 86f soft indicators, 70–71, 70t
template, 279 SPC. See statistical process control (SPC)
quantitative relations diagrams, 143–145, 144f spider charts, 253f
questionnaires, 115–117, 252 constructing, 79–80
defined, 78–79
R example of, 79f, 80–81
random sampling, 114 template, 277
R charts. See X bar and R charts, constructing sporadic variation, 183
real value-adding activities (RVA), 170–172 stakeholder analysis, 10
redundancy, eliminating, 169 approach for, 11
reengineering phase, of BPR, 211–215 examples of, 14, 15–17
clean sheet approach, 214–215 stakeholders
systematic reengineering approach, defined, 10–11
212–214 expectations for improvement of, 9–17
relations diagrams, 142–145 identifying, 11
example of, 143, 145 Kano model of, 12–14, 13f
qualitative, 142–143, 142f model of, 12f
quantitative, 143–145, 144f models for classifying, 11–12, 12f
template, 283 organizational expectations of, 12–13
relationship mapping, 43–45, 44f requirements of, 14
example of, 45–46 strategies of, 14–15, 15f
template, 280 standard deviation, 183–184
relations symbols, for matrix diagrams, 147, statistical process control (SPC). See also
147f control charts
resources basic statistics for, 183–186
bottleneck, 151 example of, 202–203
non-bottleneck, 151 process capability and, 200–202
roof-shaped matrix diagrams, 146, 146f variation and, 183
S flowcharts with, 58–62, 61f
SADT. See Structured Analysis and Design principles of, 183–186
Technique (SADT) variation, 183
sample size, deciding on, 114 strategic benchmarking, 225
Sample University, case study of, 251–262 strategies. See also strategy reviews,
sampling, 113–115 organizational
cluster, 114 emergent, 18
example of, 115 organizational, reviewing, 17–18
issues for, 114 strategy maps, 22–24
random, 114 example of, 24–25
stratified, 114 strategy reviews, organizational, 17–18
systematic, 114 tools for
scatter charts, 134–136, 134f competitive force analysis, 20–22
example of, 135–136 strategy maps, 22–25
steps for using, 135 SWOT analysis, 18–20
template, 283 stratified sampling, 114
several-leveled flowcharts, 55–58 streamlining, 167–176
creating, 56–57 eliminating bureaucracy, 168–169
example of, 58 eliminating redundancy, 169
296  Index

example of, 175–176 for improvements

reducing process cycle times, 172–174 benchmarking, 221–236
value-added analysis, 169–172 business process reengineering,
Structured Analysis and Design Technique 207–221
(SADT), 49 idealizing, 177–178
structured brainstorming, 158 quality function deployment, 179–182
support processes, 35, 35f Six Sigma, 204–207
surveys, 115–117 statistical process control/control charts,
example of, 116–117 182–203
procedure for, 116 streamlining, 167–176
swim lanes, 51 need for, 107–108
SWOT (strengths, weaknesses, opportunities, TOPP (Productivity Program of the
and threats) analysis, 18–19 Technology Industry) program, 35
example of, 19–20 traditional flowcharts, 46–51
template, 276 training, for improvement projects, 94
systematic reengineering, 209 transformation phase, of BPR, 210,
systematic sampling, 114 215–216
tree diagrams, 241–243, 242f
T creating, 242
tampering, 185 example of, 243
tools trend analysis
for analyzing performance shortcoming defined, 76
bottleneck analysis, 151–155 example of, 77–78
cause-and-effect charts, 128–131 procedures for, 76–77
critical incident, 123–125 template, 276
five whys analysis, 132–133 T-shaped matrix diagrams, 146, 146f
histograms, 136–142
is–is not analysis, 148–151 U
matrix diagrams, 145–148 u charts, for attribute data, 188
Pareto charts, 125–127 constructing, 195–196
relations diagrams, 142–145 unstructured brainstorming, 158
scatter charts, 134–136
for business process improvement model, 111t V
for collecting data about performance value-added analysis, 169–172
shortcoming variation
check sheets, 117–119 chronic, 183, 185
problem concentration diagrams, 119–122 sporadic, 183
sampling, 113–115
surveys, 115–117 W
for generating/selecting ideas
Welch, Jack, 29
brainstorming, 157–159
brainwriting, 159–160
Crawford slip method, 159–161 X
nominal group technique, 163–165 x and R charts, constructing, 191–193
paired comparisons, 165–166 X bar and R charts, constructing, 189–191,
six thinking hats, 161–163 191f
for implementing improvements X-shaped matrix diagrams, 146, 146f
ADT analysis, 240–241
force field analysis, 245–249 Y
tree diagrams, 241–243 Y-shaped matrix diagrams, 146, 146f

Related Interests