Linking Customer

and Employee
Satisfaction to the
Bottom Line
Also Available from ASQ Quality Press:

Analysis of Customer Satisfaction Data

Derek R. Allen and Tanniru R. Rao

Customer Centered Six Sigma: Linking Customers, Process Improvement,

and Financial Results
Earl Naumann and Steven H. Hoisington

Customer Satisfaction Measurement Simplified: A Step-by-Step Guide

for ISO 9001:2000 Certification
Terry G. Vavra

Improving Your Measurement of Customer Satisfaction: A Guide

to Creating, Conducting, Analyzing, and Reporting Customer
Satisfaction Measurement Programs
Terry G. Vavra

Measuring Customer Satisfaction: Survey Design, Use, and Statistical

Analysis Methods, Second Edition
Bob E. Hayes

The Trust Imperative: Performance Improvement through

Productive Relationships
Stephen Hacker and Marsha Willard

Customer Satisfaction Measurement and Management

Earl Naumann and Kathleen Giel

Performance Measurement Explained: Designing and Implementing

Your State-of-the-Art System
Bjørn Andersen and Tom Fagerhaug

Value Leadership: Winning Competitive Advantage in the Information Age

Michael C. Harris

To request a complimentary catalog of ASQ Quality Press publications,

call 800-248-1946, or visit our Web site at http://qualitypress.asq.org .
Linking Customer
and Employee
Satisfaction to the
Bottom Line
A Comprehensive Guide to
Establishing the Impact of Customer
and Employee Satisfaction on
Critical Business Outcomes

Derek R. Allen and Morris Wilburn

ASQ Quality Press

Milwaukee, Wisconsin
Linking Customer and Employee Satisfaction to the Bottom Line
Derek R. Allen and Morris Wilburn
Library of Congress Cataloging-in-Publication Data
Allen, Derek R., 1959–
Linking customer and employee satisfaction to the bottom line: a
comprehensive guide to establishing the impact of customer and
employee satisfaction on critical business outcomes / Derek R. Allen
and Morris Wilburn.
p. cm.
Includes bibliographical references and index.
ISBN 0-87389-501-0
1. Consumer satisfaction. 2. Job satisfaction. 3. Industrial management.
4. Success in business. I. Wilburn, Morris, 1953– II. Title.
HF5415.335 .A433 2002
658.3'1422—dc21 2002001314
© 2002 by ASQ
All rights reserved. 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.
10 9 8 7 6 5 4 3 2 1
ISBN 0-87389-501-0
Acquisitions Editor: Annemieke Koudstaal
Project Editor: Craig S. Powell
Production Administrator: Gretchen Trautman
Special Marketing Representative: David Luth
ASQ Mission: The American Society for Quality advances individual,
organizational, and community excellence worldwide through learning,
quality improvement, and knowledge exchange.
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 www.asq.org or http://qualitypress.asq.org .
Printed in the United States of America
Printed on acid-free paper

Quality Press
600 N. Plankinton Avenue
Milwaukee, Wisconsin 53203
Call toll free 800-248-1946
Fax 414-272-1734
www.asq.org
http://qualitypress.asq.org
http://standardsgroup.asq.org
E-mail: authors@asq.org
 Table of Contents

List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . xvii
Chapter 1 Why Customer Satisfaction? . . . . . . . . . . . . 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Deriving Attribute Importance . . . . . . . . . . . . . . . . . 5
Measurement Issues and Customer Satisfaction . . . . . . . . 6
Alternatives to Customer Satisfaction . . . . . . . . . . . . . 9
Customer Value versus Customer Satisfaction . . . . . . . . . 10
Customer Loyalty versus Customer Satisfaction . . . . . . . 11
The Future of Customer Satisfaction Research . . . . . . . . 15
Chapter 2 Employee Satisfaction and Related Phenomena . . 17
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Potential Effects of Employee Satisfaction . . . . . . . . . . 18
Some Additional Comments on Previous Work . . . . . . . . 18
Dimensions of Satisfaction . . . . . . . . . . . . . . . . . . . 20
Overall Employee Satisfaction . . . . . . . . . . . . . . . . . 23
Commitment . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Market Orientation and Service Climate . . . . . . . . . . . . 26
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

v
vi Table of Contents

Chapter 3 The Six Sigma Approach . . . . . . . . . . . . . . 33

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
The Six Sigma Process . . . . . . . . . . . . . . . . . . . . . 37
Six Sigma Tools . . . . . . . . . . . . . . . . . . . . . . . . 39
Quality Function Deployment . . . . . . . . . . . . . . . . . 41
Linking Six Sigma to Financial Performance . . . . . . . . . 46
Chapter 4 The Customer Satisfaction–Profit Link . . . . . . 51
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Linking Satisfaction to the Bottom Line . . . . . . . . . . . . 54
An Intervening Variable: Customer Retention . . . . . . . . . 55
The Impact of Customer Satisfaction in a Healthcare Setting . . 57
Customer Satisfaction, Share, and Profitability . . . . . . . . 59
A “Return on Quality” Model . . . . . . . . . . . . . . . . . 61
The ACSI Study . . . . . . . . . . . . . . . . . . . . . . . . 64
What Is Loyalty? . . . . . . . . . . . . . . . . . . . . . . . . 67
Chapter 5 Retention: A Behavioral Strategy . . . . . . . . . 71
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
An Introduction to Survival Analysis . . . . . . . . . . . . . 73
The Technical Aspects of Survival Analysis . . . . . . . . . . 74
Case Study Background . . . . . . . . . . . . . . . . . . . . 77
Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Model Development . . . . . . . . . . . . . . . . . . . . . . 79
The Significant Predictor Variables . . . . . . . . . . . . . . 80
Model Validation . . . . . . . . . . . . . . . . . . . . . . . . 80
Model Implementation . . . . . . . . . . . . . . . . . . . . . 81
Intervening to Save Profitable Customers . . . . . . . . . . . 82
Chapter 6 Key Business Outcome Metrics . . . . . . . . . . . 87
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
The Cost of Customer Acquisition . . . . . . . . . . . . . . . 88
Research Design and Measurement Implications . . . . . . . 89
The Role of Customer Retention . . . . . . . . . . . . . . . . 92
Customer Satisfaction and Market Share . . . . . . . . . . . 92
Customer Satisfaction and Profitability . . . . . . . . . . . . 95
Chapter 7 Data Preparation . . . . . . . . . . . . . . . . . . 103
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
The Relationship between Employee Attitudes and
Customer Loyalty . . . . . . . . . . . . . . . . . . . . . . 104
The Relationship between Customer Loyalty and
Business Outcomes . . . . . . . . . . . . . . . . . . . . . 111
Summary Comments . . . . . . . . . . . . . . . . . . . . . . 115
 Table of Contents vii

Chapter 8 Analysis Framework: Bivariate Relationships . . 117

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Bivariate Measures of Association . . . . . . . . . . . . . . . 121
Case Study: Canonical Correlation Analysis . . . . . . . . . 124
Causal Framework . . . . . . . . . . . . . . . . . . . . . . . 130
Chapter 9 Analysis Framework: Regression Models . . . . . 133
Dependence Modeling with Multiple Regression . . . . . . . 133
Alternative Regression Models . . . . . . . . . . . . . . . . 137
Logistic Regression Analysis . . . . . . . . . . . . . . . . . 137
Principal Components Regression . . . . . . . . . . . . . . . 142
Other Regression Models . . . . . . . . . . . . . . . . . . . 150
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Chapter 10 Analysis Framework: Causal Modeling . . . . . 153
Causal Modeling with Path Analysis . . . . . . . . . . . . . . 153
Path Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Assessing Model Fit . . . . . . . . . . . . . . . . . . . . . . 158
Case Study: PitStop Employee Satisfaction, Customer
Satisfaction, and Profitability . . . . . . . . . . . . . . . . 159
Latent Variable Path Analysis . . . . . . . . . . . . . . . . . 165
Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Assessing the Model . . . . . . . . . . . . . . . . . . . . . . 172
Chapter 11 The Future of Linkage Research . . . . . . . . . 175
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Simulation Systems: SatNav . . . . . . . . . . . . . . . . . . 175
Management Decision Tools: Loyalty Rx . . . . . . . . . . . 179
The Global Perspective . . . . . . . . . . . . . . . . . . . . . 184
Refining Our Understanding of Loyalty . . . . . . . . . . . . 189
Intangible Drivers of Customer Loyalty . . . . . . . . . . . . 191
Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . 191
Appendix: Matrix Algebra in Statistics . . . . . . . . . . . . . 195
Matrix and Vector Operations . . . . . . . . . . . . . . . . . 196
Matrix Multiplication . . . . . . . . . . . . . . . . . . . . . 197
Least Squares Regression . . . . . . . . . . . . . . . . . . . 202
Principal Components Analysis . . . . . . . . . . . . . . . . 205
Eigenvector Scaling . . . . . . . . . . . . . . . . . . . . . . 209
Eigenvector Rotation . . . . . . . . . . . . . . . . . . . . . . 211
Glossary of Terms . . . . . . . . . . . . . . . . . . . . . . . . . 213
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
 List of Figures

Figure 1.1 Malcolm Baldrige National Quality Award

evaluation criteria. . . . . . . . . . . . . . . . . . . . . . 2
Figure 1.2 Principal disparities affecting service quality. . . . . . . . 3
Figure 1.3 SERVQUAL scale dimensions. . . . . . . . . . . . . . . . 4
Figure 1.4 Customer value map. . . . . . . . . . . . . . . . . . . . . 10
Figure 1.5 Reasons for automobile repurchase. . . . . . . . . . . . . 12
Figure 1.6 Affective and cognitive dimensions of loyalty. . . . . . . . 12
Figure 1.7 Relationship between loyalty and satisfaction. . . . . . . . 13
Figure 2.1 Elements of the Minnesota Satisfaction Questionnaire. . . 19
Figure 2.2 Summary of employee satisfaction questions. . . . . . . . 23
Figure 2.3 Commitment questions. . . . . . . . . . . . . . . . . . . . 25
Figure 2.4 Items proposed by Narver, Jacobson, and Slater. . . . . . . 27
Figure 2.5 Items proposed by Kohli, Jaworski, and Kumar. . . . . . . 28
Figure 2.6 Items proposed by Deshpandé, Farley, and Webster. . . . . 29
Figure 2.7 Items proposed by Schneider, Parkington, and Buxton. . . 30
Figure 3.1 Six Sigma process variance reduction. . . . . . . . . . . . 34
Figure 3.2 Hard dollar and soft dollar outcomes of
process improvement. . . . . . . . . . . . . . . . . . . . . 35
Figure 3.3 Blue Sky Airlines: relationship between customer
and employee satisfaction and profitability. . . . . . . . . 36
Figure 3.4 The Six Sigma process. . . . . . . . . . . . . . . . . . . . 37
Figure 3.5 Six Sigma control chart. . . . . . . . . . . . . . . . . . . 39
Figure 3.6 Six Sigma sequential reliability test. . . . . . . . . . . . . 40
Figure 3.7 Quality function deployment. . . . . . . . . . . . . . . . . 41
Figure 3.8 The house of quality: Blue Sky Airlines. . . . . . . . . . . 43
Figure 3.9 Process improvement link to profitability. . . . . . . . . . 47
Figure 4.1 Five dimensions of behavioral intention. . . . . . . . . . . 56
Figure 4.2 Hypotheses relating customer satisfaction
to economic returns. . . . . . . . . . . . . . . . . . . . . . 59

ix
x List of Figures

Figure 4.3 The five steps of the ROQ quality

improvement process. . . . . . . . . . . . . . . . . . . . . 63
Figure 4.4 The ACSI model of customer loyalty. . . . . . . . . . . . 66
Figure 4.5 Reasons for repurchase behavior. . . . . . . . . . . . . . . 68
Figure 4.6 The affective and cognitive dimensions of loyalty. . . . . . 69
Figure 4.7 Quality and its effect on customer retention. . . . . . . . . 69
Figure 5.1 Illustration of attrition risk profiles. . . . . . . . . . . . . 74
Figure 5.2 Attrition rates by risk decile. . . . . . . . . . . . . . . . . 81
Figure 5.3 Attrition risk and profitability. . . . . . . . . . . . . . . . 82
Figure 5.4 Relationship between high attrition risk
and profitability. . . . . . . . . . . . . . . . . . . . . . . . 84
Figure 6.1 Plot of time series data relating customer and
employee satisfaction to market share. . . . . . . . . . . . 90
Figure 6.2 Cross-sectional data analysis: bivariate relationships. . . . 91
Figure 6.3 Causal sequence leading to profitability. . . . . . . . . . . 94
Figure 6.4 Profit margin on sales. . . . . . . . . . . . . . . . . . . . 96
Figure 6.5 Calculation of net income based on the capital
maintenance approach. . . . . . . . . . . . . . . . . . . . 97
Figure 6.6 Calculation of net income based on the
transaction approach. . . . . . . . . . . . . . . . . . . . . 97
Figure 6.7 Rate of return on assets. . . . . . . . . . . . . . . . . . . . 98
Figure 6.8 Rate of return on assets: alternate calculation. . . . . . . . 98
Figure 6.9 Rate of return on common stock equity. . . . . . . . . . . 99
Figure 6.10 Earnings per share. . . . . . . . . . . . . . . . . . . . . . 99
Figure 7.1 A common unit of analysis. . . . . . . . . . . . . . . . . . 104
Figure 7.2 Customer satisfaction and employee satisfaction
data structures. . . . . . . . . . . . . . . . . . . . . . . . 105
Figure 7.3 Financial performance as a result of
customer retention. . . . . . . . . . . . . . . . . . . . . . 111
Figure 8.1 Project timeline. . . . . . . . . . . . . . . . . . . . . . . . 118
Figure 8.2 Summary of changes in key profitability indices. . . . . . 119
Figure 8.3 Branch-level time dependence of employee and
customer satisfaction. . . . . . . . . . . . . . . . . . . . . 120
Figure 8.4 Illustration of correlation coefficients. . . . . . . . . . . . 121
Figure 8.5 Collapsing seven data files by sales district. . . . . . . . . 125
Figure 8.6 Illustration of lower triangular matrix for all survey
and financial performance data. . . . . . . . . . . . . . . . 126
Figure 8.7 Summary of canonical correlations among
data sources. . . . . . . . . . . . . . . . . . . . . . . . . . 129
Figure 8.8 Theoretical framework with significant
relationships depicted. . . . . . . . . . . . . . . . . . . . 130
Figure 9.1 Illustration of multiple regression analysis. . . . . . . . . . 134
Figure 9.2 Differences across binary independent variables. . . . . . 140
Figure 9.3 Principal components of bivariate data. . . . . . . . . . . 144
 List of Figures xi

Figure 10.1 Minimal criteria for establishing causality. . . . . . . . . . 154

Figure 10.2 Simple causal relationship. . . . . . . . . . . . . . . . . . 155
Figure 10.3 Multiple-predictor causal model. . . . . . . . . . . . . . . 156
Figure 10.4 Path analysis model with two dependent variables. . . . . 157
Figure 10.5 Quasi-confirmatory model-building approach. . . . . . . . 158
Figure 10.6 PitStop convenience stores employee
satisfaction instrument. . . . . . . . . . . . . . . . . . . . 160
Figure 10.7 PitStop convenience stores customer
satisfaction instrument. . . . . . . . . . . . . . . . . . . . 161
Figure 10.8 PitStop path analysis relating employee and customer
satisfaction and profitability. . . . . . . . . . . . . . . . . 163
Figure 10.9 Final PitStop path analysis model. . . . . . . . . . . . . . 164
Figure 10.10 Seven Peaks National Bank latent variable SEM. . . . . . 167
Figure 10.11 Seven Peaks National Bank customer
satisfaction instrument. . . . . . . . . . . . . . . . . . . . 169
Figure 10.12 Seven Peaks National Bank employee
satisfaction instrument. . . . . . . . . . . . . . . . . . . . 170
Figure 10.13 Seven Peaks National Bank latent variable
SEM results. . . . . . . . . . . . . . . . . . . . . . . . . . 173
Figure 11.1 SatNav simulation introductory screen illustrating
map types. . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Figure 11.2 SatNav output with color-coded output at
regional level. . . . . . . . . . . . . . . . . . . . . . . . . 177
Figure 11.3 SatNav output with three-dimensional
mapping feature. . . . . . . . . . . . . . . . . . . . . . . 178
Figure 11.4 Loyalty Rx . . . . . . . . . . . . . . . . . . . . . . . . . 182
Figure 11.5 Key phases in the linkage research process. . . . . . . . . 192
 Foreword

A
customer, Mr. Smith, buys a new automobile, his first new car pur-
chase. Mr. Smith is just 21 years old. He likes the car’s styling, the
car has fewer breakdowns than his old used autos, and the car gets
good mileage. The dealer, located in the city, provides good service and
takes the time to answer Mr. Smith’s questions. Service people who are
committed to the dealership are friendly and helpful in all of their dealings
with Mr. Smith. He finances the car through the auto finance company. The
finance company gives Mr. Smith the loan at a competitive rate, and he
arranges an automatic monthly payment procedure through his bank. Mr.
Smith sees his car displayed in company advertising and is pleased that the
auto company is innovative in bringing new and safer automobiles to the
public using current technology. Mr. Smith thinks the auto company has
offered superior value on all fronts. He relates his positive experiences with
his car to 10 other people and influences five of them to visit the dealer’s
showroom. His wife buys a new car, a different model, from the same auto
company and through the same dealer.
Sounds simple. This is the story of a loyal customer. The auto company
now has a customer in Mr. Smith who will contribute a significant profit in
his lifetime even though his name does not appear on the company’s bal-
ance sheet. Corporate financial officers do not always see their balance
sheets as the amalgamation of many individual customers with different
levels of satisfaction and loyalty. They should. The story is the same
whether you operate a retail store, run a financial institution, or manufac-
ture business equipment.
The simplicity of the above linkages may be the reason they are often
ignored. Today, unlike 30 years ago when customer satisfaction research

xiii
xiv Foreword

was in its infancy, virtually every company conducts some form of cus-
tomer satisfaction research and employee surveys. Customer satisfaction
studies may range from simple comment cards to well-designed relation-
ship and transactional surveys. Employee surveys range from old-fashioned job
satisfaction surveys to comprehensive employee engagement/commitment
surveys. However, even today, with the current explosion of knowledge in
these areas, companies do not always put their feedback systems in a proper
strategic context. Top management in many companies does not look at
measurement systems from the point of view of customer segmentation
and profitability.
This book by Dr. Derek R. Allen and Mr. Morris Wilburn makes a case
to research practitioners and management to focus on linkages between
customer satisfaction, employee satisfaction, and business profitability. By
bringing together all of the theoretical knowledge and empirical work done
to date on these issues, the authors encourage researchers to design these
types of studies in an integrated fashion for the benefit of the company. One
of the greatest failures in the current lack of integration is contributed by
the silo structure inherent in any business where ownership of different sur-
veys is diffused.
The book provides analytical guidance for researchers as well as con-
ceptual guidance to managers to develop an integrated view of customer
satisfaction and employee satisfaction and their link to business outcomes.
The book is clear evidence of the elevation of customer feedback research
to a professional discipline with significant intellectual rigor. It will serve
as an excellent reference tool to allow loyalty managers to better use cus-
tomer feedback information to a strategic advantage.

Dr. T. R. Rao
CEO/President
Market Probe
 Preface

T
his book is intended for advanced service quality managers and mar-
keting researchers with more than a modest exposure to statistical
data analysis. Our objective is to provide the reader with a funda-
mental understanding of how customer satisfaction and employee attitudes
may be empirically linked to substantive business outcomes such as prof-
itability. Above all, this book is intended for the practitioner and, as such,
relies upon numerous industry examples to illustrate key points.
The scope of this book ranges from a brief treatment of customer and
employee satisfaction measurement to in-depth discussions concerning ana-
lytical approaches to linking customer satisfaction and profitability. We have
avoided as much as possible an overly technical treatment of these subjects.
Instead, a more descriptive, managerial approach was taken in an effort to
facilitate an understanding of how sophisticated analytical techniques can be
applied and, more important, what they can tell us about the data. Nonethe-
less, we cannot avoid some technical discussions since our objective is to
demonstrate how to link customer and employee satisfaction data to key
business measures.
We have attempted to include actual industry examples to illustrate
how techniques were applied, how the results were interpreted, and, finally,
how critical strategic marketing decisions were affected. We have also
addressed some of the most problematic aspects of service quality data and
internal business metrics, including data scaling and scaling variables such
as customer profitability.
This book provides the advanced service quality manager or marketing
researcher embarking on customer satisfaction research with a comprehen-
sive overview of how these data may be related to critical business outcomes.

xv
xvi Preface

Perhaps more important, researchers with mature customer satisfaction mea-

surement systems may use the techniques described in this book to maximize
the value of their existing programs. Indeed, more and more frequently,
researchers managing large customer and employee satisfaction programs are
being required to demonstrate their systems’ value. A more poignant question
has been posed by senior management with increasing frequency: What is the
value of customer and employee satisfaction research?
No technique or methodology can guarantee a strong link between cus-
tomer satisfaction and key business outcomes. This book can, however,
ensure that appropriate scales, variables, and assumptions are used. Further,
we have included a wide variety of case studies that will facilitate an under-
standing of the types of variables one must consider when attempting to link
business metrics to psychometric data. This book makes a concerted effort
to address employee satisfaction and its role in the development of customer
satisfaction and profitability, customer retention, and market share. As such,
we feel it makes a unique contribution to the service quality management
and human resources fields.
The first four chapters of the book introduce customer satisfaction,
employee satisfaction, the Six Sigma approach to quality management, and
the links between satisfaction, profitability, and customer retention. The fifth
chapter introduces a unique application of survival analysis using customer
attrition data. Chapter 6 provides a review of key business performance vari-
ables, and chapter 7 describes how data should be prepared for the type of
analysis described in chapters 8 through 10. Chapter 8 introduces bivariate
approaches to establishing empirical linkages between customer satisfaction,
employee satisfaction, and financial performance measures. Chapter 9 intro-
duces a series of multivariate dependence models beginning with multiple
regression analysis and includes discussions on the use of logistic regression,
principal components regression, and nonlinear regression. Chapter 10 intro-
duces causal modeling with both manifest and latent variables.
The final chapter of this book describes a number of cutting-edge
applications that have been used to link both customer and employee satis-
faction to key business outcomes. Among these are decision support soft-
ware and simulation packages that permit managers to pose “what if?”
questions using their data. The use of this type of tool will facilitate the
ability of managers and researchers alike to test the efficacy of models link-
ing customer satisfaction to corporate profitability.
Theoretical Foundations: Chapters 1–4
Applications: Chapters 5–10
Summary: Chapter 11
 Acknowledgments

T
his book would not have been possible without the theoretical and
technical foundations laid by many applied and academic researchers
over the past 25 years. The completion of this work was greatly facil-
itated by the administrative skills of Sandy Cummings, who produced all of
the graphics and spent many hours ensuring that editorial changes were
implemented. Maya Hughes provided insightful editorial input and project
management during the development of this book, and to her we are
extremely grateful.
A number of Market Probe project management staff helped proof
this book. Among them were Amy Bastic, Jill Carnick, Deana Gillespie,
Chandreyee Mittra, Megan Murphy, Heather Nagel, Bob Peterson, Lori
Richards, Debashis Sengupta, Ron Sincere, and Barry Tochterman. We
would also like to thank Karen Ethington for reviewing the accounting
practices associated with the business performance metrics described in
chapter 6. Thanks also go to the anonymous ASQ reviewers whose insights
and valuable comments enhanced this volume’s contribution to the cus-
tomer satisfaction research literature. Special thanks to Leayn and Paul
Tabili at New Paradigm for their outstanding prepress work. Finally, we
would like to express our gratitude to ASQ Quality Press editorial staff
members Annemieke Koudstaal and Craig Powell.

xvii
 1
Why Customer Satisfaction?

INTRODUCTION
This book focuses on the relationships between customer satisfaction,
employee satisfaction, and tangible business outcomes such as market
share, revenue, and profitability. How customer satisfaction research devel-
oped and the role it plays in business are described in this chapter. As a con-
struct, customer satisfaction enjoys an enviable niche in terms of its role in
corporate incentive systems and other reward programs. The conventional
wisdom is that since it is used so extensively in gauging performance, cus-
tomer satisfaction must be indicative of business success . . . or the lack
thereof. But this logic is somehow circular and fails to present scientific
evidence for such a relationship.
In this book, we provide a synthesis of applied and academic research
with the aim of empirically demonstrating the relationship between cus-
tomer and employee satisfaction and substantive outcomes such as corporate
profitability. In order to set the stage, we first treat customer satisfaction
(chapter 1) and employee satisfaction (chapter 2). By establishing a sound
understanding of these constructs and how they are measured, we can more
effectively demonstrate how they affect key business outcomes such as prof-
itability and market share.
That customer satisfaction is inexorably related to the quality revolution
in the United States is practically incontrovertible. The formalization of cus-
tomer satisfaction as a quality component in national competitions such as
the Malcolm Baldrige National Quality Award has further validated the cus-
tomer satisfaction research agenda. As shown in Figure 1.1, customer satis-
faction constitutes a significant part of the Baldrige Award criteria.

1
2 Chapter One

Information
and
analysis Strategic
Leadership (75 pts.) quality
(95 pts.) planning
(60 pts.)

Human
resources
management
(150 pts.)
Customer focus
and satisfaction
(300 pts.) Management of
process quality
(140 pts.)

Quality and
operational results
(180 pts.)

Figure 1.1 Malcolm Baldrige National Quality Award evaluation criteria.

This chapter provides a succinct history of customer satisfaction research

and reviews more contemporary attempts to link satisfaction to retention and,
ultimately, corporate profitability.
The first forays into the measurement of customer satisfaction occurred
in the early 1980s. These typically involved assessing the drivers of satis-
faction. Early works by Oliver (1980), Churchill and Surprenant (1982),
and Bearden and Teel (1983) tended to focus on the operationalization of
customer satisfaction and its antecedents. By the mid-1980s, the focus of
both applied and academic research had shifted to construct refinement and
the implementation of strategies designed to optimize customer satisfaction,
according to Zeithaml, Berry, and Parasuraman (1996, 31).
Rigorous scientific inquiry and the development of a general service
quality theory can be attributed to Parasuraman, Berry, and Zeithaml (1985).
Their discussion of customer satisfaction, service quality, and customer
expectations represents one of the first attempts to operationalize satisfac-
tion in a theoretical context. Parasuraman, Berry, and Zeithaml proposed
that the ratio of perceived performance to customer expectations was key to
 Why Customer Satisfaction? 3

maintaining satisfied customers. The disparity between performance and

expectations was presumed to take five forms, as shown in Figure 1.2.
Several years later, Parasuraman, Berry, and Zeithaml (1988) pub-
lished a second, related discussion that focused more specifically on the
psychometric aspects of service quality. Their multi-item SERVQUAL
scale is considered to be one of the first attempts to operationalize the cus-
tomer satisfaction construct. The SERVQUAL scale focused on the perfor-
mance component of the service quality model in which quality was
defined as the disparity between expectations and performance. The battery
of items used in the SERVQUAL multi-item scale is still used today as a
foundation upon which instrument development is often based. The pri-
mary areas considered in the scale are depicted in Figure 1.3.
The five primary components of the model each represent a distinct
aspect of the customer relationship. For example, the tangibles dimension
includes the appearance of facilities, equipment, personnel, and communi-
cation materials. Similarly, the reliability dimension addresses the service
provider’s ability to perform dependably and accurately. The responsiveness
component of the SERVQUAL architecture involves the service provider’s

GAP

Management perceptions
Consumer expectations
of consumer expectation

Management perceptions Intended level of

of consumer expectations service quality

Intended level of Service quality experienced

service quality by consumers

Service quality experienced What is communicated about

by consumers the service to consumers

Consumer perceptions of
Consumer expectations
service delivery

Figure 1.2 Principal disparities affecting service quality.

4 Chapter One

Tangibles

Empathy Reliability

Assurance Responsiveness

Figure 1.3 SERVQUAL scale dimensions.

willingness to help customers and provide prompt service. Assurance is

identified as the knowledge base and courtesy of employees and their abil-
ity to instill trust and confidence. Finally, empathy relates to the service
organization’s caring and individualized service.
Throughout the 1980s, both applied and academic researchers focused
on these (and other) issues and their effects on overall customer satisfaction
according to Zeithaml, Berry, and Parasuraman (1996, 31). That is, the pri-
mary research question involved which of the five areas was most impor-
tant vis-à-vis customer expectations. Much of the earliest applied work
involving the derivation of attribute importance involved stated measures.
It was not uncommon to encounter surveys in which for every item both
importance and performance measures were sought. The gap between these
two measures was considered instrumental in resource allocation. Large
gaps demanded the most attention. It is important to note, however, that
Parasuraman, Berry, and Zeithaml (1988) employed regression analysis to
assess the effect of each dimension relative to a dependent measure in their
introduction of the SERVQUAL model. The use of regression analysis and
other dependency models to derive the importance of attributes relative to
an outcome measure is now considered de rigueur and is described in detail
in the following pages.
 Why Customer Satisfaction? 5

DERIVING ATTRIBUTE IMPORTANCE

Derived importance models represent a cornerstone of customer satisfac-
tion research. The derivation of attribute importance currently represents
the norm in the customer satisfaction research industry. Today, few if any
consultants advocate the stated importance framework in which customers
are asked to rate (or rank) the importance of service or product quality
issues. Its shortcomings have been illustrated with the frequently cited air-
line safety example in which stated and derived importance metrics lead to
disparate conclusions. In the airline safety example, we contrast the impor-
tance of five issues:
• Food quality
• Cleanliness
• Safety
• Attendant courtesy
• Comfort
When asked which of these issues is most important to them, most con-
sumers will indicate that safety is their primary concern. After all, who would
respond that food quality is more important than personal safety? As most
who are familiar with this example will recall, the derived importance frame-
work employs multiple regression to mathematically determine the impact of
each issue on a single outcome variable: overall satisfaction with the airline.
There is some confusion surrounding the implication of the word
importance in the dependence model, or multiple regression, context. Take,
for example, the reaction of product managers at a world-class manufac-
turing company when told that product quality was not important. Their
initial inclination was to dismiss the results and walk out! The simple fact
is that the use of the term importance tends to cloud the true meaning of
what we are trying to convey. In the regression context, importance implies
the magnitude of impact on an outcome variable such as overall satisfac-
tion or repurchase propensity. Once they understood this, the managers and
engineers at the manufacturing company described above were actually
quite pleased. Their products were—and still are—considered best in class
and enjoy stellar customer satisfaction scores.
So, exactly what are we conveying when we say something is important
in a dependence model context? It should be clear that derived importance
results provide managers with road maps for maximizing an outcome vari-
able such as overall satisfaction. In order to achieve the greatest increase in
6 Chapter One

Table 1.1 Hypothetical key drivers.

Quarter 1 Quarter 4
Service/Product Satisfaction Key Driver Status Key Driver Status
Teller courtesy No Yes
Competitive rates No No
Waiting time for teller No Yes
Statement accuracy No No
Parking space Yes No
Branch proximity No No

overall satisfaction (or any other outcome variable), we focus on the predic-
tor variables that have the greatest effect on it. More accurately, we isolate
the predictor variables that covary most strongly with the outcome variable.
It is this covariation that is at the heart of derived importance models.
Derived importance models are probably erroneously labeled. A more
appropriate name might be marginal resource allocation models. It is
important that we consider only marginal resources since reallocating
existing resources based upon a derived importance model will not neces-
sarily yield the desired results. Consider the extreme example of a retail
bank that blindly reallocates existing resources based upon its key driver
analysis. Assume the simplified set of key drivers shown in Table 1.1.
If bank management erroneously decides to reallocate resources based
upon the set of first quarter key drivers, it may choose to reduce the number
(and quality) of the teller staff in order to free funds for the acquisition of
adjacent land for parking. This reallocation of resources yields a completely
different set of drivers in the fourth quarter. The new drivers reflect the
effects of management’s decision to reallocate existing rather than marginal
dollars. Since teller staffing was cut back, waiting time increased and, as a
result, customers became less satisfied with this aspect of their relationship
with the bank. In fact, it appears from these data that the dissatisfaction was
sufficiently acute to affect overall satisfaction. Such a turn of events under-
scores the dynamic nature of key drivers. It also illustrates why key driver
analysis should really be considered as a marginal resource allocation tool.

MEASUREMENT ISSUES AND

CUSTOMER SATISFACTION
When we consider the linkage between customer satisfaction and profitability,
how we measure customer satisfaction emerges as a significant issue. Given
our desire to establish a relationship between customer satisfaction and key
 Why Customer Satisfaction? 7

business outcomes, the measurement of the former is a salient issue. If we fail

to adequately measure satisfaction, our chances of establishing an empirical
link with, for example, profitability or market share may be lessened.
Over the past 20 years or so, researchers have experimented with a
wide variety of approaches to measuring customer satisfaction. Among
these are gap measures, letter grade scales, expectation scales, and Likert-
type numeric scales with anchored endpoints. Each approach has advocates
who defend it tenaciously. It seems clear that the Likert-type numeric scale
with endpoint anchors (“very satisfied” to “very dissatisfied”) is favored by
both academicians and researchers working in applied settings.
The multipoint scale yields more data variability than, for example,
binary scales that elicit “yes” or “no” responses (Hayes 1998, 70–71). For
organizations performing at stellar levels (for example, top-two-box scores
of 90% or greater on a 5-point scale), it may be worth considering either a
7-point or 10-point scale. There are at least two reasons for this. Consider
a hypothetical organization with 75 national sales districts. The company
sells and services a premium product that is universally considered “best in
class.” Assume that this company has a mature customer satisfaction mea-
surement (CSM) and tracking program that relies upon a 5-point scale.
Feedback is provided on a quarterly basis at the sales district level. If most
sales districts receive quarterly scores of better than 90%, differentiating
performance levels among them will be difficult at best. It is the ability to
discriminate between top performers and poor performers that adds tremen-
dous utility to CSM programs that are linked to compensation and bonus
plans. Clearly, if the reporting metric produces a distribution of scores so
narrow that statistically differentiating performance levels is impossible,
then the program will have only modest utility.
In reality, choice of measurement scales is frequently determined by uni-
variate reporting requirements. A company that routinely receives 90% top-
two-box scores on a 5-point scale will likely only enjoy about an 85%
top-two-box score on a 7-point scale. On a 10-point scale, the same com-
pany would expect a score of only about 75%. In addition, the distribution
of districts around the average top-two-box score will be more diffuse. The
wider distribution of scores around the mean gives us more discriminating
power. The enhanced discriminating power enables us more reliably to iso-
late poor performers.
The second reason a 7-point or 10-point scale is preferred involves
covariance. In general, it is easier to establish covariance between two vari-
ables with greater dispersion, or variance, around their means. It is this
covariance that is critical to establishing strong multivariate dependence
models used in key driver assessments. It will also prove to be critical in
our efforts to establish a strong link between customer satisfaction and key
8 Chapter One

business outcomes such as profitability and market share. Thus, from a

model development perspective, the 10-point scale is preferred. Sophisti-
cated statistical software such as Jöreskog and Sörbom’s LISREL 8 struc-
tural equation modeling package, for example, assumes that data are
ordinal if five or fewer scale levels are present.
In an empirical review of several scale types, Wittink and Bayer (1994)
concluded that a 10-point dependent measure was preferred. The authors
also maintained that a 2-point experiential predictor variable scale should
be adopted. This scale type solicits a yes or no response to statements
involving specific service or product experiences. For example, “Did the
service person provide you with an accurate invoice?” can only be
answered with a yes or no. While the 10-point dependent measure is well
accepted in both academic and industry research settings, the 2-point pre-
dictor variable scale has been met with some trepidation. One reason is that
models based upon this measurement approach tend to have lower predic-
tive (R2) capacities.
Not surprisingly, the configuration that Wittink and Bayer advocated
has not been widely adopted. Nonetheless, their comparison of the 5-point
and 10-point dependent variable scales is valuable. Wittink and Bayer
relied upon a total of 13 criteria. These included issues involving the
respondent, such as simplicity and understandability, and statistical issues
such as response bias, power, and sensitivity. Response bias involves the
extent to which target respondents are excluded due to inability or refusal
to participate. Power involves the extent to which the scale can detect
changes over time (or differences among groups). Finally, sensitivity was
defined as the degree of room for improving customer satisfaction. Given
this set of criteria, Wittink and Bayer concluded that the 10-point endpoint-
anchored measurement scale was preferred to the 5-point alternative.
We must concede that skewness remains a problem even with the 10-
point scale; this problem may never be fully circumvented in customer sat-
isfaction research. Nonetheless, there are ways to lessen the impact of
skewed data. One simple approach is to provide extreme endpoints. Rather
than a scale ranging from “satisfied” to “dissatisfied,” we can make the
anchor labels more difficult to achieve. For example, labeling the endpoints
as “delighted” and “very disappointed” has been used with some success.
Others have suggested a traditional 5-point scale ranging from “very dis-
satisfied” to “very satisfied” with the addition of a sixth point labeled sep-
arately as “delighted.” This approach also seems to yield slightly less
skewed distributions but poses the danger of losing interval-level proper-
ties. That is, it is a 6-point scale with three labels; the difference between a
“5” rating and a “6” is ambiguous. Is the distance from “very satisfied” to
 Why Customer Satisfaction? 9

“delighted” the same as the distance between a rating of a “4” and a “5”?
An additional option that is (regrettably) infrequently employed in applied
customer satisfaction research involves data transformation. While a wide
variety of transformations are available, our interest focuses on the distrib-
ution that is highly skewed.
All things considered, most statisticians working with customer satis-
faction research—particularly those involved in model development—
advocate scales with more points. This is because of the increased variance
and better chance of demonstrating covariance among key variables. The
final scaling decision is frequently dictated by past measurements. A com-
pany with a stable tracking system that has used a 5-point scale for many
years should consider carefully the implications of losing continuity in its
search for more variance. Clearly, the two must be weighed judiciously; a
company with stellar customer satisfaction ratings might be inclined to pur-
sue more scale points in an effort to enhance its ability to differentiate top
performers from poor performers. On the other hand, there would be few
compelling reasons to explore an expanded scale for a company with more
modest scores and little interest in multivariate statistical models.
There are no unequivocal answers when it comes to the number of
points for a scale. Each case is characterized by the current distribution of
scores, the need for tracking continuity, benchmarking, and, in many cases,
personal preference. It should be clear that fewer than 5 points veers dan-
gerously from the interval properties assumed by most of the multivariate
statistical procedures described in this book. Similarly, more than 10 points
is very uncommon and may be unmanageable for the respondent.

ALTERNATIVES TO
CUSTOMER SATISFACTION
It should come as no surprise that in an effort to understand consumer
behaviors some researchers have advocated alternatives to customer satis-
faction. Ostensibly, these new constructs are better predictors of critical
outcomes such as repeat purchases and, inevitably, corporate profitability.
The two principal alternatives to customer satisfaction include customer
value perceptions and customer loyalty. The former measure involves con-
sumer evaluations of product (or service) price and quality. The ratio of
these two perceptions (price and quality) yields a value assessment. Cus-
tomer loyalty measurement involves more emotive measures and is consid-
ered to be dependent upon customer satisfaction. Each is discussed in more
detail in the following pages.
10 Chapter One

CUSTOMER VALUE VERSUS

CUSTOMER SATISFACTION
Customer value perceptions have been offered as a viable alternative to cus-
tomer satisfaction. Gayle (1994, 8–11) proposed that consumers’ value per-
ceptions represented a natural extension of customer satisfaction. The
reason companies win or lose customers, Gayle argued, lies in those cus-
tomers’ value perceptions.
Ostensibly, the two primary advantages of customer value management
include the ability to link customer value perceptions to market share and the
measurement context. With respect to the latter, customer value perceptions
are measured in the market as a whole. That is, consumers assess a com-
pany’s products or services relative to those of other competitors. This mea-
surement framework naturally permits comparisons among key competitors.
Gayle operationalized value as the ratio between price and quality, as
shown in Figure 1.4. The two axes of the figure represent perceived price
and perceived quality. The center diagonal line represents the fair market
value. Being above the line is indicative of worse value. Conversely, a com-
pany whose product or service is encountered below the line represents a
better value.
In the example shown in Figure 1.4, ABC Microchips represents a worse
value. Not only is it perceived to have lower-quality products, it is also

Higher
Worse Value
Perceived Price

ABC Microchips

e
e lin XYZ Microchips
alu
ir v
Fa

Lower Better Value

Worse Better
Perceived Quality

Figure 1.4 Customer value map.

 Why Customer Satisfaction? 11

assumed to be considerably more expensive. As a result, it is positioned

above the fair value line and presumably would enjoy a much lower share
than its competitor, XYZ Microchips. It should be clear that this type of
analysis solicits evaluations of all competitors from respondents. While this
is certainly appropriate in an industrial setting, wherein a purchasing agent is
familiar with various suppliers, it is not clear how well value analysis works
with consumers who have little knowledge of competitors’ prices and qual-
ity. It could be argued that only perceived price and quality are required; how-
ever, this may change as consumers engage in information gathering. Still,
customer value management has a significant following and appears to rep-
resent both a challenge and complement to customer satisfaction research.

CUSTOMER LOYALTY VERSUS

CUSTOMER SATISFACTION
As the customer satisfaction research environment has matured, a variety of
rival and complementary constructs have emerged. Among these is customer
loyalty. There appears to be no consensus among researchers concerning the
operationalization of loyalty. A large part of loyalty may involve an attitudinal
state. It is not, for example, exclusively a behavior. Dick and Basu (1994) sug-
gested that customer loyalty is a combination of behaviors and attitudes. In
short, the authors maintained that loyal customers are those who had both a
favorable attitude and repeated purchases. Such an approach allows for “spu-
rious loyalty,” which arises when purchases are made repeatedly by a cus-
tomer who does not have a particularly favorable attitude toward the brand.
When considered exclusively as an attitudinal state, loyalty certainly
has many desirable behavioral outcomes, such as higher repurchase and
lower attrition (customer defection) rates. Repurchase activities or switch-
ing behaviors may be manifestations of loyalty. The behaviors are not loy-
alty, but they may be an outcome of loyalty. An aversion to switching and
high repurchase rates do not necessarily mean that a customer base is loyal
per se. Clearly, many other factors can cause these desirable behaviors,
including proximity and price, as shown in Figure 1.5.
If loyalty is not a behavior, nor the same as satisfaction, what exactly
is it? Recent developments in the operationalization of loyalty have yielded
attitude scales designed specifically for the industrial buyer and consumer
markets. Loyalty may have two primary dimensions: affective and rational,
as shown in Figure 1.6. The affective dimension of loyalty has emotional
underpinnings and involves human interaction. In contrast, the cognitive
dimension of loyalty includes evaluations of the business relationship that
involve price, proximity, timeliness, and so on.
12 Chapter One

Closest dealership

Cheapest brand

Repurchase

Loyalty to dealership

Only dealer with

model available

Figure 1.5 Reasons for automobile repurchase.

Affective loyalty

Customer loyalty

Cognitive loyalty

Figure 1.6 Affective and cognitive dimensions of loyalty.

Very few applied customer satisfaction programs currently accommo-

date a fully-developed operationalization of loyalty. Fewer still attempt to
differentiate between the affective and cognitive components of loyalty.
Table 1.2 presents a summary of several questionnaire items we have used
in an effort to capture both dimensions of loyalty. These were developed
based upon qualitative and quantitative experiences across a diverse range
of industries in both consumer and business-to-business settings.
In most cases, customer satisfaction is a necessary, but not suffi-
cient, condition for loyalty. We believe that satisfaction and loyalty are
 Why Customer Satisfaction? 13

two different constructs. Satisfaction is directed specifically at product

or service attributes and may be a relatively more dynamic measure. In
contrast, loyalty is a broader, more static attitude toward a company in
general. As described earlier, it may subsume both rational and emo-
tional elements and is clearly affected by satisfaction.
As shown in Figure 1.7, a causal chain is suggested. First, brand image,
service quality, product quality, and price are four critical predictor variables.

Table 1.2 Questionnaire items relating to loyalty.

Cognitive Component Items
“For our firm, staying in the relationship with XYZ Co. is a matter of necessity.”
“Discontinuing our relationship with XYZ Co. would result in a loss of business
for our company.”
“Our company stays in its relationship with XYZ Co. because of the rewards and
benefits it brings us.”
Affective Component Items
“The people at XYZ Co. are just like us.”
“XYZ Co. staff have our best interests at heart.”
“A lot of the employees at XYZ Co. are my friends.”

Brand image

Affective
loyalty
Service
quality issues

Customer Customer
satisfaction retention

Product Cognitive
quality issues loyalty

Value

Price
perception

Critical predictor Intermediate Ultimate dependent

variables variables variable

Figure 1.7 Relationship between loyalty and satisfaction.

14 Chapter One

The first set of intermediate variables involves value and customer satis-
faction. Value is shown to be a function of both product quality and price
perceptions and directly affects customer satisfaction. The second set of inter-
mediate variables involves the two dimensions of loyalty described earlier.
Note that brand image perceptions directly affect the emotional component
of loyalty, while value perceptions and customer satisfaction have an impact
on the more rational aspects of loyalty. Finally, customer satisfaction and the
two loyalty measures directly affect customer retention.
Customer retention is typically measured as an attitudinal item. That is,
one or more questionnaire items relating to the respondent’s probable future
behavior frequently serve as a measure for customer retention. In some
industries, it is becoming increasingly common to encounter actual mea-
sures of retention in models similar to that depicted in Figure 1.7. Clearly,
behavioral measures of retention and its tangents (that is, share of wallet)
require a mature customer database that captures actual customer loss.
The implication of Figure 1.7 is that customer retention is dependent
upon a complex intermingling of effects. It is likely, for example, that some
level of reciprocal causation exists between the affective and cognitive
components of loyalty. In effect, they cause one another. The final outcome
of the hypothetical causal chain shown in the figure is customer retention.
In some instances customer retention has been used as a surrogate for prof-
itability. Whenever possible it is preferable to use individual-level profita-
bility metrics. For example, banks typically track profitability at the household
level. Unfortunately, most companies rarely have this type of information
at the customer level. However, this situation is changing. Financial services
organizations typically have mature profitability measures in place and can
gauge the future worth of virtually any customer. One result of this advan-
tage is that the financial services industry has experienced a tremendous
increase in activity with respect to the development of models aimed
specifically at linking customer satisfaction and profitability. On the other
hand, organizations that have little contact with their customers are in a
much less appealing situation.
It will be especially interesting to assess the relative effects of cus-
tomer loyalty and customer satisfaction with respect to key business out-
comes. Whether, for example, customer loyalty emerges as a more potent
predictor of profitability than customer satisfaction will be of great impor-
tance. Of course, in the framework presented in Figure 1.7, customer satis-
faction is an antecedent of customer loyalty. As such, it is likely that
customer satisfaction will enjoy an important role in the development of
corporate profitability even if it is not as strong a predictor as loyalty. This
is pure conjecture at this point, unfortunately, since there is no evidence to
suggest that the relative impact of these constructs has been tested.
 Why Customer Satisfaction? 15

Figure 1.7 is not only a conceptual schematic of how loyalty and cus-
tomer retention develop. The relationships posited in the figure can actually
be tested using a broad class of techniques known as causal modeling. This
analytical approach permits us to develop complex causal sequences
involving multiple dependent variables. Sophisticated reciprocal relation-
ships like the one depicted between affective and cognitive loyalty are pos-
sible, and can be tested statistically. We will never be able to establish
causality using cross-sectional data, but the causal modeling techniques
introduced later in this book allow us to test whether the causal relation-
ships we posit are consistent with the data.

THE FUTURE OF CUSTOMER

SATISFACTION RESEARCH
The future of customer satisfaction research will most likely be to continue
to establish empirical relationships with key business outcomes. The mea-
surement of customer profitability and its surrogates will play an important
role in future research. Customer retention appears to be emerging as a
viable surrogate for profitability when the value of individual customer
relationships cannot be quantified. Similarly, actual customer behaviors
represent a reasonably good substitute for the types of profitability mea-
sures that, for example, the financial services industry enjoys. The extent to
which these metrics reside in a common data warehouse will largely deter-
mine the ease with which these data can be integrated and related to both
retention and profitability.
Many companies have developed robust models that permit manage-
ment to test the effects of changes in customer and employee satisfaction
on customer retention and profitability. These tools may include cost–benefit
analyses that match the cost of increasing customer (or employee) satisfac-
tion to the resulting increased profits. To the extent possible, customer sat-
isfaction researchers should have such a tool in mind when developing their
survey instruments. In this framework, customer satisfaction research is
part of a bigger picture in which the intent is more explicit: to increase cor-
porate profits. After all, the corporate interest in customer satisfaction is not
happenstance; there is an implicit assumption that increasing customer sat-
isfaction will yield greater profits. And yet, without an empirical link
between customer satisfaction and key business measures such as prof-
itability, this relationship remains a mystery.
There have been numerous attempts to produce simulation applications
that permit ‘what if?’ scenarios using customer satisfaction data. One of the
most comprehensive studies was conducted by Dillon, White, Rao, and
16 Chapter One

Filak (1997) using American Express data. The authors’ simulation model
permitted managers to assess the impact of changing customer satisfaction
on overt behaviors like share of spending, card retention, and spending vol-
ume. Once a structural equations model was developed and tested, it was
migrated to a spreadsheet environment. Using a series of simple com-
mands, managers could test the effects of changing satisfaction with various
product and service quality issues. Retention, spending volume, and share
of wallet were outcome variables that changed as the service and product
quality issues were manipulated in the simulation.
The remainder of this book focuses on techniques that can be used to
develop statistical models for predicting key metrics such as retention or
profitability. We include a chapter on employee satisfaction because it
appears to be an issue of substantive importance with respect to corporate
profitability in particular. This is followed by a detailed discussion of the
Six Sigma process that was developed at Motorola. Interestingly, customer
satisfaction enjoys an important role in the Six Sigma architecture. Chapter
4 addresses current research aimed at linking customer satisfaction and key
business outcomes. Chapter 5 and chapter 6 focus on customer retention
and the isolation of key business metrics, respectively. These are followed
by a detailed discussion of data preparation in chapter 7. Chapters 8, 9, and
10 introduce increasingly rigorous techniques used to link customer satis-
faction and employee satisfaction to financial performance measures. These
three chapters introduce bivariate measures, multiple regression tech-
niques, and causal modeling, respectively. Finally, chapter 11 presents a
futuristic view of customer satisfaction research and describes simulation
software advances, decision support systems, and problems associated with
global data collection.