Building Models for Marketing Decisions

by Leeflang, Peter S. H.; Wittink, Dick R.; Wedel, Michel; Naert, Philippe A.

• ISBN13: 9780792378136
• ISBN10: 079237813X
• Format: Paperback
• Copyright: 2000-07-01
• Publisher: Kluwer Academic Pub

Summary

The market environment is changing rapidly. Prior to scanner data, ACNielsen, the major supplier of information on brand performances, said its business was to provide the score but not to explain or predict it. Now, model-based insights are not only demanded by managers, but can also be meaningfully provided. It is common for managers in many countries to receive market feedback frequently, quickly and in great detail due to the use of scanners and computers. With advances in information technology and expertise in modeling, IRI introduced model-based services in the US that explain and predict essential parts of the marketplace. ACNielsen followed, and marketing researchers have been developing increasingly valid, useful and relevant models of marketplace behavior ever since. Models that provide information about the sensitivity of market behavior to marketing activities such as advertising, pricing, promotions and distribution are now routinely used by managers for the identification of changes in marketing programs that can improve brand performances. Building Models for Marketing Decisions describes marketing models that managers can use as an aid in decision making. It has long been known that even simple models outperform judgments in predicting outcomes in a wide variety of contexts. More complex models potentially provide insights about structural relations not available from casual observations. Although marketing models are now widely accepted, the quality of the marketing decisions is critically dependent upon the quality of the models on which those decisions are based. In this book, which is a revision and expansion of Naert and Leeflang's Building Implementable Marketing Models (1978) , the authors discuss in detail the model-building process. They distinguish four parts in this process: specification, estimation, validation and use of models. Throughout the book, the authors provide examples and illustrations. This book will be of interest to researchers, analysts, managers and students who want to understand, develop or use models of marketing phenomena.

Author Biography

Peter S.H. Leeflang (1946) is Professor of Marketing at the Department of Economics at the University of Groningen, The Netherlands, and member of the Royal Netherlands Academy of Arts and Sciences.Dick R. Wittink (1945) is the General George Rogers Clark Professor of Management and Marketing at the Yale School of Management, USA, and Professor of Marketing and Marketing Research at the University of Groningen, The NetherlandsMichel Wedel (1957) is Professor of Marketing Research at the Department of Economics, University of Groningen, The Netherlands.Philippe A. Naert (1945) is the Dean of TIAS Business School and Professor of Marketing, Tilburg University, The Netherlands.

Table of Contents

Preface	xiii
PART ONE: Introduction to marketing models	1	(46)
Introduction	3	(10)
Purpose	4	(3)
Outline	7	(3)
The model concept	10	(3)
Classifying marketing models according to degree of explicitness	13	(8)
Implicit models	13	(1)
Verbal models	13	(2)
Formalized models	15	(3)
Numerically specified models	18	(3)
Benefits from using marketing models	21	(16)
Are marketing problems quantifiable?	21	(3)
Benefits from marketing decision models	24	(4)
Building models to advance our knowledge of marketing	28	(4)
On the use of a marketing model: a case study	32	(5)
A typology of marketing models	37	(10)
Intended use: descriptive, predictive, normative models	37	(3)
Demand models: product class sales, brand sales, and market share models	40	(1)
Behavioral detail	41	(3)
Time series and causal models	44	(1)
Models of ``single'' versus ``multiple'' products	45	(2)
PART TWO: Specification	47	(252)
Elements of model building	49	(36)
The model-building process	49	(6)
Some basic model-building terminology	55	(11)
Specification of behavioral equations: some simple examples	66	(19)
Models linear in parameters and variables	66	(1)
Models linear in the parameters but not in the variables	67	(12)
Models non-linear in the parameters and not linearizable	79	(6)
Marketing dynamics	85	(16)
Modeling lagged effects: one explanatory variable	85	(11)
Modeling lagged effects: several explanatory variables	96	(1)
Selection of (dynamic) models	97	(1)
Lead effects	98	(3)
Implementation criteria with respect to model structure	101	(22)
Introduction	101	(1)
Implementation criteria	102	(8)
Models should be simple	102	(3)
Models should be built in an evolutionary way	105	(1)
Models should be complete on important issues	105	(2)
Models should be adaptive	107	(1)
Models should be robust	108	(2)
Can non-robust models be good models?	110	(5)
Robustness related to intended use	115	(5)
Robustness related to the problem situation	120	(3)
Specifying models according to intended use	123	(34)
Descriptive models	123	(7)
Predictive models	130	(14)
Normative models	144	(13)
A profit maximization model	144	(7)
Allocation models	151	(3)
Appendix: The Dorfman-Steiner theorem	154	(3)
Specifying models according to level of demand	157	(22)
An introduction to individual and aggregate demand	157	(7)
Product class sales models	164	(3)
Brand sales models	167	(4)
Market share models	171	(8)
Specifying models according to amount of behavioral detail	179	(42)
Models with no behavioral detail	180	(1)
Models with some behavioral detail	180	(15)
Models with a substantial amount of behavioral detail	195	(6)
Modeling competition	201	(20)
Competitor-centered approaches to diagnose competition	202	(6)
Customer-focused assessments to diagnose competition	208	(3)
Congruence between customer-focused and competitor-centered approaches	211	(4)
Game-theoretic models of competition	215	(6)
Stochastic consumer behavior models	221	(30)
Purchase incidence	222	(4)
Introduction	222	(1)
The Poisson purchase incidence model	222	(1)
Heterogeneity and the Negative Binomial (NBD) purchase incidence model	223	(1)
The Zero-Inflated Poisson (ZIP) purchase incidence model	224	(1)
Adding marketing decision variables	225	(1)
Purchase timing	226	(5)
Hazard models	226	(3)
Heterogeneity	229	(1)
Adding marketing decision variables	230	(1)
Brand choice models	231	(15)
Markov and Bernouilli models	232	(7)
Learning models	239	(1)
Brand choice models with marketing decision variables	240	(6)
Integrated models of incidence, timing and choice	246	(5)
Multiproduct models	251	(16)
Interdependencies	252	(4)
An example of a resource allocation model	256	(2)
Product line pricing	258	(3)
Shelf space allocation models	261	(3)
Multiproduct advertising budgeting	264	(3)
Model specification issues	267	(32)
Specifying models at different levels of aggregation	267	(14)
Introduction	267	(1)
Entity aggregation	268	(11)
Time aggregation	279	(2)
Pooling	281	(1)
Market boundaries	282	(4)
Modeling asymmetric competition	286	(5)
Hierarchical models	291	(4)
A comparison of hierarchical and non-hierarchical asymmetric models	295	(4)
PART THREE: Parameterization and validation	299	(224)
Organizing Data	301	(22)
``Good'' data	301	(4)
Marketing management support systems	305	(3)
Data sources	308	(8)
Data collection through model development: A case study	316	(7)
Estimation and testing	323	(118)
The linear model	324	(37)
The two-variable case	324	(1)
The L-variable case	325	(2)
Assumptions about disturbances	327	(3)
Violations of the assumptions	330	(18)
Goodness of fit and reliability	348	(13)
Pooling methods	361	(8)
Generalized Least Squares	369	(7)
Simultaneous equations	376	(7)
Nonlinear estimation	383	(6)
Maximum Likelihood Estimation	389	(7)
Maximizing the likelihood	389	(2)
Example	391	(1)
Large sample properties of the ML-Estimator	392	(3)
Statistical tests	395	(1)
Non- and semiparametric regression models	396	(12)
Introduction	396	(1)
Advantages and disadvantages of the parametric regression model	397	(1)
The nonparametric regression model	397	(5)
The semiparametric regression model	402	(6)
Illustration and discussion	408	(5)
Subjective estimation	413	(28)
Justification	413	(3)
Obtaining subjective estimates	416	(12)
Combining subjective estimates	428	(5)
Combining subjective and objective data	433	(3)
Illustration	436	(5)
Special topics in model specification and estimation	441	(38)
Structural equation models with latent variables	441	(10)
Outline of the model and path diagram	441	(8)
Seemingly unrelated regression models	449	(1)
Errors-in-variables models	449	(1)
Simultaneous equations	450	(1)
Confirmatory factor analysis	450	(1)
Mixture regression models for market segmentation	451	(7)
Introduction	451	(1)
General mixture models	452	(1)
Mixture regression models	453	(2)
Application	455	(1)
Concomitant variable mixture regression models	456	(1)
Latent Markov mixture regression models	457	(1)
Time-series models	458	(15)
Introduction	458	(1)
Autoregressive processes	459	(2)
Moving average processes	461	(1)
ARMA processes	462	(1)
Stationarity and unit root testing	463	(2)
Integrated processes	465	(1)
Seasonal processes	465	(2)
Transfer functions	467	(3)
Intervention analysis	470	(3)
Varying parameter models	473	(6)
Validation	479	(44)
Validation criteria	480	(2)
Statistical tests and validation criteria	482	(2)
Face validity	484	(3)
Model selection	487	(13)
Introduction	487	(1)
Nested models	488	(4)
Non-nested models	492	(3)
Causality tests	495	(5)
Predictive validity	500	(8)
Illustrations	508	(9)
Validation of subjective estimates	517	(6)
PART FOUR: Use / Implementation	523	(56)
Determinants of model implementation	525	(20)
Organizational validity	526	(8)
Personal factors	526	(2)
Interpersonal factors: the model user - model builder interface	528	(4)
Organizational factors	532	(2)
Implementation strategy dimensions	534	(11)
Introduction	534	(1)
Evolutionary model building	535	(3)
Model scope	538	(5)
Ease of use	543	(2)
Cost-benefit considerations in model building and use	545	(20)
Tradeoffs	546	(1)
The cost of building models	547	(1)
Measuring benefits	548	(5)
Some qualitative examples	553	(3)
General observations	556	(9)
Models for marketing decisions in the future	565	(14)
Examples of recent developments in model building	565	(3)
The role of models in management decisions	568	(2)
A broader framework	570	(9)
Bibliography	579	(38)
Author Index	617	(20)
Subject Index	637