Probabilistic Subset-Conjunctive Models for Heterogeneous Consumers
Published 11/1/2005
Author: Kamel Jedidi and Rajeev Kohli
View this content
Executive Summary
The authors present two generalizations of conjunctive and disjunctive rules that consumers use when making complex decisions, unfamiliar decisions, and decisions involving large sets of alternatives. The first generalization relaxes the assumption that an acceptable alternative must be satisfactory on all attributes (conjunctive rule) or on at least one attribute (disjunctive rule). The second generalization relaxes the assumption that consumers have deterministic preferences. The two generalizations are combined to formulate a probabilistic subsetconjunctive rule, by which an acceptable alternative is satisfactory on a subset of attributes and an attribute level is satisfactory with a probability that can differ from one level to another. Subset-conjunctive rules allow for evaluations of alternatives in the presence of incomplete attribute information, they provide a decision maker flexibility in choosing how stringent the decision criteria are for screening alternatives, and they subsume as special cases conjunctive and disjunctive rules and conjunctive rules over any proper subset of the attributes.
The authors describe two methods for inferring probabilistic subset-conjunctive rules. The methods enable a marketer to identify customer segments that are revealed in the process of rule inference but are otherwise unobserved. Each segment can have its own probabilistic subset-conjunctive rule. The first method requires consumers to assess whether each of a set of alternatives is acceptable or unacceptable. The second method requires them to select, at most, one alternative from each of several choice sets. The latter method assumes that choice is made in two steps. In the first step, consumers screen alternatives using a probabilistic subset-conjunctive rule. In the second step, they either reject all screened alternatives or choose a single alternative that they like best. Unlike other two-step choice models in the literature, the present choice model does not require that consumers make independent judgments in the two steps. For example, alternatives with a greater consideration probability can have a greater choice probability, all else being equal. The authors use the proposed models to infer market segments and the associated subset-conjunctive rules for consumers of household batteries and personal computers. They describe how the predictions of product consideration, consumer choice, and market shares differ when the proposed models are used instead of linear models and how they differ in the insights they provide into consumer decision processes.
Biography
Kamel Jedidi is Professor of Marketing in the Graduate School of Business at Columbia University. He has a bachelor’s degree in Economics from the Faculté des Sciences Economiques de Tunis, Tunisia, and master’s and doctoral degrees in Marketing from the Wharton School, University of Pennsylvania. He has published more than 30 articles in leading marketing and statistics journals, the most recent of which have appeared in Journal of Marketing Research, Marketing Science, Management Science, International Journal of Research in Marketing, and Psychometrika. His substantive research interests include pricing, product design and positioning, diffusion of innovations, market segmentation, and the long-term impact of advertising and promotions. His methodological interests lie in multidimensional scaling, classification, structural equation modeling, and Bayesian and finite-mixture models. He was awarded the 1998 IJRM Best Article Award and the Marketing Science Institute 2000 Best Paper Award.
Rajeev Kohli is a professor in the Graduate School of Business at Columbia University. He has a doctoral degree in Applied Economics and Decision Sciences from the University of Pennsylvania; an MBA from Northern Illinois University; and a bachelor’s degree in Electrical Engineering from BITS, Pilani, India. His research interests are in models of consumer preference and choice, techniques for new product development, Internet technology and personalization, analysis of algorithms, and combinatorial optimization. He has published articles on these subjects in leading journals in marketing, management science, operations research, mathematical psychology, computer science, and discrete mathematics. He teaches courses on new product development and marketing models to MBA, executive MBA, and doctoral students. He also teaches executive programs on innovation, creativity, and new product development in the United States, Asia, and Europe.
J Marketing Research, Volume 42, Number 4, November 2005
View Table of Contents.