About AMA

Executive Summary
Marketing decision makers, retailers, and multiproduct/multicategory manufacturers typically set prices and levels of other promotional activities across several different product categories. To understand the consequences of their decisions fully, these firms need to understand (1) the total impact of a marketing activity within a product category, which requires an understanding of the impact on purchase incidence, brand choice, and purchase quantity, and (2) the impact of marketing variables in one category on the sales in other product categories. This calls for an integrated approach to modeling the three decisions across categories.

This article provides an integrated utility-maximizing framework for households' multicategory purchase incidence, brand choice, and purchase quantity decisions. On a store visit, households maximize utility across several categories subject to an overall budget constraint. They then choose to purchase a subset of categories. Within each category, a specific brand and an associated quantity level are chosen. The authors derive the marginal and joint purchase incidence probabilities, the conditional brand choice probabilities, and the conditional quantities of the chosen brands. Under their assumptions for the stochastic terms in the specification, the authors obtain closed-form expressions for the household demands, joint purchase incidence, and brand choice probabilities, along with conditional purchase quantities. The authors show that the modeling framework accounts for coincidence and for complementarity/substitutability across categories. In addition, they account for unobserved heterogeneity in the estimation. Thus, the three key features of multicategory models proposed in the literature are all reflected in their framework.

Although it is also possible to decompose the total effect of marketing actions using several submodels, the integrated approach proposed here makes it possible to compute consumer welfare measures corresponding to marketing actions. For example, if a retailer decides to drop a product from the shelf, not only can the impact of such actions on the firms involved be computed (as would an approach with separate models), but it is also possible to compute the impact of that move on consumer welfare. For example, in complementary product categories, withdrawing a product from one category might lower consumer welfare not only for that category but for the complementary category as well. Another advantage of the integrated approach is that it is much more parsimonious (i.e., the number of parameters estimated would rapidly increase with categories and brands in the submodeling approach).

The authors provide an empirical application of the methodology using four categories from the Stanford market basket data—laundry detergent, fabric softener, toilette tissue, and paper towel. The results indicate that cross-category effects, measured by the cross-category price elasticities between brands in those categories, are small in general, and they arise mostly through consumers' joint purchase incidence decisions rather than at the purchase quantity decision conditional on brand choice, indicating that the effects of coincidence are larger than the effects of complementarity in these data. Another implication of the results is that prima facie evidence from descriptive statistics cannot be relied to draw conclusions about the substitutability/complementarity across categories. As the authors show, two categories that appeared to be substitutes in the raw data are actually unrelated after controlling for household heterogeneity and the effects of marketing activities.

Biography
Inseong Song is Assistant Professor of Marketing in Hong Kong University of Science and Technology. He received his PhD from the University of Chicago. He is interested in studying dynamics of demand for new products, consumer purchase behavior, and firms' competitive behaviors. Previous research has appeared in Management Science, Journal of Marketing Research, and Quantitative Marketing and Economics.

Pradeep K. Chintagunta is Robert Law Professor of Marketing in the Graduate School of Business at the University of Chicago. He is interested in studying strategic interactions among firms in vertical and horizontal relationships. His research also includes measuring the effectiveness of marketing activities in pharmaceutical markets, investigating aspects of technology product markets, and the analysis of household purchase behavior.