Product Attributes, Cross Elasticities and Dynamic Market Structure in a Category with Many Products

Abstract

Understanding changes in a market and the ways in which brands compete with one another (market structure) are critically important especially in product categories with changes in product variety. Over time, products offer different combinations of attributes and marketing messages. This leads to changes in competitive structures. In characterizing a category's market structure, researchers need to account for (potentially) large and changing numbers of products while allowing for rapid changes in the marketing environment. Building on existing literature, we propose a model that accounts for dynamics, temporal changes in product variety and competition. A challenge with that approach is (cross-effects) parameter explosion when variety increases. We leverage a dataset with detailed product descriptions that contains information on intrinsic characteristics of the product (e.g., fat content) to construct “distances” between products in attribute space. Cross elasticities are then specified as functions of these distances. Not only does this help us resolve the dimensionality problem, it also helps us study the evolution of the market structure over time at the individual product level. We provide an empirical illustration using data from the yogurt product category which has hundreds of products. Price and advertising are major drivers that we include while accounting for their possible endogeneity. Besides providing a dynamic description of a market, the proposed model framework will serve as an useful input to managers for forecasting the consequences of changing the attributes of products, as well as forecasting the sales of new products with a set of attributes spanned by those in the market.

Keywords: Product differentiation, dynamic linear model, market structure, elasticity

JEL Classification: M31