Product Line Management as Dynamic, Attribute-level Competition

Product lines are composed of SKUs from the same manufacturer which compete directly with one another. Understanding how consumers choose among numerous alternatives of brand, size and format is complicated by the multitude of possibilities among which they can switch over time. We approach this problem in terms of dynamic attribute satiation: that households seek out specific attributes over time, and that their switching patterns will explicitly reflect this tendency. We formulate a dynamic model of sequential choice which relies on the following insight: although households may entirely disagree on the relative merits of two products, they are far more likely to agree on how similar they are. That is, while product preferences may differ markedly across households, product positionings will not. The resulting model offers a number of unique benefits over and above parsimoniousness. Operationalizing product similarity in terms of observed attributes – information typically available in existing scanner panel data – describes the data better than sidestepping such attribute information in lieu of unobserved attributes. Estimated “part worths” enable one to characterize how substitutable or complementary two products are from the consumption perspective of households: for example, in one category, we find brand name to be a crucial determinant of how similarly two products are perceived, though in another apparently similar category, size is the critical attribute. While on the one hand this suggests that manufacturers may be able to capitalize upon a strong brand name by umbrella branding, it also suggests that products may cannibalize one another’s sales if other attributes in the product category are perceived to be unimportant. Switching probabilities depend on household-level inertia as well as on inter-product similarities, allowing an understanding of both these effects as well as their interplay. This stands in contrast with state dependence models that ignore inter-product similarities, which force different products to be always complements for a variety-seeking household, and always substitutes for an inertial household. The model accounts for patterns of asymmetric substitutability, and suggests that complementarity and substitutability should not be construed solely at the category level. Key-words: State Dependence, Inertia, Variety-seeking, Dynamic Choice, Market Structure, Similarity Structure