Modeling consumption patterns in the attribute space: Theory and evidence of hybrid behavior

This paper presents and estimates an attribute-based choice model where pure inertial and pure variety seeking behaviors are incorporated as, sometimes transient, particular cases of a general utility function. In light of the evidence of several state-dependent patterns, the standard framework is extended by considering non-linear partial utilities that allow for a wide spectrum of behavioral patterns. Under the inverted-U hybrid utility assumption, the consumer behaves inertially among unfamiliar items, but eventually switches to variety-seeking behavior when a high level of familiarity with the implied attributes is approached. An empirical analysis is run using scanner databases for fabric softener, saltine cracker, and catsup product categories. Non-linear specifications provide the best fit of the data and reveal that the gradual trend, to seek variety as the level of familiarity with the purchased items increases, can be forecasted with higher predictive ability. The results confirm significant differences in the purchasing patterns across segments, as well as relevant segments presenting hybrid behavior for most attributes.