Capturing Consumption Flexibility in Assortment Choice from Scanner Panel Data

This study develops and estimates a model to empirically identify two behavioral effects (namely consumption flexibility and state dependence) that may underlie temporal and horizontal assortment choice, using scanner panel data where consumption information is unavailable. The proposed approach permits a consumer's consumption utility to be dependent on previous consumptions, thus capturing state dependence both across purchase occasions and within horizontal assortments. Moreover, consumers' purchase and consumption decisions are modeled at two distinctive and sequentially related stages, which allows for incorporating the effect of consumption flexibility. The model is estimated on scanner panel data of yogurt purchase. It is found that the two captured effects provide strong empirical support with face validity for the temporal and horizontal assortment choice patterns observed in the data. The behavioral insights derived from estimating the proposed model can also be translated into significant managerial implications.