Identifying price sensitive consumers: The relative merits of demographic vs. purchase pattern information

We have very limited knowledge of the antecedents of household-level price sensitivity. In this paper we systematically examine the price sensitivities of a large number of households across multiple product categories to attempt to uncover the antecedents of price sensitivity. Using ERIM scanner panel data provided by A. C. Nielsen, we estimate the price sensitivity of each household in the panel for each of five product categories in two market areas. We test two competing structural models that link household price sensitivity to a series of demographic and shopping pattern variables. We find that the shopping pattern variables have substantially greater predictive validity in determining a household's price sensitivity. Thus, our model suggests that shopping pattern variables, commonly available to retailers through means such as scannable "loyal shopper" cards, provide retailers with better information about household-level price sensitivity than, typically much more difficult to procure, household-level demographic data.