Using weekly scanner data from 2010 representing 16293 stores (of which 9899 are food stores, 3185 are mass merchandisers, and 3209 are drug stores) in the United States, we estimate store-specific category demand models in 3 different food categories (cereal, coffee, cola). Specifically, we estimate parameters representing baseline category sales, as well as the dependence of category demand on category marketing activities (price, display and feature). We then relate the estimated 16293 sets of store-level demand parameters to geodemographic characteristics of stores’ zip codes, as well as the local competitive intensity faced by stores, in second-stage regressions.
To provide visual perspective on the extent of geographic dispersion in the estimated price sensitivities across stores, we use different shades of red (darker shade denoting greater price sensitivity) for different stores in the Houston metropolitan area. We also show the degree of dispersion across all stores on the national map of the United States.
Using different symbols for different store formats (hexagons for food stores, squares for mass merchandisers, triangles for drug stores), we also report the geographic dispersion separately for the three store formats. Using data visualizations such as these, retail chain managers can better understand and rationalize the incentives for store-level pricing over zonal pricing in the three categories under study. City planners, on the basis of the revealed differences in estimated marketing mix sensitivities across different city neighborhoods, can better understand the real-world implications of approving zoning permits for retail stores of different store formats at different locations in the city. For example, there could be different tax revenue implications, on account of differential price sensitivities of consumers for different store formats, for approving a zonal permit for a food store versus a mass merchandiser in a given zip code.