How assortment size, distinctive allocations, and local market density affect convenience store transaction performance

Convenience stores set store assortment sizes and allocate assortment weights across categories to either follow or be distinct from overall chainwide profiles. Such stores operate in local markets that vary in population density and implicit density of competing stores. Using over 3.2 million transactions from a proportional stratified random sample of 235 stores of a large convenience store retailer in Taiwan, we capture assortments and transaction performance. We also obtain local market population densities. We develop a cosine similarity measure to capture distinctiveness in store-level cross-category assortment allocations. Regression models with endogeneity checks reveal that, controlling for assortment size, distinctiveness in cross-category assortment weight allocation has a modest positive effect on store unit volume but no effect on revenue. Decomposing store volume into transaction components reveals that store assortment size and, to a lesser extent, the degree of distinctiveness in cross-category allocations have greater positive effects on store transaction counts than on mean basket sizes. Local market density has a small positive effect on store transaction count, but a much stronger negative effect on store mean basket size. We interpret these results and discuss their implications. These results extend supermarket findings about assortment size sales effects to the convenience store format and go further by examining its effects on each transaction component. The study breaks new ground by examining the effects on the basic store volume components of transaction count and mean basket size of both a) distinctiveness in store-level cross-category allocations of assortment weights and b) local market density.