The selection and pricing of retail assortments: an empirical approach

Abstract

Determining the best assortment to carry is one of the most central problems in retailing. The key decisions are what items to stock and how to price them. In this article, we develop and test an empirical (or nonparametric) approach that simultaneously addresses the selection and pricing problems. The approach is then applied to the problem of selecting an optimal assortment of backpacks from a field of eight available items. The empirical procedure generates data from a calibration sample of shoppers that we use to determine supposedly optimal assortments (of size k = 2, 3, 4, and 5). In a validation analysis on a new sample of shoppers, from the same parent population, it is shown that in this instance the new approach does yield significantly more profitable retail assortments, in all cases (e.g., for all assortment sizes), when compared with more traditional approaches. Additionally, the experimental approach predicts sales and profitability more accurately than traditional approaches.