Every House a Warehouse: An Inventory-Theoretic Model of Retail Shopping Behavior

Every House a Warehouse: An Inventory-Theoretic Model of Retail Shopping Behavior
Author(s): Metters, Richard and Semple, John
Year: 2000
Paper Number: GBS-DIA-2000-001
Goizueta Department: Decision and Information Analysis

Abstract

We explain household shopping behavior using a flexible stochastic inventory framework that links store choice, purchase cost, purchase quantity, purchase timing and consumption within a single dynamic system. In our approach, households operate like warehouses and make shopping decisions to minimize their total inventory costs. We show that store choice and purchase quantity are governed by a form of generalized (s, S) control policy that holds for a much richer class of stochastic distributions than could be determined from previous works. Our model predicts the existence of two types of shoppers: those loyal to a single store; and those who can be classified as “store switchers,” i.e., those who make substantial purchases at different stores. For both segments, store choice and purchase quantities are shown to depend on the household’s fixed and variable costs of shopping at the various stores. For store switchers, our model predicts more complex relationships involving store choice, basket size, purchase timing, and lags of these variables. All of our model’s normative predictions are supported by an empirical analysis of syndicated panel data on 246 households collected over two years. Our findings support the importance of modeling store choice as a dynamic process. This work suggests the importance of matching targeted retail promotions with time-dependent shopping behavior.