Impact of Wal-Mart Supercenter Entry on Consumer Purchase Behavior: An Empirical Investigation

Abstract

Supermarkets operate in an increasingly competitive environment. The rapid growth of alternative retail formats has transformed not only the competitive structure of the industry, but also the way in which consumers shop. The biggest challenge to the industry is coming from none other than the world's largest retailer: Wal-Mart. Although a relatively new player, Wal-Mart through its supercenter format has become the nation's largest grocer and is cited by supermarket managers as their biggest concern in the coming years. Despite the dramatic proliferation of supercenters, relatively little is known about the impact it has on the performance of a traditional grocery store or how it changes consumer buying behavior. This paper provides an empirical study of entry by a Wal-Mart supercenter into a local market. Using a unique frequent shopper database from a supermarket, we study the impact of Wal-Mart's entry on household purchase behavior. The database records purchases for over 10,000 households before and after Wal-Mart's entry. Our primary focus in this paper is to analyze Wal-Mart's impact on the two key household decisions: store visit and basket size. We develop a joint model of inter-purchase time and basket size and allow for a structural break at the time of competitive entry. The model allows us to evaluate the impact of Wal-Mart on household store visit frequency and basket size, while allowing for consumer heterogeneity. We investigate the shopping and demographic characteristics of the consumers that are most likely to shift purchases to Wal-Mart. Our results show that the incumbent store lost 17% volume - amounting to a quarter million dollars in monthly revenue - following Wal-Mart's entry. Decomposing the lost sales into components attributed to store visits and in-store expenditures, we find that the majority of these losses were due to fewer store visits with a much smaller impact on basket sizes. This in turn suggests that strategies designed to increase store traffic could be quite effective in mitigating losses to Wal-Mart. Interestingly, we find that a small proportion of customers account for a large proportion of the losses. For example, 10% of the households account for 45% of the store's lost revenue, while 20% of the customers account for almost 70% of the lost revenue. Profiling these households in terms of their observed characteristics, we find that distance to store, while useful, explains little of the variation in household heterogeneity. Households that respond to Wal-Mart are likely to have an infant and pet in the family, and are more likely to be weekend shoppers. Furthermore, we find that these households are large basket consumers, confirming the findings in Bell and Lattin (1998) that large basket buyers are more likely to choose an EDLP operator. On the other hand, households that spend a large proportion of their grocery expenditures on fresh produce, seafood, and home meal replacement items are less likely to defect to Wal-Mart. Using a holdout sample, we find that these shopper characteristics can be quite useful in identifying potential defectors to Wal-Mart. Implications and strategies for supermarket managers to compete with Wal-Mart are discussed.