A heavily used competitive tactic in the grocery business is the weekly advertising of price reductions in newspaper inserts and store fliers. Store managers commonly believe that advertisements of price reductions and loss leaders help to build store traffic by diverting customers from competing stores, thereby increasing store volume and profitability. It is therefore not surprising that grocery retail planners across competing stores expend considerable thought on what items to advertise each week and at what levels of prominence. What is surprising, however, is that we marketing scientists do not know much about the manner and extent to which feature advertising in a competitive environment influences where and how customers shop. The marketing science literature has not even been able to establish that feature advertising has a substantial impact on store choice, let alone the more operational question of which categories are better at drawing consumers away from one store and into a competing store. In this paper we employ a stochastic choice modeling framework to propose and empirically estimate a disaggregate, consumer-level model of the effects of feature advertising on store choice. We use this model to understand which categories are more influential drivers of store traffic and better at diverting consumers from competing stores.