This study shows that individuals’ grocery shopping habits are increasingly useful in predicting their credit card payment behaviors and that such increased predictive power can translate into increased profits for businesses. Building on previous work, we identify five general grocery buying habits associated with payment behaviors: (1) shopping the same day of the week, (2) relying on a shopping budget, (3) buying the same brands and categories consistently, (4) taking advantage of deals and offers promotional, and (5) purchasing of healthy products. Knowing the Five Habits of Grocery provides guidance on how to turn raw grocery data into input for flexible machine learning models, which we use to assess the increasing predictive power of grocery data. We found that incremental predictive gains from grocery data, as well as standard data sets used by issuers, range from 0.2% to 9.4%, depending on the data environment that issuers face in different credit markets. Furthermore, simulations of issuer credit span decisions show that the marginal effect on issuer earnings ranges from 0.3% to 15.2% and is the largest effect for consumers without a stable credit history. This suggests that grocery data may enable credit card issuers to offer credit to consumers who currently have limited or no access to credit. We also discuss a boundary case in which grocery data may not have an incremental value. Overall, this study highlights how consumer data from a seemingly unrelated domain can help address a focal domain management problem.
Keywords: habits, grocery shopping, consumer finance, credit card, data economics