The study of individual differences has always been a cornerstone of behavioural sciences and consumer research. We adopted the concept of entropy from information theory and applied it to measure the extent to which a series of shopping patterns is consistent or random, therefore captured individual differences between customers in retail settings. We first defined a form of normalised entropy and demonstrated its general properties with a simulation, results of which showed that entropy can effectively reflects diversity in long-term purchasing patterns. With the use of real-life transactional data from a major pharmacy chain in the United Kingdom (N = 16,132), we then investigated the relationships between entropy and spending, while also considering personality traits and shopping attitudes. Findings from both frequentist and Bayesian regression models showed that entropy had strong relationships with spending and saving: customers who were more consistent in their shopping patterns tended to spend more in total, but spent less per transaction and saved more from discounted items. Notably, entropy clearly had stronger effects than personality scales or shopping attitude measures in explaining spending and saving. Furthermore, decision trees showed that entropy was the most important features when predicting spending even when non-linear relationships were considered, while mixed effect models showed that the relationships between entropy and spending could be found when analyses were conducted with quarterly analyses.

Zoom Invitation Link

Meeting ID: 952 6867 5965

Passcode: 444933