­­­­­Investigating the Loyalty of Customers of Dairy Products in Chain Stores Using Data Mining

Abstract This study investigated the loyalty of dairy product customers in chain stores using data mining and investigated the effective factors through logit regression. The statistics used in this study, including the number and amount of purchases by customers, the purchased brand of dairy product, the time and date of purchase, and the available demographic information, were obtained from the convenience store data warehouse daily from the beginning of March 2019 to April 2020. The total number of customers surveyed in the study was 18,232 people. Using the K-means method, customers were clustered. According to the results of the final scenario of 3 clusters, 91% of the active customers were in the first and second clusters. Based on the results, the silhouette coefficient of the third cluster was selected as the superior cluster. According to the results of the recency, frequency, monetary, and continuity (RFMC) model, 50% of the active customers could be grouped according to the leading model. Therefore, the model coverage was appropriate. According to the results of the logit model, older age, access to parking and facilities, quality of service and staff behavior, offering traditional products, and variety of dairy products҆ brands increased the likelihood of customers’ ҆ presence in the loyal category, and being married or male and having a higher educational level increased the likelihood of customers’ ҆ presence in lower loyalty ranks.