Customer Scoring and Competition - The Case of Retail Format Competition

Customer scoring is widely practiced in industries such as mail order, telecommunications or financial services to identify customers for specific marketing activities. It has been suggested in the literature to score customers on the basis of the value of their transactions per unit of waiting time while taking into account the possibly heterogeneous covariation of waiting times and purchase volumes. We extend this scoring approach by conditioning waiting times and purchase volumes on individual customers' interactions with a competitor. In the literature, it has been repeatedly proposed to include interactions with competitors when scoring customers. On the basis of individual purchase data of customers with two retailers over a period of two years, we score customers using a hierarchical Bayesian method, which incorporates the possibility that purchases with a competing retailer affects both waiting times and purchase volumes with the focal retailer. We test this model using simulated data and demonstrate that the model identifies individual-level and group-level parameters well. It is then applied to a real-life data set. We find significant effects of interactions with the competitor on purchase timing with the focal retailer on the individual level. This results in differences in customer scores when competition is included. Our results indicate that it is highly relevant to account for competition when scoring customers. Implications and directions for future research are discussed