Marketing segmentation of banks’ corporate clients based on data mining technique

Abstract

In recent years, the banking services market has been developing dynamically, experiencing a sharp increase in competition. Banks that provide maximum profitability for each client gain the most significant competitive advantage. The sales model in corporate banking is associated with personal interaction between bank employees and business owners, and the subsequent establishment of individual service conditions. However, this approach is often ineffective when a bank faces the issue of maximising the efficiency of business activities. This study aims to segment a bank’s corporate client base and develop a pricing strategy for each of the groups that have been singled out in the process. The study sample consisted of 4,500 corporate clients of a Ukrainian bank who were active users of euro accounts. The k-means data mining algorithm was used to develop marketing segments. The optimal number of clusters was determined by weighing the results of calculating 26 indices from the NbClust package and the bank’s business requirements. Six similarity groups were found during the calculation of the algorithm. The study found that clusters 1 and 2 were a concentration of unprofitable customers for whom an introduction of a service fee was urgently needed. Marketing segments 3 and 4 were customers who did not record net losses but with whom it was deemed necessary to work to improve their profitability. The remaining segments were ‘healthy’ users of euro accounts. With regard to these customers, it was recommended no additional service fees should be imposed. The proposed methodology makes it possible for a bank to remain attractive in a competitive environment while not incurring unnecessary costs.