Clustering of Bank Branches Offering Debit Cards Bundled with Insurance

Selling debit cards with insurance through bank branches currently faces the challenge of offering debit cards that meet customers' needs in each group and each area. This research aims to cluster the bank branches that sell the debit cards with personal accident insurance of a bank and analyze the factors affecting the sales of such debit cards in each group. The data of sales of debit cards with accident insurance sold through 919 branches in 2022-2023 were used in this research. The data used for grouping consisted of quantitative and qualitative data, such as debit card sales, the number of sale staff in each branch, demographic data by area, and branch locations. The data were divided into 2 sets by branch locations: branches in Bangkok and its vicinity and branches in the provinces. Clustering the bank branches was done by using the K-prototypes method, and the appropriate number of groups was considered by using the elbow method. The grouping results showed that branches in Bangkok and its vicinity could be grouped into 2 groups: the group of branches in the area with high debit card sales and the group of branches in Bangkok with low debit card sales. Branches in the provinces could be grouped into 3 groups: the group of branches in the provincial cities with very high sales, the group of branches in tourist and industrial areas with high sales and the group of branches outside the province with high sales. From the analysis of factors affecting sales in each group using the decision tree method and the feature importance, it was found that the most important factors affecting sales were population density, average income per capita, and the number of sale staff in each branch. The results of this research can be used to develop appropriate marketing strategies for customers in each branch group.