Customer future profitability assessment: A data-driven segmentation function approach

One of the important tasks in customer relationship management is to find out the future profitability of individual and/or groups of customers. Data mining-based approaches only provide coarse-grained customer segmentation. It is also hard to obtain a high-precision structure model purely by using regression methods. This paper proposes a data-driven segmentation function that provides a precise regression model on top of the segmentation from a data mining approach. For a new customer, a structure model constructed from profit contribution data of current customers is adopted to assess the profitability. For an existing customer, external information such as stock value performance is taken into the regression model as well as historical trend prediction on the profit contribution. In addition, this paper shows how the proposed approach works and how it improves the customer profitability analysis through experiments on the sample data.