A Hybrid Model of Data Mining and MCDM Methods for Estimating Customer Lifetime Value

Due to competitive environment, companies want to create a durable relationship with their customers. Building effective customer relationship management, companies should estimate customer lifetime value (CLV). CLV is normally calculated in terms of recency, frequency and monetary (RFM) variables. Allocating resource to customers’ segments regarding to communication channel based on CLV can be helpful for managers .In this paper, a hybrid model for estimating CLV based on RFM variables incorporated with data mining and multi criteria decision making (MCDM) methods has been proposed. The proposed methodology contains three phases: data understanding and collection, data preprocessing, modeling and data analyzing in which Fuzzy Analytical Hierarchy Process (FAHP) has been used to determine RFM variables’ weights. Then, K-means clustering method as a data mining method was employed in order to customer clustering and segmentation. Customer clusters were then ranked using MCDM method. Based on this ranking the company’s marketing and advertising resources have been allocated to different clusters. Finally, the proficiency of the model was shown by conducting a case study of cosmetics industry