High Value Customer Profile, Using Data Mining Techniques

Abstract Studying about customer segmentation and creating a customer ranking plan has attracted more attention in recent years. In this regard, this project tries on providing a methodology for customer segmentation depending on their value driver parameters which will be extracted from transaction data. The objectives of this project is to identify the High Value Customer Profile, using data mining techniques such as classification and clustering approaches. In the first phase, the data will be cleaned and patterns will be developed. In the second phase, the data will be profiled and clustered to identify High Value Customer Profile. Background and the Problem Domain Companies are increasingly interested in identifying customers who†¦show more content†¦For example, according to Selden and Colvin (2003) FedEx Corporation is categorizing its clients as â€Å"good†, â€Å"bad† and â€Å"ugly† based on their turnover potential and is charging higher prices from less profitable customers while providing enhanced services to more profitable customers. Numerous industry experts and researchers support this practice. Therefore, it is argued that by treating all its customers equally a company is not only wasting its resources on unprofitable customers, but also is underserving the profitable customers and risks losing them (Selden and Colvin 2003, Venkatesan and Kumar 2004, Gupta and Lehmann 2005). Therefore, by knowing the fact that 80% of business of an organization often comes from 20% of their customers (the Pareto law), it is important for organizations to identify a model to classify the customers as high value customers or not. Most of the past researches (Kim et al. 2006, Khajvand M. and Tarokh M. J. 2011) have performed customer segmentation based on the traditional Customer Relationship Management (CRM) variables. However, these variables might not be sufficient to identify the profile of High Value Customers since traditional CRM variables do not consider the online activities of customers and also the dispute that a customer might influence other