Forecasting Customer Lifetime Value: A Statistical Approach
Keywords:
data mining, customer lifetime value, customer relationship management, truncated data, hazard function model
Abstract
We propose a method of forecasting customer lifetime value using the customer usage database, sampling strategy, segmentation, modeling, and validation techniques in data mining. The highly heterogeneous customer database being mined will allow the inclusion of uncertainty components in the estimation of customer lifetime value. A hazard function model based on a truncated lifetime data of customers can provide adequate information to compute the value of the incentive to be offered and the length of the lock up period for the customer.
