Abstract: The economic value of retail customers is determined together by purchase frequency and monetary value. This article employs the NBD model to fit purchase frequency and the gamma-gamma model to fit the average monetary value. Based on the Bayes theorem we can calculate the expected future purchase frequency and average monetary value in a given customers fast purchase behavior,and the customer's future economic value is the multiplication of these two expected values. This article applied the above stodrastic models to fit empirically the frequent shopper card data of a retailer. The results show that the models not only can accurately fit the data of purchase frequency and monetary value,but also can accurately predict customers future economic value. The modeling method is of great value to retailers in enhancing customer analysis and improving customer management.

Key words: future customer economic value, NBD model, gamma-gamma modle