Abstract:This paper carries out customer value stratification in the traditional retail industry and adopts correspondingmarketing strategies for customers of different values which not only reduces the marketing cost but also enhances the marketing effects and ultimately increases the enterprise profit significantly.Based on the classical RFM model the calculationmethod of index F is modified then the weights of model indexes are determined by entropy weight method and finally thecustomers are divided into different value levels by K-means clustering analysis.According to the relevant data of a largesupermarket with the value stratification of customers in the observation period this paper analyses the profit-making situation of different types of customers and verifies the rationality of the clustering results which proves customers with highervalue level create more profits.Finally corresponding marketing strategies are given to different types of customers.