Research On The Prediction Of Economic Value Of Retail Member Customers Based On Product Classification

With the constant changes in the commodity,consumer markets and the rapid development of online channels,the retail industry has entered into the era of meager profit.Under the circumstances of limited market,accurate marketing based on individualized behavior of customers has become a key measure for retail companies to obtain profits.At the same time,the combination of information technology and the retail industry has brought new opportunities for the development of the retail industry.Customers are no longer the same faces,but have different and personalized needs.A highly marketable company should spend resources and energy to track market trends,and monitor the dynamic needs of customers,how to identify the future value of customers,determine the most economical customer,has become a problem that must be solved for the enterprise precision marketing.Forecasting customer's future value through random probability model has been verified in many fields,especially in the retail field.The most typical model is the Pareto/NBD model,and the BG/NBD model,combined with the Gamma-Gamma model,only need to refer to the four aspects of customer transaction information,you can identify the possibility of customer loss and the customer's future economic value.However,because the model ignores the explanatory variables of other customer's personalized behaviors and lacks important information about the classification of commodities.It is difficult to provide decision support for the customer's personalized relationship management and precise product category marketing.Therefore,the purpose of this study is to develop a economic value prediction model for retailer member customers based on product classification by expanding the variable factors of the stochastic probability model and referencing more personalized customer behavior information to develop customer dynamic value for a retail enterprise that sells multiple commodities at the same time.Providing multi-aspect and valuable digital supports for personalized relationship management,and precision merchandise marketing.Completed the following tasks:(1)Constructed a variable factor system for the retail enterprise customer value forecasting model that simultaneously sells multiple categories of commodities,and added commodity category factors to the random probability model;(2)Considered that the customer's future economic value is its sum of economic value expectations in various commodity categories,and established a retail customer's member customer economic value prediction model based on commodity classification.(3)In order to verify the validity of the model,the 2 years actual consumption behavior data of a member of a certain supermarket is selected,and the data is divided into the observation period and the verification period,and the results obtained by the model and the results obtained by the traditional random probability model are compared with the customer actual economic value to verify the validity of the model.(4)Analyze the model results from individual customer level and product category level,and propose application suggestions.The results show that:(1)The retailer's economic value prediction model for member customers based on commodity classification can accurately match the customer's future economic value based on the customer's past purchase behavior,and has strong explanatory power and prediction of customer value ability.(2)Compared with the traditional model,the retailer's economic value prediction model for member customers based on commodity classification improves the application value of the random probability model without reducing the prediction accuracy of the original model.The results can be applied to personalized customer relationship management and precise product category marketing,which is more suitable for customer value research in the retail industry that sells multiple types of products at the same time.