Empirical Study On Customer Segmentation Of A Cosmetics Company

The improvement of the quality of life has made cosmetics popular with the public.As a fast-moving consumer goods,it has strong replaceability and fierce competition.Reasonable and effective customer segmentation can analyze the value of different customers,identify high-quality customers and potential high-value customers,provide decision support for corporate marketing,and improve corporate profitability and competitiveness.The main work is to combine G's actual business,build customer lifetime value model,including customer's current value,potential value and loyalty value,and conduct research on customer segmentation.Customer's current value is the composite score of improved RFM model,using entropy method.Logistic regression algorithm and BP neural network algorithm are combined to predict the probability of customers purchasing the four basic products of AR,OI,Lip and FU in the future.Product forecasting probability times average profit,then sum them up,we can take the calculation as potential value.Using the same method,the prediction probability of future returning purchase is used as the customer loyalty value.The weights of three indexes are determined by entropy method,then the composite score of CLV is calculated.Based on the CLV model,using K-Means algorithm for customer segmentation,the customers are classified into 8 categories.At the same time,the RFM model is used for customer segmentation.The results of the two kinds of customer segmentation are analyzed and compared,and the marketing strategies are formulated and implemented.The conclusions are as follows:(1)Based on the CLV model,customers with different values are explored,and the customer's preference for the product is further understood.Among them,the seventh,sixth,second and fourth types of customers buy more skincare products,others prefer makeup products.(2)Customer segmentation based on CLV model is more reasonable.Compared with the RFM model,the customer segmentation based on the CLV model obtains more homogeneous customer bases;and the high-value customers and the low-value customers can be distinguished better.(3)The customer segmentation method based on CLV model improves customer retention rate and company profit.10000 customers were randomly selected to implement corresponding marketing strategies according to different categories.74% of customers repurchased in July,and the marketing effect was better.