Research on E-commerce Customer Value Segmentation Model Based on Network Behavior

Abstract

The traditional customer segmentation model is based on the value of the customer's consumed data, and the customer's consumption habit is obtained to predict its potential consumption value, and then the marketing strategy and customer retention strategy are determined. According to the characteristics of e-commerce enterprises and the recordability of historical network behavior of e-commerce customers, this paper constructs the AFCS customer segmentation model based on the traditional customer segmentation model customer value matrix which represents the existing value of e-commerce customers, added two potential value factors representing e-commerce customers, one is the total number of clicks of users who represent the activity of e-commerce customers, the other is the total number of user collections and shopping carts representing the potential purchase intention of users. Then the AFCS model is tested with K-Means, SOM and SOM + K-Means, the experimental results prove that the AFCS model based on the SOM + K-Means algorithm is superior to the AFCS model using the SOM or K-Means algorithm alone, and its customer segmentation results are more accurate, which can provide reference for effective customer retention strategies and targeted marketing for e-commerce.