Customer Segmentation Study Of CRM Based On Improved RFM Model

In the global economic era,the management and maintenance of customer relations is a concern of enterprises.That is to say,only by improving the viscosity of enterprise customers and utilizing the customer relationship management system,can we help enterprise customers to carry out segmentation and realize the in-depth exploration of enterprise customers' preferences and meet the real needs of enterprise customers.RFM model based on the latest customer purchase interval,consumption frequency and consumption amount data is widely used in enterprise customer relationship management system.However,the RFM model ignores the periodicity of customer purchases and thus cannot predict the trend of customer behavior over time.In this paper,an improved RFM model and a customer segmentation algorithm considering customer purchase or purchase cycle are presented,and a prototype analysis and design of customer segmentation based on the improved RFM model is presented.The main work of this paper includes:(1)Through literature research,this paper expounds the research status of customer relationship management,RFM model and customer segmentation at home and abroad,and finds the deficiencies of the current research work.(2)Summarize relevant theories used in this paper,including customer relationship management,RFM model,conceptual principle and process of customer segmentation,etc.(3)In view of the current lack of customer relationship management in the process of customer buy a cyclical problem,improved RFM model,the gap between recent purchase,purchase frequency and consumption amount of RFM model is modified to include customer relationship length,the modified latest purchase,consumption frequency and consumption amount and between customers buy periodic five dimensions of RFM model improvement.Among them,the modified latest purchase interval takes into account the expectation of the last purchase interval,and the length of customer relationship is defined as the time interval between the first purchase and the last purchase to express customer loyalty.The customer purchase cycle reflects whether the customer purchases regularly.(4)Aiming at the deficiency of k-means algorithm,the experiment of customer segmentation was carried out on two real data sets.The values of five dimensions of customer segmentation obtained based on the improved RFM model were firstly analyzed and determined from three aspects of contour value,c-h value and elbow rule.The customer is then divided into different classes using the improved k-means algorithm.Finally,the effectiveness of the proposed model and algorithm is proved by analyzing and comparing the different effects of the improved RFM model and other customer segmentation in the process of customer segmentation through the recommendation evaluation index.(5)According to the general form of customer segmentation module of enterprise customer relationship management system,the improved RFM model proposed in this paper was integrated into enterprise customer relationship management system,and an application prototype of customer segmentation module of customer relationship management system was designed.