Research On Identification,Early Warning And Countermeasures Of Lost Customers

The basic economics principle tells us that when the technology of the product tends to mature,the market competition will shift from the technical competition to the competition of how to enhance the added value of the product.The Porter Competitive Theory gives three coping strategies,namely C ost leadership strategy,differentiation strategy and specificization strategy.Differentiation includes two aspects,one is the differentiation of products,and the other is the differentiation of customers.The so-called differentiation of customers refers to providing customers with standardized products and maximizing them for different customer groups.Additional services that meet the preferences.That is to say,in the face of relatively mature technology products,when enterprises choose to implement differentiated strategies to enhance competitiveness,the key to the premise and success is to classify customers and understand the relevant groups of customers based on the classification.Demand characteristics.China's broadband market is a relatively mature market for such products.The competition in the broadband market is basically the competition of the adjustment of the existing customers.How to establish its own competitive advantage in this kind of stock adjustment competition,the differentiation strategy may be both realistic and Effective strategy choices.In order to overcome the sample data problems that may be encountered in the process of developing a differentiated competitive strategy,this paper adopts a reverse thinking research method,which attempts to reveal the reasons for customer churn by examining the common characteristics of the lost customers,thus revealing the enterprise strategy and operation.The problems existing in the above,and the corresponding countermeasures are proposed from the strategic level,so as to achieve the purpose of stabilizing customers and attracting new customers.Specifically,taking B company as the research object,taking the broadband customer information in the B company's management analysis system database as sample data,using C5.0 decision tree,neural network and Logistic regression model to identify and analyze the customer status.On this basis,the optimized model-C5.0 decision tree is used to construct the early warning system for the lost customers,and the targeted customer retention strategy analysis is carried out according to the C5.0 decision tree early warning rules,so as to reduce the probability of customer churn.Enhance the appeal of external customers.The analysis results show that(1)the neural network model,Logistic model and C5.0 decision tree model are applicable to enterprise customer state discriminant analysis.In general,the three types of model judgment and prediction effects are effective,but in comparison,C5.0 decision tree The model has the best effect.The overall accuracy of the model is 82.7%.The proportion of users who are predicted to be broadband normal users and actually broadband normal users is 91.5%.The proportion of users who are predicted to be broadband loss users and actually lose broadband users is 66.5%.The hit rate is 66.5%.81%,coverage rate is 66.5%;(2)the most important factors affecting customer status selection include:billing mode name,broadband rate,brand,network access duration(month),whether to pay,active,billed income,etc.;(3)Characteristic broadband customers are more likely to become lost customers:billing mode name=monthly subscription;network access duration(month)<=146;whether credit=0;rate>100M;active=0;brand=enterprise broadband;=2.77 yuan;age between 20-30 years old;(4)effective measures to solve customer churn is based on the premise of classification,combined with the specific characteristics of the classification of customers to accurately policy.