The Management And Application Of Customer Value Based On Data Mining

With the increasing competition of market, enterprises’operation ideas transferfrom the“product”to“customer”, the customer is becoming the competing resourcesof the companies. Enterprises are getting more important for maintain and developrelationships with customers. Customer Value Management, as a core part of theCustomer Relationship Management, has become the very focus of enterprises. Thepremise of effective CVM is that enterprise can clearly identify the customer.However, with the deepening application of business system, enterprises haveaccumulated a large amount of customer data. These data contains lots of knowledgeand information, how to tap these information effectively to provide correct supportfor enterprises decision making? With the growing IT innovation, the advantages ofthe new data processing and analysis methods are increasingly embodied. Therefore,the application of data mining provides a new method for CVM.The paper, firstly, reviewed and commented in detail previous studies oncustomer value, customer life value and customer value management, defined theconcept of customer value and customer value management, relying on the customervalue correlation theories, deeply analyzed the sources of customer value,classification and creation mode. Secondly, according to the lack of the customervalue management theories and the lack of the customer value evaluation method,construct the fundamental hierarchy framework of CVM (the source layer, the valueanalysis layer, the data mining layer, the results layer and the application layer),analyzed the contents of CVM ( width management, length management and depthmanagement), established the customer value assessment system (historical value,current value and potential value), also gave the method for customer valuecalculation. Subsequently, to the core part of width management—customersegmentation, discussed the deficiencies of the traditional customer segmentationmethods, put forward the multi-indicator RFM method based on the study of previousresearches, using the method of factor analysis to identify potential confoundingeffects, combined with the clustering algorithm in data mining technology to segmentcustomers; for the customer retention and win back customer in length management,combined with the decision tree to construct customer churn models. Provides a basisfor customer retention and win back; for the cross-buying in depth management, combined with the Apriori algorithm, obtained the association rules of customerscross-buying, provide a basis for the depth management of CVM. Finally, using thecustomer data of a large paper enterprise to verify the width, length and depth ofCVM, the result shows that using the data mining technology can effectively carry outthe customer value management.