Research On B2C E-commerce Customer Segmentation Based On Two Step Clustering

In recent years, the Internet and e-commerce has been developing rapidly, B2C e-commerce businesses highlights the fierce competition between the significant role of enterprise customers, is the enterprise to obtain and maintain a competitive advantage of important factors. Companies focus gradually shifted, companies will no longer targeting the product to customers, greater efforts into the development of potential customers and maintain current customers, using differentiated, personalized marketing tools to improve customer satisfaction, thus increasing the profitability of enterprises, and the realization of personalized precision marketing premise is effective customer segmentation. The effectiveness depends on customer segmentation models and methods used in science and rationality. Average customer segmentation mainly through the customer’s one-dimensional attributes, customers something of value, RFM model, a combination of factors such as simple taxonomy, it is difficult to deep segmenting customers. Too simple and rough models and methods will hardly meet the increasingly diverse business requirements and customer needs, customer segmentation applied reasonable and appropriate customer segmentation techniques produce urgently needed, and good at dealing with high-dimensional, large data sets of data mining techniques to be shipped born and widely used. Data mining techniques can be as high accuracy, high efficient processing complex models and data, and thus the customer base for accurate classification and description of the types of clients fully reflect the comprehensive features, corporate marketing departments and meticulous detail, which can understand and analyze the customer’s characteristics can be achieved even on the customer’s dynamic tracking, then made effective marketing strategies and plans to appropriate, attentive service firmly seize the customer, to achieve the ultimate goal.After elaborating on customer segmentation theory and the analysis of the basis for its segment, methods, requirements, limitations, and in the B2C e-commerce environment changes, based on the statistical properties of population, psychological characteristic attributes, consumer behavior and customer value and the impact of four aspects of the network, the paper established a B2C e-commerce customer segmentation index system for cluster analysis reference when selecting indicators.In addition, the paper selected cluster analysis as the method for B2C e-commerce customer segmentation,the cluster analysis principles, characteristics and divided to do a more fully described and summarized, focusing on the K-means algorithm, SL hierarchical clustering algorithm, DBSCAN algorithm and two-step hierarchical clustering algorithm, the various algorithms described in more detail, the use of simulation were compared in several ways comparison results are given, and ultimately determine the two-step hierarchical clustering algorithm as the completion of B2C e-Business customer segmentation process approach.Finally, the two-step hierarchical clustering algorithm in the field of B2C e-commerce customer segmentation application to get data from a Web site, after data cleaning and data transform data preparation process, combined with the index system and the actual situation, to determine the final clustering participation analysis of customer targets, using IBM’s Clementine software to complete a two-step hierarchical clustering algorithm clustering of data, statistical information on the clustering results and results analysis? and gives some suggestions.