Internet technology enables companies to capture new customers, track their performances and online behavior, and customize communications, products, services, and prices. Analyses of customers and customer interactions for electronic customer relationship management (e-CRM) can be performed by way of using data mining (DM), optimization methods, or combined approaches. One key issue in the analysis of access patterns on the Web is the clustering and classification of Web documents. Generally, the classification has its base on analytical models which assume a pre-fixed set of keywords (attributes) with predefined list of categories. This assumption is not realistic for large and evolving collections of documents such as World Wide Web. We propose a new approach to solve the problem of unknown number of evolving categories. The approach begins with the classification of test documents into a set of initial categories. A working prototype system which is based on Fuzzy Clustering CRM (FC-CRM) has been developed and presented to validate the proposed approach and illustrate how it handles the dynamic inflow of new documents. 2008 Elsevier B.V. All rights reserved.