Information providers in Internet are facing a highly competitive environment and a lack of insight into the customer's current and potential value and the customer's behaviour and requirements. Customer Relationship Management (CRM) is a customer-focused business strategy designed to optimise profitability, revenue and customer satisfaction that enables to develop valuable customer relationships and create a competitive position. Most of its success, however, relies on the ability to properly segment and profile customers for differentiated product offerings and service levels. We address the need of an information provider that offers a subscription-based access to Scientific and Technological databases. The service collects, for accounting purposes, data on system usage through an application server. These data are much richer than web logs data usually collected by web servers and permit the recognition of the single events that take place inside a user session. We show how these data can be successfully mined to extract knowledge on user behaviour. We apply a clustering algorithm to the accounting data, summarized at the user level, to obtain a customer segmentation. The results show the existence of different strategies of usage of the service resources. We show how these segmentation results serve to focus the attention on specific set of customers to understand and address their common needs, opportunities or threats. The obtained user profiles provide the basis for targeted marketing campaigns, customer retention and service improvement. Finally the data mining results are integrated in a wider process of Customer Relationship Management already implemented by the information provider.
展开▼
