Benefits of the combined clustering and classification methods for better understanding of Telecom customers' needs and behavior

Complex and dynamic environment of Telecoms and their efforts to strive for in the fierce competition fosters operators to enhance their processes and to be more effective then their competitors. Creation of a concept of Customer Relationship Management (CRM) facilitates these processes supporting operator's holistic management of relationships with customers. Customer segmentation and identification of properties and characteristics inside particular segment like: profits, loyalty, quality of customer services etc. are only possible with suitable Data Warehouse-DW and Data Mining support. Therefore, the paper discusses the role of a suitable technical support like Data Warehouse, OLAP (OnLine Analytical Processing) and Data Mining in CRM [1] and presents a research on customers clustering using CDR data. With this research, analyzing and accurately classifying the individual clusters, based on different methods, and comparing them with distribution of the targeted variables for the whole ensemble, it was found out numerous, earlier hidden, information which contributed to detect some negative trends, insider fraud and potential churners. These results helped managers to develop creative marketing campaigns, improve the tariff policy and customer loyalty for the mutual benefit of the customers and its organization.