Customer Demographic Segmentation Based On Telecom Behavioral Data

Abstract

In the modern world, digitalization becomes ubiquitous and covers almost every aspect of the business and daily life. Telecom services providers have a major role in these processes due to their involvement in collecting, storing and processing enormous amounts of customer data. This also includes personal telecom services usage data, which if correctly interpreted, might be used for many different purposes. Using telecom data to predict certain demographic characteristics of the customers is helpful in more than one aspect: 1) It could add the acquired knowledge into customer segmentation to better target different customer groups. 2) Such data could be used in cases where traditional historic data is not available- the potential strength of predicting customer credit worthiness based on behavior data is still not fully explored. 3) Last but definitely not least, is the use of data for verifying customer identification in fraud detection. In this paper, an overview of some successful use of telecom data for non-telecom services is shown, as well as with a set of real telco data, statistical techniques are used to demonstrate the relation between mobile telecom services usage and subscription owners’ age. Use of alternative customer data could have enormous implication both on traditional predictive models and could alter the role of the telecoms, making them one of the most important information sources for financial institutions, which operate with sensitive customer data.