tsmg

A leading Indian telecom company having countrywide market penetration wanted to develop and implement segment wise customer strategy.

The challenge was to identify repeatable patterns in customer behavior using call transactions data for post-paid subscribers, create homogeneous groups of customers on the basis of these patterns. 

One of the roadblocks faced during the project was poor quality or almost no availability of customer demographic data. Tata Strategic developed an innovative solution by creating segmentation model applying clustering techniques on historical calling data and billing data for a representative cohort of customers. 

These customers were grouped into homogeneous segments using Ward’s clustering methodology; the clustering parameters being call frequency, call duration, type of call, usage of value added services, SMSes etc. Descriptive statistics for different parameters were evaluated across the time of day and the day of week. The exercise also included the identification of statistical distributions and the hourly patterns of the variables call duration, frequency of calls and type of calls. Following this study, cluster specific predictive models were developed. 

This exercise helped the telecom company in establishing the mapping or alignment of new customers to appropriate segments after an initial tracking period of their calling pattern. This in turn facilitated the deployment of most effective strategy for a segment on a proactive basis.