Machine learning for telecoms: From churn prediction to customer relationship management

The use of machine learning algorithms in churn prediction for telecom customers has shown its effectiveness in the last decades. Thus, uplift modeling, as a prescriptive analytic machine learning technique, may help telcos to reduce churn by setting up effective retention campaigns targeting only customers who are more suitable to be retained. In this paper, we compare three uplift models used in churn prediction contexts in three different sectors having many similarities to telecoms. Results gave some recommendations for building effective uplift models especially in terms of the quantity and quality of data sets. Therefore, the study also gave an empirical proof of uplift modeling being a good decision-making tool for customer relationship management.