Abstract—This paper focuses on the application of a machine learning algorithms such as the logistic regression, to firstly, derive insights from the data to identify the factors that drive churn, secondly to identify which customers are highly likely to churn and their probability of churn and thirdly, to develop a retention strategy that reduced churn and maximized the revenue of the company. The final chosen model was the logistic regression model with a high churn prediction accuracy of 75.3%. The top five significant variables for driving churn was “FiberOptic”, “MonthtoMonthContract”, “DSL”, “OneYearContract” and “StreamingMovies”. By implementing the retention strategy, the business was able to reduce the churn rate by 40% and more than double their overall profit. By spending $88,000, the retention strategy was able to retain a revenue of $893,908.50, which is ten times more than the amount of money spent on the retention strategy.
Index Terms—Machine learning algorithms, churn, logistic regression, retention strategy, factors, churn reduction, revenue maximization.
Carol Anne Hargreaves is with the Department of Statistics and Applied Probability, National University of Singapore, Singapore (e-mail: stacah@nus.edu.sg).
[PDF]
Cite: Carol Anne Hargreaves, "A Machine Learning Algorithm for Churn Reduction & Revenue Maximization: An Application in the Telecommunication Industry," International Journal of Future Computer and Communication vol. 8, no. 4, pp. 109-113, 2019.
