Practical Methods for Predicting Customer Retention

This study examines methods for analyzing and forecasting the retention of active subscribers in the telecommunications industry using various criteria for subscriber activity. The results demonstrate that the retention dynamics of an active subscriber base can be effectively modeled using a decreasing power function. This allows for medium-term forecasting based on initial subscriber activity data. However, it is important to note the potential limitations in the effectiveness of the proposed approach for long-term forecasting, associated with changes in subscriber churn dynamics over time.