Distance to churn predictive solution seeks to identify the expected time of churn in an open-ended timeline. This helps in developing strategies with long time foresight on when to retain the customers proactively. Also, helps to retain ‘valued’ customers by generating a long-term foresight on the possibility of churn. The data that is captured for the solution comprises customer details, behavioral data and POS/Campaign data.
The initial phase involves value based segmentation and identification of KPIs and then deriving a survival model to identify triggers/indicators of silent churn.
The Churn Summary calls out active customer base vis-à-vis churners based on dimensions like month wise, card type, promo codes and point-of-sale. The transaction trends depict the customer status, time to event, card type, POS and promo code based filtration to observe what the trend looks like and analyze the cause of peaks and valleys.
The predictive modelling allows identification of churn based on risk category, predicted month of churn, card type, balance amount, promo codes, changes in amounts quarter wise, transaction count in specific quarters which are indeed the significant variables impacting churn as spun out by the model employed. The three process steps involved are – Data Analysis : The descriptive analytics on customers’ demographics, transaction and derived data capturing the changes in consumers’ behavioral trend, Base segmentation : customer segmentation based on their value will help reducing generalization on the solution and predictive modelling : survival model equation and ‘distance to churn’ for proactive retention of HVCs most efficiently. The business impact thus achieved is foresight into ‘time of churn’ for high value customers to help proactive retention and target them who are more responsive.