A major online music subscriber company wanted to predict the likelihood of churn among its subscribers to identify “at risk” subscribers.

• 5 member ADT team worked closely with client team for 2 months to build the model.
• An exhaustive set of variables were tried and the model was rigorously tested over large volumes of data
• Well-Structured project execution plan & strict adherence to guidelines & timelines were maintained.
• Effective & efficient communication was built with client to constantly provide and receive feedbacks for smooth execution of projects.
• ADT had developed a churn prediction model that was also leveraged to estimate the Life Time Margin (LTM) of subscribers.
• A complex and robust technique “Survival Analysis” was used to predict churn amongst subscribers
• An exhaustive list of variables and variable transformations were tried in the model.

The model predicted 50 pc of the churners in the 1st three deciles consistently and which helped the client take proactive actions to save their “at risk” subscribers.