CHURN PREDICTION MODEL IN A MEDIA BROADCASTING COMPANY

Created a Churn Prediction model (Machine Learning - Classification) that predicts the likelihood of customers to leave the subscription service in the following 1 to 3 months. The model reached an overall accuracy > 90% and a precision in identifying Customers that were going to churn > 50%, giving the opportunity to CRM, Retention & Campaign team to be very targeted in sending marketing campaigns only to customer really willing to churn. This maximized profitability of Prevention, Retention, Upselling and Downgrade Marketing strategies.