Customer Churn Prediction Model Reduced Churn by 29%
The Challenge
Windstream was experiencing 4.2% monthly churn on their business broadband segment — above industry average and trending upward. Their retention team was entirely reactive, calling customers only after they had already submitted cancellation requests. Win-back rates were under 20%.
Our Solution
Block Logic built a customer churn prediction model using 90 days of behavioral, usage, support, and billing data across 47 input features. The model identifies at-risk customers 45 days before likely cancellation with 79% precision. The output feeds into a Salesforce workflow that triggers personalized retention offers, prioritizes outreach by predicted lifetime value, and tracks outcomes to continuously improve model accuracy.
Results Delivered
Monthly churn on the target segment dropped 29% within 6 months. The model identifies 340+ at-risk customers per month. Annual revenue retained attributable to the model: $3.4M.
"We went from chasing cancellations to preventing them. The model identified behavioral patterns our entire team had completely missed."
Ready to Be Our Next Success Story?
Every engagement starts with a conversation. Let's explore what we can achieve together.