MasterControl flags customer churn 90 days early with 85% accuracy

The workflow before

MasterControl's customer success team judged account health with simple rules of thumb. Product usage, support tickets, contract and payment history, and training data sat in separate systems, so building a view of any one account was manual work. Warning signs were usually spotted close to renewal, when there was little time left to act.

What we built

Over 12 weeks we built a churn prediction model on AWS SageMaker trained on 3 years of customer history. It combines product usage trends, support ticket patterns, contract and payment history, firmographics and training completion into one score per account, predicting churn within a 90-day window.

The model runs behind a SageMaker endpoint and feeds a dashboard for the customer success team, with SHAP explanations showing why each account is flagged. We paired it with intervention playbooks matched to the main risk drivers: falling engagement, rising ticket volume, low adoption of advanced features, contract reductions and declining training completion.

Before and after

What's next

Next is pushing scores and alerts into Salesforce so account managers act inside their daily workflow, and retraining on intervention outcomes. Protecting even a few points of churn on $200M of recurring revenue is worth about $2M a year.