About

This canvas is a comprehensive churn prediction and upgrade propensity ML pipeline that analyzes 3.5M user events from a SaaS product to identify success signals and build predictive models for user engagement and upgrade likelihood. The workflow ingests event data, performs feature engineering at the user level, trains leakage-aware success and early-warning models, segments users into behavioral cohorts (Champions, Power Users, AI Adopters, Deployers, Explorers, Casual, Churners), and generates upgrade propensity scores with actionable targeting recommendations for product teams.