About

This canvas analyzes Zerve's user upgrade behavior by engineering behavioral features and segmenting users into four cohorts (Dead Stars, Power Users, Hidden Gems, Dormant) using a novel Workflow Diversity Score (WDS), then trains a logistic regression model that achieves 99.1% accuracy in predicting paid conversions. The key insight: Dead Stars—users with high event volume but low feature diversity—are the only converting segment (1.82% upgrade rate), revealing that tight specialization combined with high engagement is the strongest signal for monetization readiness.