Designed a 5-table relational database and conducted a 21-step exploratory analysis on synthetic SaaS customer data. Identified key pre-churn behavioral patterns using logistic regression; discovered 90-day support ticket drop as the strongest churn predictor — explained through Learned Helplessness theory. Built 5-dashboard Power BI report including a Customer Success Action Center, flagging 366 at-risk accounts with personalized intervention recommendations. Key insight: 85% of churn occurs in Year 1 — onboarding optimization is the highest-ROI intervention.