A production audit of a live Next.js SaaS uncovered systematic billing and analytics failures. MRR was computed as subscriber count times a hardcoded price, ignoring team seats, annual plans, and provider-side adjustments. Stripe invoice webhooks silently failed to resolve subscriptions because the payload structure changed across API versions. PayPal renewals refreshed billing dates only for past-due subscriptions, letting active ones drift. Test accounts polluted subscriber counts, MRR, churn, and conversion across 10+ admin surfaces that had no shared exclusion function.

A methodology for building subscription analytics systems that goes past basic MRR tracking. Covers the full lifecycle: immutable payment event ledgers, subscription state machines with multi-provider tie-breaking, 4-factor behavioral health scoring (0-100), 19-driver churn prediction via logistic regression, Monte Carlo revenue simulation with P10/P50/P90 fan charts, automated retention interventions with Bayesian adaptive thresholds, and a second tier of advanced mathematical methods including survival analysis, conformal prediction, causal inference, and optimal experimental design. Include