Indie SaaS Retention — Churn Analysis, Email Automation, and Habit-Forming UX

"Users are signing up but MRR isn't growing." "People vanish after the trial." These are retention problems, and they're far more common in indie SaaS than acquisition problems. This article covers a practical retention stack built on Supabase, Resend, and Flutter.

Defining Churn Prediction Signals

Churn doesn't happen suddenly — there are leading indicators. The most reliable signals:

Signal Threshold (example)

Login frequency drop < 1 login in the past 7 days

Core feature inactivity 0 uses of the primary feature in 30 days

Support ticket spike 3+ tickets in 14 days

Incomplete onboarding Profile completion < 50% after 3 days

Track these in a user_events table and compute a risk score automatically on a schedule.

Auto-detecting At-Risk Users with Supabase + pg_cron

-- View: users with no login in the past 7 days create or replace view churn_risk_users as select u . id , u . email , u .…