The 3-signal churn score that predicts cancellations 12 days early
Most SaaS founders find out a customer is churning the day they cancel. By then it's too late to save them.
Here's a simple scoring model we use to flag at-risk accounts before they hit the cancel button:
The 3 signals that matter most:
Login frequency drop — compare last 14 days of logins vs. the prior 30-day average. A drop of 50%+ is signal #1.
Core action decay — track the ONE action that correlates with retention in your product (e.g. messages sent, reports generated, tasks completed). A 2-week decline flags the account.
Support sentiment — any support ticket tagged "confused," "not working," or "too expensive" in the last 30 days adds risk weight.
Scoring it:
Each signal = 1 point if triggered.
Score of 2+ = "at risk" → trigger a save sequence (personal check-in email + usage tips, NOT a generic discount blast).
Score of 3 = "critical" → founder/CS lead reaches out personally within 24 hours.
Teams running this simple 3-point model catch 60-70% of cancellations while there's still time to intervene — instead of finding out via a cancellation email.
Start simple: pull login timestamps and your one core action from your database, run the math in a spreadsheet weekly. You don't need fancy tooling to start saving revenue — you need the right signals and a fast response process.
