| Customer | Plan | Risk | Recovery play |
|---|---|---|---|
| Sarah Kim | Pro | High | Send offer → |
| Mark Lowe | Basic | Low | Thank-you email → |
| Olivia Hunt | Premium | Medium | Check-in email → |
| James Avery | Premium | Medium | Add loyalty credit → |
| Ethan Miles | Basic | High | Call customer → |
Retain scores every account against your own usage, billing and support history, then runs the recovery play automatically — before the cancel button is ever found.
| Customer | Plan | Risk | Recovery play |
|---|---|---|---|
| Sarah Kim | Pro | High | Send offer → |
| Mark Lowe | Basic | Low | Thank-you email → |
| Olivia Hunt | Premium | Medium | Check-in email → |
| James Avery | Premium | Medium | Add loyalty credit → |
| Ethan Miles | Basic | High | Call customer → |
The platform
A score on a dashboard never saved an account. Retain connects the prediction to a play, runs it where the customer already is, and reports what the save was worth.
The model trains on your product events, invoices and tickets — not an industry average. It explains every score with the three signals that moved it.
Discount, check-in, onboarding nudge or a task for the CSM. Pick the play per risk band and Retain executes it across email, in-app and your CRM.
Median warning before a cancellation
Of at-risk revenue recovered in the first quarter
From connecting your data to the first scored cohort
Warehouse, Stripe, Postgres, Segment, HubSpot and webhooks. Nothing leaves your infrastructure if you self-host.
Send us a sample export and we'll return a scored cohort with the three drivers behind it, within two working days.