Each active subscriber carries a churn score, refreshed on the hourly maintenance pass, and a band: high, medium or low. High-risk subscribers are listed on your dashboard, and high and medium subscribers are listed under Retention AI, worst first, with the factors that drove the score.
Where to find it
On the app's Overview under At-risk subscribers, under Retention AI → Spot who might leave, and as an At risk badge and Risk signals on each row in Subscribers.
Options
The score is worked out from real signals on your store, not from an AI model:
| Factor | What it means |
|---|---|
| Subscription is paused | A paused subscriber is closer to canceling than an active one. |
| Failed payments | One or more billing attempts failed. |
| Payment recovery exhausted | The retry ladder ran out without a successful charge. |
| Only the first order so far | New subscribers are the most fragile. |
| Next billing date is overdue | The next charge is more than two days late. |
Example
What to do with it
- Keep your best save offer ready for the cancel flow, so a subscriber on the way out sees a reason to stay.
- Reach out before the cancel, not after. A skipped delivery or a paused month keeps a subscriber that a cancellation loses.
- Read the factors. A list of at-risk subscribers who all failed a payment last cycle is a dunning problem, not a product problem.
Plan required
Related articles