Aggregate churn rate is a population statistic. Churn risk is the per-customer version — the probability for one specific subscriber. Knowing the population rate tells you how big the problem is; knowing per-customer risk tells you who to call this morning.
How churn risk gets scored
- Behavioral signals — drop in email opens, no recent portal logins, multiple consecutive skips.
- Subscription actions — pause requests, frequency downgrades, swap requests, complaints in support tickets.
- Payment health — recent failed charge, expiring card, hit credit limit.
- Tenure and cohort — first-90-days customers carry inherently higher risk; long-tenured loyalists carry lower.
Most subscription analytics tools express this as a low / medium / high tag or a 0–100 score. The exact number matters less than the relative ranking — you want a list, sorted by risk, to act on.
What to do with high-risk customers
- Email a save offer. A discount, a swap, a pause option — give the customer flexibility before they cancel.
- Trigger a personal outreach. For higher-value subscribers, a personal email from a success manager beats automation.
- Adjust their plan. Suggest a lower-frequency cadence or smaller pack size based on their consumption signals.
- Capture feedback. A one-question survey ("Is everything okay with your subscription?") opens the conversation before the cancel button does.
Avoid two common mistakes
First, do not over-intervene on customers who are simply paused or on a slow cadence — making them feel watched is a fast way to push them to cancel. Second, do not assume the highest-risk customers are the most valuable to save. Some are best released (price-sensitive churners, bad-fit customers) so the team can focus on saving the ones with strong LTV potential. See churn prediction model for the modeling side and win-back campaigns for what happens after.