What is MRR at risk, and how to calculate it
MRR at risk is the recurring revenue from customers showing signs they're about to leave. Here's a working definition, how to calculate it, and why it's a better weekly number than last month's churn.
Author
Vinicius Aguiar
Published
Reading time
3 min read
Short answer: MRR at risk is the monthly recurring revenue from customers who haven't cancelled yet but show the signs that usually come before a cancellation — a failed payment, a scheduled cancellation, a downgrade, usage that dropped, or no usage at all. Churn tells you what you lost last month. MRR at risk tells you what you can still save this month.
A working definition
There's no standard formula, so be explicit about yours. A practical one:
MRR at risk = Σ MRR of paying customers with at least one risk signal
Count each customer once, even if they have several signals, and use their current MRR — what they pay per month today, with annual plans divided by 12.
Which signals to count
| — | Source | Strength |
|---|---|---|
| Scheduled cancellation | Billing (cancel at period end) | Very strong — the decision is made |
| Failed payment | Billing (past_due) | Strong, but often recoverable |
| Never active | Product usage | Strong — they never got value |
| Silent for weeks | Product usage | Strong for products used weekly |
| Usage dropped sharply | Product usage | Medium — could be a slow week |
| Downgrade | Billing | Medium — contraction, often before churn |
Billing signals are easy to get and arrive late. Usage signals take a join between your billing and your analytics, and arrive weeks earlier. A version of MRR at risk built only from billing mostly measures churn that is already happening.
How to calculate it, step by step
Step 1: Take every paying customer and their current MRR
From your billing system, include subscriptions that are active, trialing with a card, or past due. Normalize every price to a monthly amount.
Step 2: Attach the signals
Mark each customer with the billing signals from the table above. Then join usage from your product analytics — this week's activity, last week's, and the last date seen — and mark the usage signals.
Step 3: Sum, and keep the breakdown
Add up the MRR of every customer with at least one signal. Keep the list behind the number: the total is for the weekly meeting, the list is for the people making calls.
Step 4: Divide by total MRR
MRR at risk % = MRR at risk ÷ total MRR
The percentage lets you compare weeks while the business grows.
Why it's a better weekly number than churn
Churn rate is a lagging indicator. By the time it moves, the customers are gone, and the only thing left to do is explain it.
MRR at risk moves first. When it rises, you know revenue is about to leave, and you know which accounts it's coming from. It turns churn from a number you report into a list you work through — ordered, if you sort it by MRR, by what each call is worth.
A worked example
A business with $50,000 MRR and 200 customers finds, this week:
- 3 customers with a scheduled cancellation, $1,800 MRR in total
- 5 with a failed payment, $1,100
- 9 paying but never active, $2,400
- 4 whose usage dropped sharply, $3,600
MRR at risk is $8,900, or 17.8% of MRR. The biggest single slice isn't in billing at all: it's the $2,400 from customers who pay and never used the product, plus the $3,600 whose usage is fading.
Tracking it without a spreadsheet
Keep calculates this continuously: it reads Stripe for revenue and billing signals, reads PostHog or an activity API for usage, and alerts in Slack on each at-risk account with the MRR at stake — ordered by that number.
FAQ
Is MRR at risk the same as churn forecast?
No. A forecast estimates how much will churn. MRR at risk is the revenue attached to customers showing risk signals today; some will be saved, some will leave. It's a work list, not a prediction.
Should failed payments count toward MRR at risk?
Yes, but track them separately. Failed payments are often recovered automatically by retries, so they behave differently from customers who stopped using the product.
How often should I calculate it?
Weekly is a good rhythm for most teams, with alerts for individual accounts as the signals appear.
What's a normal MRR at risk percentage?
There isn't a reliable public benchmark, and it depends heavily on which signals you count. Track your own trend week over week instead of comparing with other companies.
Read next
How to detect churn risk from Stripe data
What your Stripe account can tell you about churn before it happens, what it can't, and a step-by-step way to build an at-risk list — with or without product usage data.
Voluntary vs involuntary churn: what's the difference, and which to fix first
Involuntary churn is a payment that failed. Voluntary churn is a customer who decided to leave. They have different causes, different signals, and different fixes.
"Pays and never used" is worse than "usage dropped"
Every churn dashboard is built to catch a drop. The most dangerous customer on your books has no drop to catch.