How much MRR do you actually lose to failed payments?
Last reviewed 2026-10-07
Cards expire, balances run dry, issuers decline, and a subscription quietly stops paying without anyone cancelling. Most founders have never put a number on it. This guide covers the two benchmarks people quote, why neither is your number, and how to work out your own from the Stripe account you already have.
The two numbers everyone quotes
Search for this topic and you will meet the same two figures everywhere, usually with no source attached. Here they are with their sources, their dates, and their problems.
| Claim | Who published it | When | Based on |
|---|---|---|---|
| Subscription businesses lose on average 9% of MRR to failed payments | Baremetrics | Published 3 June 2026, updated 10 July 2026 | A benchmark page whose figures come from 119 US B2B SaaS companies using Baremetrics Recover in May 2026 |
| Declined card payments account for 50% of all customer churn | PYMNTS Intelligence with FlexPay | 17 November 2023 | A survey of 200 subscription executives; the article attributes the 50% itself to market research |
Both are plausible. Neither is evidence about your account, and it is worth being precise about why.
Why the benchmarks are weaker than they look
The Baremetrics figure comes from a page built on their own Recover customers — 119 US B2B SaaS companies, 94% of them running Stripe, measured over a single month. These are businesses that noticed a failed-payment problem and bought a tool for it. A company losing a rounding error to declines does not show up in that sample, because it never became a customer. The direction of that bias is obvious and it runs upward.
The same page reports a median attempted recovery rate of 12.7% across that sample, which is a useful reality check in the other direction: even with a paid tool running, most of what fails in a given month is not recovered that month.
The PYMNTS figure has a different weakness. It comes from a survey of 200 executives rather than from transaction data, it is nearly three years old at the time of writing, and the article itself attributes the 50% to market research rather than to its own sample. The more interesting line in the same piece is that one failed payment in four is caused by something the customer did — which means three in four are not.
Our own position: quote these if you must, but treat them as an order of magnitude, not a measurement. A vendor publishing a number that makes its product look necessary is not lying, but it is not neutral either. We sell a recovery tool eventually, so apply the same scepticism to us.
What your own number is made of
Before measuring anything, it helps to separate three piles that often get added together and should not be.
- Permanently lost: invoices that failed and will not be paid. Written off, voided, or attached to a subscription that was cancelled while unpaid.
- Still being retried: invoices that failed but where Stripe has not given up. Some of this will arrive. Counting it as lost inflates your number.
- Recovered: invoices that failed at least once and were eventually paid. This is money your retry settings already saved, and it belongs in the win column, not the loss column.
A headline figure that merges all three is the most common way these numbers get exaggerated, including by people with no intention of exaggerating. Separating them is also what tells you whether to act: a large recovered pile means your retries work, a large still-retrying pile means wait, a large permanently-lost pile means something is wrong.
How to check your own number, for free
You do not need a tool for a first pass. Stripe already reports this, and the fastest honest answer is sitting in your dashboard.
- 1.Open Billing, then Revenue recovery, then the analytics view. Stripe shows payment failure rates, recovery rates, and the customers whose payments recently failed.
- 2.Note the window Stripe is showing you. A failure rate for last month answers a different question from a loss total for the last year.
- 3.Export your invoices for the last twelve months and filter to those with a failed attempt that are still unpaid. That is your permanently lost plus still retrying, before splitting them.
- 4.Check the decline codes on those failures. Expired and incorrect cards call for a different fix than insufficient funds, and hard declines call for no fix at all.
What Stripe's analytics page does not describe is a breakdown of losses by decline family, which is the step that turns a number into a decision. That is the gap our audit fills, and it is the only reason we built it.
What to do once you have the number
If the number is small, stop. You have spent twenty minutes and bought yourself the right to ignore an entire category of advice. That is a good outcome and nobody will write a blog post about it.
If it is large, the split tells you what to do next. Expired and incorrect cards are recovered by getting the customer to enter a new one, which means emails and an update page, not more retries. Insufficient funds are recovered by retrying at a better moment, which Stripe's Smart Retries already attempts. Hard declines — lost, stolen, fraudulent, do-not-honour — are not recovered at all, and time spent chasing them is time wasted.
The first action, in nearly every case, is free: switch on Stripe's automatic customer emails if you have not. A failed payment nobody mentioned to the customer is the cheapest recovery available, and a surprising number of accounts have it turned off.
Questions people ask
- What percentage of MRR do SaaS companies lose to failed payments?
- Baremetrics publishes an average of 9% of monthly recurring revenue, on a page updated 10 July 2026 whose benchmark figures come from 119 US B2B SaaS companies using its own recovery product. Treat it as an upper-leaning order of magnitude: the sample is made of companies that had already bought a tool for the problem.
- How much of churn is caused by failed payments?
- A PYMNTS Intelligence study with FlexPay, published 17 November 2023, reports that declined card payments account for 50% of all customer churn. It is a survey of 200 executives rather than transaction data, and the article attributes the figure to market research rather than its own sample.
- Is involuntary churn the same as failed payments?
- Close, but not identical. Involuntary churn is the subscription ending without the customer choosing to leave, usually because payment failed and never recovered. A failed payment that is later paid is not churn at all. That is why the recovered pile has to be kept separate.
- Can I see this in Stripe without any extra tool?
- Yes. Stripe's revenue recovery analytics show failure rates, recovery rates and recently failed customers. What it does not document is a breakdown by decline family, so you will have to read the decline codes on the failed invoices yourself to know which failures are worth chasing.
Sources
- Baremetrics — Subscription Payment Recovery Benchmarks (2026) — published 3 June 2026, updated 10 July 2026, checked 7 October 2026
- PYMNTS — Half of Subscription Cancellations Are Due to Failed Card Payments — published 17 November 2023, checked 7 October 2026
- Stripe Docs — Revenue recovery overview — checked 7 October 2026