---
title: "Involuntary Churn: Recovering Failed Payments | MarginFix"
description: "How to measure involuntary churn from your billing exports, recover failed payments in the right order, and value each saved subscriber against CAC."
url: https://marginfix.ai/insights/involuntary-churn-failed-payments/
lastmod: 2026-10-13
---

[Home](/) › [Insights](/insights/) › Involuntary churn: how failed payments leak subscription revenue, and how to win it back

Field guide · 12 min read

# Involuntary churn: how failed payments leak subscription revenue, and how to win it back

![Portrait photograph of Cristian Bragau](/assets/founder/cristian-bragau-2026-10.jpg) **[Cristian Bragau](/about/cristian-bragau/)**, Founder & Lead Auditor · Reviewed Oct 2026 · verified against the base of 74 audits

The short answer

Involuntary churn is the subscribers you lose because a payment failed, not because they chose to leave. Measure it from your billing exports: split ended subscriptions by reason and sort failed charges by decline code. Then recover it in order: card updater, retry schedule, dunning emails, in app prompts, outreach before cards expire, and a wider payment method mix.

## Why failed payments hide inside your churn rate

A subscriber whose card expired did not decide to leave, yet a single churn rate counts them the same way as a subscriber who clicked cancel. They also call for different work: a cancellation points at the product or the price, while a failed payment points at how billing handles a declined card. That workflow is called [dunning](/glossary/dunning/), and its definition and the basic involuntary churn formula live on that page. This guide is the how to: measure your involuntary share from your own exports, fix it in the right order, and value what you keep.

Retention leaks showed up in 41% of the 74 marketing and margin audits MarginFix has published, and leaky retention took 7% of all audited marketing spend. Failed payments are one part of that leak. In case 52, [Weak payment recovery was leaking revenue as involuntary churn](/cases/weak-payment-recovery-was-leaking-revenue-as-involuntary-chu/), a subscription business let subscribers it could have kept lapse instead of retrying their payments; the audit identified $52k a year in margin, with findings in 5 days.

Two separate figures from the 74 published audits. They measure different things and do not combine.

| Measure | Figure |
| --- | --- |
| Audits where a retention leak was found | 41% |
| Share of audited marketing spend that went to leaky retention | 7% |

Outside our audits, one public figure comes from a payment processor. Stripe wrote in January 2024 that a quarter of lapsed subscriptions are due purely to payment failures. Stripe does not publish the method behind that number, so treat it as one processor’s figure, not a benchmark for your business. Your own share is sitting in your billing exports.

## Step 1: Export ended subscriptions and failed charges

You need two exports from your billing system for the last full quarter. The first lists every subscription that ended, with its end date, plan, price and the reason it ended. The second lists every failed renewal charge, with its date, amount, decline code and whether a later attempt paid it. In Stripe, for example, a subscription’s cancellation reason reads payment_failed when it was canceled automatically after a failed payment, and cancellation_requested when it was canceled on request. Other billing systems keep an equivalent field or status.

Check your settings before you count. Once retries run out, Stripe cancels the subscription, marks it unpaid, leaves it past due or pauses it, depending on your settings, so a business that does not cancel will not find those subscribers among its cancellations at all. Count subscriptions left unpaid, past due beyond your retry window, or paused after a failed payment as involuntary losses until they pay again.

## Step 2: Compute the involuntary share and the recovery rate

Three numbers come out of those exports. The involuntary share is the subscriptions that ended after a failed payment nobody recovered, divided by all subscriptions that ended. The first attempt failure rate is the renewals that failed on the first try, divided by all renewals due. The recovery rate is the failed renewals that a later retry, email or card update paid, divided by all failed renewals. Compute each by count and by dollars: one failed annual renewal can weigh as much as a year of monthly ones, so the two can tell different stories.

Leave the latest month or two out of the recovery rate. Stripe’s recovery analytics notes that the rate can dip for the current and previous month while automatic retries are still running, so a recent month looks worse than it will end up.

**Worked example (an example, not a client figure):** say a business ends a quarter with 1,200 subscriptions gone. 840 were canceled on request and 360 ended after a failed payment that was never recovered, so the involuntary share is 30%. In the same quarter 1,500 renewals failed on the first attempt and 1,140 were paid later, a recovery rate of 76%. The 360 unpaid renewals are the pool this guide works on. The 840 cancellations need different fixes, such as the pause flow in case 53, [The pause and skip flow was leaking straight to cancellation](/cases/the-pause-and-skip-flow-was-leaking-straight-to-cancellation/).

## Step 3: Sort the failures by what fixes them

Each failed charge carries a decline code from the card issuer, and Stripe adds an advice code that says whether to try again. Sort the failures into four groups, because each group has its own fix. Expect a large generic group: Stripe notes that card issuers classify most declines as generic, which is why the advice code tells you more than the reason text.

Failed payments by what fixes them. Example codes are Stripe’s; other processors name them differently.

| Group | Example codes | What fixes it |
| --- | --- | --- |
| Card expired or replaced | expired_card | The card updater, then a prompt to the subscriber |
| Temporary decline | insufficient_funds, or advice code try_again_later | A retry a few days later |
| Do not retry | stolen_card, lost_card, pickup_card, or advice code do_not_try_again | A new payment method from the subscriber |
| Subscriber must act | authentication_required, incorrect_number | An email or in app prompt with a direct link |

## The six recovery levers, in order

Work through the levers in this order. The first two need nothing from the subscriber. The next three ask the subscriber to act, so they only work on what the first two missed. The last one changes what can fail in the first place. Check whether your billing system splits recovered payments by method, as Stripe’s analytics do (retries, emails and other), so you can see what each lever adds.

## Lever 1: Turn on the card updater

Card networks run updater services that pass a reissued card’s new number or expiry date to the businesses that store it. Visa describes its account updater as a way to keep stored credentials current and prevent avoidable declines, and Mastercard’s automatic billing updater keeps a global repository of card numbers and expiry dates for registered merchants. Stripe said in February 2023 that about 40% of cardholders had replaced a card in the prior year, and that most people do not update the card details they have stored with businesses. Coverage is not universal: Stripe says automatic updates are widely supported for cards issued in the United States, vary by country elsewhere, and depend on the issuing bank taking part. Confirm the updater is on with your processor, then count how many saved cards it changed last quarter.

## Lever 2: Retry on a schedule, not every day

Retry temporary declines over days, not hours. Stripe notes that card networks limit how many times you can reattempt a single charge, recommends at most eight retries for charges that allow them, and warns that issuers may read extra retries as possible fraud. Its own default is 8 tries within 2 weeks, timed by a model, and Stripe wrote in January 2024 that “The best time to retry many failed subscription payments is days into the future.” Do not retry hard declines: Stripe will not run those retries until the subscriber adds a new payment method, so they belong to the next three levers. This is the lever that was too lax in case 52, where the flow let subscribers lapse instead of retrying them.

## Lever 3: Send a dunning email after every failure

Send an email after each failed payment, not once. Stripe’s setting does exactly that: it emails the subscriber after each failure, says the payment failed, and lets them update the payment method so the charge can be retried. Make the link open the form that updates the card, name the plan at risk and the date access ends, and send it from a name the subscriber knows. Keep the reason neutral: Stripe advises presenting a lost or stolen card decline to the customer as a generic decline. Emails matter most for hard declines and authentication failures, where no retry can succeed without the subscriber.

## Lever 4: Prompt inside your product

A subscriber who still logs in during the retry window is easier to reach in your product than in an inbox. Stripe’s guidance for declines when the customer is not present is to notify them by email or by a notification in your app so they come back to your site or app. Show a banner with the update form on their next visit, and remove it the moment the payment succeeds.

## Lever 5: Reach out before the card expires

The cheapest failure is the one that never happens. Stripe can email subscribers one month before the card on file expires and send reminders before a renewal date, and its analytics guide suggests reminders ahead of payment dates when the failure rate is high. Start with annual plans: one failed annual renewal puts a full year of revenue on a single charge.

## Lever 6: Widen the payment method mix and choose the end state

A second card on file does not help by itself: Stripe retries on the first available payment method in a fixed order, so the subscriber’s default has to change. Bank debits have no card to expire, but they still fail for insufficient funds, and Stripe does not retry them by default; with retries turned on, ACH Direct Debit is retried only for insufficient funds, at most twice within 40 days. Then choose what happens when every lever has failed. Stripe lets you cancel the subscription, mark it unpaid, leave it past due or, where the subscription is eligible, pause it. A paused subscription generates no invoices until it resumes, which keeps the door open without billing someone who is not paying. For B2B accounts that pay by invoice rather than a card on file, an unpaid invoice is a collections question; the [B2B SaaS audit page](/industries/b2b-saas/) shows how billing is read apart from retention.

## What a saved subscriber is worth against CAC

A subscriber you keep after a failed payment is worth more than a new subscriber bought at the same [CAC](/glossary/cac/), because their acquisition cost is already paid. Their contribution margin counts from the first month you keep them, less the cost of the retries and emails. A new subscriber first has to repay what it cost to win them, which is the [payback period](/glossary/payback-period/).

Stripe reported in November 2024 that, in its data, a monthly subscription saved from involuntary churn typically lasts another seven months on average. Use your own cohorts when you have them; the example below borrows Stripe’s figure only to show the arithmetic.

**Worked example (an example, not a client figure):** say a $50 monthly plan has a 60% contribution margin, so each subscriber brings in $30 of margin a month. A saved subscriber who stays seven more months is worth $210 of contribution margin, less what the retries and emails cost. A new subscriber at a CAC of $180 brings in the same $30 a month, but the first six months only repay the $180, so over the same seven months the new one adds $30. If a better flow saved a third of the 360 subscribers in the earlier example, that is 120 subscribers and $25,200 of contribution margin from one quarter’s failures, with no new media.

Saved revenue also moves the ratios the board reads. Observed lifetimes get longer, which improves [LTV:CAC](/glossary/ltv-cac/) from the top rather than by cutting spend, and [MER](/glossary/mer/), total revenue over total marketing spend, rises with no new media at all. The reverse is just as real: in case 17, [Growth spend was outrunning retention](/cases/growth-spend-was-outrunning-retention/), acquisition was healthy on day one but churn erased it by month three ($51k a year identified, findings in 5 days). Ranked by dollars a year, involuntary churn is a [margin leak](/glossary/margin-leak/) like any other, and among the retention leaks it is usually the quickest to fix, because the fix is a setting and a message rather than a new offer.

## Check it every month

Rerun the three numbers every month after billing closes, with the same definitions each time, and add the recovered dollars by method. When a number moves, look at the four decline groups first: the group that grew tells you which lever to check. The published [subscription audits](/industries/subscription/) show where the rest of a subscription budget leaks, and [where SMB marketing budgets leak](/insights/where-smb-marketing-budgets-leak/) sets retention beside the other four leak categories.

## Questions about involuntary churn

### How much of our churn is involuntary?

It depends on the business, so measure it: divide the subscriptions that ended after an unrecovered failed payment by all subscriptions that ended in the same quarter. Count subscriptions left unpaid after retries, not only cancellations. For an outside reference, Stripe wrote in January 2024 that a quarter of lapsed subscriptions are due purely to payment failures, without publishing its method. [Dunning and involuntary churn, defined →](/glossary/dunning/)

### How many times should we retry a failed payment?

A handful of times, spread over days, not daily. Stripe recommends at most eight retries for a charge that allows them, because card networks limit reattempts and issuers may read extra attempts as possible fraud; its default is 8 tries within 2 weeks. Never retry a hard decline such as a stolen card: it needs a new payment method from the subscriber.

### Is it cheaper to keep a subscriber than to win a new one?

Usually, because the kept subscriber’s acquisition cost is already paid. Compare the contribution margin they produce over their expected remaining months, less the cost of retries and emails, with what a new subscriber produces after repaying CAC. For example, at $30 of monthly margin and a $180 CAC, a new subscriber needs six months just to break even. [Payback period, on contribution margin →](/glossary/payback-period/)

[See how the audit reads lifecycle and retention data →](/how-it-works/)

## More field notes

[How to audit your media agency’s programmatic fees, Working with agencies](/insights/audit-media-agency-programmatic-fees/) [The marketing audit process: 8 steps to run one on your own numbers, Audit method](/insights/marketing-audit-process/) [How to cut wasted ad spend: six checks, in the order to run them, Paid media](/insights/how-to-reduce-wasted-ad-spend/) [How to reconcile GA4 and ad platform revenue with Shopify or Stripe, Measurement](/insights/reconcile-ga4-ad-platform-revenue/) [How to measure incrementality with a simple holdout test, Measurement](/insights/how-to-measure-incrementality/) [Pricing audit: where discounts eat your margin, Pricing](/insights/pricing-audit/)

[Request an audit →](/book/)
 [See the benchmark](/benchmark/)