Involuntary Churn Calculator
This involuntary churn calculator sizes the revenue you lose to failed card payments, and only to failed card payments. You give it your subscriber count, your average revenue per subscriber, your monthly decline rate, what you currently recover through retries and dunning, and what you think you could recover. It returns monthly revenue at risk, revenue recovered, revenue lost, the upside at your target recovery rate, the annualised figure, and your effective involuntary churn rate. Every benchmark on this page is sourced to a document we fetched and dated, and where the industry has no reliable public number we say so rather than filling the gap.
Your book
Per month.
Your payment failures
The share of renewal charges that fail on the first attempt.
Of failures you win back through retries and dunning.
Lost to failed payments
$27,636
$2,303.00 a month, on $122,500 of MRR.
Involuntary churn rate
Getting recovery from 53% to 71% is worth $10,584 a year, or $882.00 a month. That is revenue you have already earned and already lost once.
Why renewals fail
| Reason | Share |
|---|---|
| Insufficient funds or credit | 44.4% |
| Card data errors: wrong expiry date or wrong CVV2 | 20.6% |
| Other declines | 15.2% |
| Lost or stolen card | 10.4% |
| Suspected fraud | 9.4% |
Assumes a steady book and a constant decline rate. Annual churn is compounded month on month, which is the only correct way to annualise a rate, and is why this figure is lower than tools that multiply the monthly number by twelve.
Worked example
A US subscription business with 2,500 active subscribers at $49 per subscriber per month. Roughly 4% of renewal attempts decline each month. Its current retry and dunning sequence recovers about 53% of those failed invoices. It wants to know what moving to 71% recovery is worth, and what its involuntary churn rate actually is once recovery is accounted for.
Monthly recurring revenue is 2,500 x $49 = $122,500. At a 4% decline rate, $4,900 of that is at risk every month. At 53% recovery, $2,597 comes back and $2,303 is lost, which is $27,636 a year. Lifting recovery to 71% would bring back $3,479 instead, an extra $882 a month, or $10,584 a year. The effective involuntary churn rate is the decline rate multiplied by the share that never recovers: 4% x 47% = 1.88% of subscribers per month, or about 47 subscribers. Compounded over twelve months that is 20.37% of the base, not the 22.56% you get by multiplying by 12. At the 71% target, involuntary churn falls to 1.16% monthly and 13.07% annualised, about 29 subscribers a month.
Involuntary churn is a plumbing problem wearing a pricing problem's clothes
Most churn dashboards give you one number. Recurly's network data, updated with July 2026 figures, puts overall subscription churn at 3.60%, voluntary at 2.34% and involuntary at 1.25%. Recurly labels those median annual rates, and the label matters more than the values do: roughly 35% of all churn happens because a card failed, not because a customer decided to leave. Most calculators for this search never separate those halves, which is strange, because they respond to completely different work.
Voluntary churn is a product, pricing and onboarding conversation. Involuntary churn is a payments conversation: retry timing, decline code handling, account updater coverage, stored credential indicators, gateway routing. A team that reads a blended 4% and spends a quarter rebuilding onboarding has aimed at the wrong half of its number. If your involuntary slice is small, tuning dunning is busywork.
This calculator refuses to infer the split. It takes your decline rate as an input, because your processor can tell you that number and no benchmark can. Effective involuntary churn is decline rate multiplied by the share of declines that never recover, compounded annually rather than multiplied by twelve: 1.88% monthly is 20.37% over a year, not 22.56%. Recurly also shows involuntary churn falling steeply as average revenue per customer rises, from 1.30% in the $10 to $25 tier to 0.18% above $250. If you sell a $12 app subscription, failed payments are probably your largest leak. If you sell a $400 seat, this page is a rounding error.
What is actually failing, and how much of it comes back
Ethoca ran a study with one card issuing bank across five merchant customers to find out what the generic 'do not honour' response code really represents. The average breakdown: insufficient funds 44.4%, card data errors covering wrong expiry date and wrong CVV2 20.6%, other declines 15.2%, lost or stolen 10.4%, and suspected fraud 9.4%. That research is from March 2017, the most granular public breakdown we could verify, so treat it as shape rather than this year's numbers.
The shape is the useful part. Almost half of declines are a cash timing problem, a fifth are a card data problem, and only about a tenth are fraud. Merchants guess the opposite, and Ethoca found that 52% of orders merchants rejected as fraudulent were good orders they could have fulfilled. You will not see this in your own reporting: as Stripe puts it, card issuers categorise most declines as generic, which is why your decline report has one enormous unhelpful bucket at the top.
Recovery follows the same split. Recurly found that insufficient funds has the highest recovery rate of the top five reasons, that the three most common decline messages recover at over 45%, and that invalid card number, the only hard decline in the top five, still recovers at over 20%, though that is a new card rather than a retry working. Timing is tight: 90% of recovered transactions land within the first 10 days, most in a two to twelve day window. A dunning sequence still running in week four is collecting sunk cost.
Set your target against your own decline mix, not a vendor's case study. Stripe will not retry nine decline codes at all: incorrect_number, lost_card, pickup_card, stolen_card, revocation_of_authorization, revocation_of_all_authorizations, authentication_required, highest_risk_level and transaction_not_allowed. If many of your declines sit in that list, no retry schedule moves your number.
The retry rules are contractual, not advisory
Visa's Business News bulletin AI10325, published 3 September 2020 and effective 17 April 2021, groups authorization response codes into four categories and tells acquirers and merchants to manage reattempts by category. Category 1 means the issuer will never approve and the merchant is not permitted to reattempt at all. Category 2 means the issuer cannot approve at this time, and merchants may reattempt up to 15 times in 30 days. It moved codes 03, 62, 78 and 93 from Category 1 into Category 2 so merchants could retry them, and is explicit that code 14, invalid account number, must not be reattempted using the same account number.
Mastercard signals per transaction instead of by category, using Merchant Advice Codes. Stripe normalises both networks into a single advice_code field with three values: do_not_try_again, try_again_later and confirm_card_data. That field, read per decline, should decide whether attempt number four happens. Stripe recommends a maximum of eight retries and defaults Smart Retries to eight tries within two weeks, while a custom schedule caps at three. Its reason is worth repeating to anyone selling an aggressive retry engine: issuers can read heavy retrying as fraud, raising declines on legitimate charges. For merchants on interchange plus pricing, Adaptive Acceptance blocks some attempts before they reach the network, avoiding excessive retry penalties and cases where authorization is unlikely.
A second rule set affects your baseline before any retry happens. Visa Acceptance states that any transaction initiated by a merchant as a follow-on to an initial cardholder-initiated transaction must follow the merchant-initiated transaction framework, that Visa's mandate dates from 2017, and that Mastercard revised its own framework in late 2021 to cover eight use cases. That means flagging the initiator and transaction type correctly on every renewal, and compliance is associated with higher authorization rates. If your integration does not, your decline rate is inflated before you tune a single retry.
Account updater is the cheapest recovery you will ever buy
Roughly a fifth of declines in the Ethoca breakdown were card data errors, and that bucket is two problems wearing one label. It combines wrong expiry dates with wrong CVV2 values, and only the expiry half is the network's to fix for you. That is what Visa Account Updater and Mastercard Automatic Billing Updater exist to do. The CVV2 half belongs to checkout, since stored credential renewals generally do not submit a CVV2 at all, so do not expect an updater to touch it.
Visa's merchant fact sheet spells out the mechanics. Issuers submit changes within two business days of a permanent change becoming active in their authorization system, and Visa strongly encourages daily files. Merchants enrol through their acquirer, submit account numbers a few days before billing, and must update billing records within five days of a response. Responses cover account number updates, expiration date updates, closed account advices and contact cardholder advices. Triggers include expiration, lost or stolen replacement, account closure and portfolio conversions from Mastercard, American Express or Discover to Visa.
Coverage differs by processor. Stripe performs automatic card updates as part of the platform, with support wide in the United States across most American Express, Visa, Mastercard and Discover cards issued here, though it cannot say in advance which cards support it. Braintree covers Visa, Mastercard and Discover for Braintree Direct merchants based in the US, but it is off by default, pricing depends on your pricing model, and prepaid cards plus Apple Pay and Google Pay tokens are excluded. Adyen's Real Time Account Updater fetches updated details during a decline and immediately retries. Ask your processor whether it is on today, which brands it covers, and the per inquiry charge.
Reading your own number honestly
Watch the period on any benchmark you compare yourself against. This calculator reports involuntary churn monthly and annualised, while Recurly's 1.25% is a median annual rate. The defaults show why that matters: 4% monthly declines with 53% recovery annualises to 20.37%, roughly sixteen times Recurly's median. That gap is not a finding about your business. It is a warning that the 4% placeholder comes from 2017 card-not-present data rather than subscription renewals, and that your own renewal decline report is the only input here worth trusting.
Be careful with the target recovery rate. Recurly published a case where optimised retry strategies moved recovery from about 53% to about 71%, at an enterprise big-box retailer with membership subscriptions and real engineering behind the change. It suggests businesses in the low to mid 50s aim for the high 60s to low 70s. That is not a default, and typing 90% into the target field produces a number nobody should take to a budget meeting.
If your annualised upside is smaller than the annual cost of the vendor you are evaluating, the answer is no. If it is several multiples larger, work in cost order rather than vendor order: fix your stored credential indicators, confirm account updater is on, tune the retry schedule inside Visa's fifteen-in-thirty limit and Stripe's eight-retry guidance, and only then pay for a prediction layer. Each step changes your decline rate or your recovery rate, so re-run this after each one rather than stacking gains measured from the same starting point.
Card account updater coverage at three processors, taken from their own current documentation, checked 4 September 2026. None of the three publishes a per-inquiry price, so the cost column reports what the documentation actually says rather than a figure we cannot verify.
| Fee | Card brands covered | How you get it | Stated cost | Known exclusions |
|---|---|---|---|---|
| StripeStripe describes support as widely available in the United States, with international coverage varying by country. | Most American Express, Visa, Mastercard and Discover cards issued in the US | Runs as part of the platform, surfaced via the payment_method.automatically_updated event | Not listed as a separate charge in the cards documentation | Coverage depends on issuer participation, and Stripe states you cannot identify in advance which cards support automatic updates |
| BraintreeBraintree sends all vaulted Visa, Mastercard and Discover cards to the issuers in batches of 600,000, then moves to rolling updates tied to expiry, upcoming recurring payments and other activity. Accounts holding over 2 million vaulted payment methods only include cards expired within 13 months in the initial request. | Visa, Mastercard, Discover | Off by default, must be requested from Braintree, for Braintree Direct merchants based in the US or transacting primarily with US customers | Documentation states pricing varies by pricing model and directs merchants to contact Braintree | Prepaid cards, and cards processed through Apple Pay or Google Pay |
| AdyenReal Time Account Updater retrieves updated details during the decline and retries immediately, which is the only one of the three that closes the loop inside a single authorization attempt. | Visa, Mastercard, Cartes Bancaires | Real Time Account Updater is synchronous and needs no integration work; Batch Account Updater uses request and result files | Not published in the account updater documentation | No American Express or Discover coverage is listed, and Cartes Bancaires is a French domestic network, so US merchants get Visa and Mastercard |
Assumptions and limits
- The subscriber count and ARPU defaults, 2,500 and $49, are arbitrary placeholders chosen to make the arithmetic legible. They carry no source. The outputs also assume a flat subscriber base, a stable decline rate, and that every recovered invoice belongs to a subscriber who would otherwise have left for good. Growing businesses will understate the opportunity and shrinking ones will overstate it.
- The default 4% decline rate comes from Ethoca's 2017 finding that physical goods merchants often run 3% to 4% and that one digital goods merchant measured approximately 4% when counted by unique cardholder. Per-unique-cardholder is the right unit for a monthly renewal, but the data is nine years old and covers card-not-present broadly, not renewals. Annualised, it implies far more involuntary churn than Recurly's median, so treat it as a high placeholder and replace it with your processor's renewal decline report.
- The 53% and 71% recovery defaults come from a single published Recurly case, an enterprise big-box retailer with membership subscriptions, dated 30 April 2026. One case is not a distribution. They are anchors for the slider, not expectations for your business.
- Recurly labels its churn figures median annual rates. This calculator reports involuntary churn both monthly and annualised, and only the annualised output is comparable to Recurly's 1.25%. Comparing a monthly figure to an annual benchmark is the same period error this page exists to correct, so we label the period on every number.
- Everything here goes stale on a schedule. Network reattempt rules change by bulletin, processor account updater terms and pricing change without notice, and vendor recovery benchmarks are marketing artefacts that get refreshed annually. Re-check the Visa bulletin and your processor's documentation before making a purchasing decision on these numbers.
Frequently asked questions
- What is a good involuntary churn rate?
- There is no single good number, and the first thing to check is the period. Recurly's network data from July 2026 puts involuntary churn at 1.25% against total churn of 3.60%, and labels both median annual rates, so involuntary failures account for roughly 35% of all subscription churn across its book. The same data shows the rate depends heavily on price: 1.30% in the $10 to $25 average revenue per customer tier falling to 0.18% above $250. Compare your annualised figure to those, never your monthly one. The ratio of involuntary to total churn is the more honest comparison anyway, since it does not depend on how the period is defined.
- How do I calculate involuntary churn separately from voluntary churn?
- Involuntary churn is your monthly decline rate multiplied by the share of those declines you never recover. If 4% of renewals decline and you recover 53% of them, your effective involuntary churn is 4% x 47% = 1.88% per month. Subtract that from your blended churn rate and what remains is voluntary. The one thing to avoid is annualising by multiplying by twelve. Churn compounds, so 1.88% monthly is 20.37% of the base over a year, not 22.56%. This calculator does the compounding for you, which is the main reason it exists.
- How many times can you retry a declined subscription payment?
- It depends on the decline. Visa's Business News bulletin AI10325, effective 17 April 2021, groups response codes into four categories. Category 1 means the issuer will never approve and you are not permitted to reattempt at all. Category 2 means the issuer cannot approve at this time and you may reattempt up to 15 times in 30 days. Visa also states that response code 14, invalid account number, must not be reattempted with the same account number. Mastercard signals per transaction through Merchant Advice Codes rather than by category. Separately from the network ceilings, Stripe recommends a maximum of eight retries and defaults its Smart Retries to eight tries within two weeks, warning that heavy retrying can look like fraud to issuers and raise declines on your legitimate charges.
- What percentage of failed subscription payments can you recover?
- Recurly published a case where optimised retry strategies moved recovery from approximately 53% to approximately 71% for an enterprise retailer with membership subscriptions, and suggests businesses currently in the low to mid 50s target the high 60s to low 70s. By decline reason, Recurly found the three most common decline messages recover at over 45%, insufficient funds recovers best of the top five, and invalid card number still recovers over 20%, though that is usually a new card rather than a successful retry. Timing constrains the ceiling: 90% of recovered transactions land within the first 10 days of the failure, so sequences running past three weeks add very little.
- Does a card account updater actually reduce failed payments?
- It addresses a specific slice, and a smaller one than vendors imply. Card data errors accounted for 20.6% of declines in Ethoca's study, but that bucket mixes wrong expiry dates with wrong CVV2 values, and an account updater only supplies the new expiry date or account number. It does not supply a CVV2, and stored credential renewals generally do not submit one anyway, so the expiry half is the part an updater actually recovers. Within that half it works without any customer contact: Visa Account Updater has issuers submit changes within two business days of a permanent change going live, and requires enrolled merchants to update their billing records within five days of a response. Stripe runs automatic card updates as part of the platform with wide US coverage across Amex, Visa, Mastercard and Discover. Braintree offers it on Visa, Mastercard and Discover but off by default, excluding prepaid cards plus Apple Pay and Google Pay tokens. Adyen offers a real time version that retries inside the same decline. Confirm yours is active and re-measure for a full billing cycle before you evaluate a dunning vendor.
Go deeper
Other calculators
Know your number before you buy a solution for it
Once you have your annualised involuntary churn figure, the next question is whether your current processor is the reason it is that high. Account updater coverage, stored credential handling and retry controls vary more between processors than pricing does, and they are rarely on the comparison page. See how the major US processors handle recurring billing before you pay a third party to patch around one.
Looking for something else? See free payment tools and calculators.