Marketing

Subscription Churn Calculator

Split churn into voluntary and involuntary — because a third of it is usually failed payments, which is a billing problem you can fix without touching the product.

Subscription Churn Calculator

Results recalculate instantly on every keystroke. Nothing you type is transmitted.

Churn
Cause
Value
$
Cohort
Monthly Churn Rate
churned ÷ starting customers. Not all of it is a choice.
Involuntary Churn
Voluntary Churn
Recoverable by Dunning
Churn After Dunning
Annual Churn Compounded
The 12× Error
Constant-Churn Lifetime vs a Real Cohort

What this result does not account for

  • Period rates exclude within-period joiners; cohort rates do not.
  • The cohort curve is simplified to a first month and a mature rate.
  • Customer churn only — revenue churn is the MRR page.
Zero-Server Execution Updated 11 Aug 2026 Reviewed by Imran S. Qureshi, CFA IEEE-754 Double Precision

In short: A third of churn is a billing problem. Of 28 customers lost, 9.52 were involuntary; dunning recovers 4.00 of them, cutting monthly churn from 2.3333% to 2.0000% with no product change at all.

Formula

churn = customers lost ÷ customers at start

annual = 1 − (1 − monthly)12, NOT monthly × 12

[('involuntary', 'failed payments — a billing problem'), ('voluntary', 'a decision to leave — a product signal'), ('dunning', 'recovers a large share of involuntary churn'), ('constant churn', 'overstates lifetime against a real cohort')]

Worked Example

  1. Divide customers lost by customers at period start.
  2. Split into involuntary and voluntary.
  3. Apply a dunning recovery rate to the involuntary share.
  4. Compound monthly churn for the annual figure — never multiply.
  5. Compare the constant-churn lifetime against a real cohort curve.

28 of 1,200 customers lost is 2.3333% monthly churn. At a 34% involuntary share, 9.52 of those were failed payments rather than decisions to leave; a 42% dunning recovery saves 4.00 customers, cutting churn to 2.0000% and preserving 595.84 of monthly recurring revenue. Compounded, 2.3333% monthly is 24.6720% annually — not the 28.0000% that multiplying by twelve suggests. And a real cohort retaining 70% in month one and 98.5% thereafter survives 27.82 months against the 42.86 the constant-churn model claims: 1.5404 times too long.

Strengths & Limits Of This Model

Where this engine is strong

  • Separates the billing problem from the product problem
  • Compounds annual churn and prices the 12x error
  • Tests constant churn against a real cohort curve

Where it stops

  • Simplified cohort curve
  • Excludes downgrades

Risk & accuracy notice. A single blended churn rate applied forever overstates average customer lifetime substantially, because real cohorts churn hardest in month one. Every lifetime value built on that assumption inherits the error.

Practical Use Cases

Finding cheap churn reduction

Isolating the involuntary share.

Building a dunning business case

Pricing recovered customers in MRR.

Reporting annual churn

Compounding rather than multiplying.

Sanity-checking LTV

Testing the constant-churn lifetime against a cohort.

Separating signals

Reading voluntary churn as the product measure.

Methodology & Editorial Standards

Churn is customers lost over customers at period start, with rates above one hundred per cent refused because they indicate within-period joiners being counted against a base that never held them. The involuntary share is separated and dunning recovery applied to it alone, since it is a billing rather than a product problem. Annual churn is COMPOUNDED and the naive twelve-times figure shown beside it. A cohort survival curve is summed to give an empirical lifetime and compared against the constant-churn reciprocal, exposing how far that model overstates.

Computation runs in IEEE-754 double precision at full internal precision; rounding to two decimal places occurs strictly at the display layer, so no cumulative drift enters the result. All monetary outputs use accounting presentation — grouped thousands, two decimals, negatives in parentheses — so figures can be transcribed directly into a model or working paper. Division-by-zero and out-of-domain inputs return an em-dash rather than a misleading number.

This engine was reconciled against an independent reference implementation and hand-verified for the worked example above before release. Our full five-stage review process is published on the About Us page.

Imran S. Qureshi, CFA Head of Quantitative Modelling · ApexConverter

Performance-marketing unit economics and contribution-margin analysis. Last reviewed: 11 August 2026.

Disclaimer. This calculator is provided for informational and modelling purposes only and does not constitute financial, tax, legal, medical, or engineering advice. Verify all figures with a qualified professional before acting on them.


Subscription Churn Calculator — 10 Expert FAQs

10 analyst-written answers to the questions practitioners actually ask — optimised for voice and answer-engine retrieval.

What is the difference between voluntary and involuntary churn?

Voluntary churn is a customer deciding to leave; involuntary churn is a payment failing — an expired card, insufficient funds, a fraud block. The second is a billing problem and is far cheaper to fix, yet an aggregate churn number hides it completely.

How much churn is involuntary?

Commonly around a third, and in consumer subscriptions sometimes more. Because it requires no product change to address, it is almost always the highest-return retention work available and almost always the least glamorous.

What is dunning?

The process of recovering failed payments — retrying cards on an intelligent schedule, notifying customers before expiry, and using account-updater services that refresh card details automatically. It routinely recovers a large minority of involuntary churn.

Why can't I multiply monthly churn by twelve?

Because each month churns a share of the customers who survived the previous month, not of the original base. Compounding gives a lower figure than multiplication, and at high monthly rates the naive method produces annual churn above one hundred per cent — which is impossible.

What is the constant-churn fallacy?

Assuming a single churn rate applies for the whole customer lifetime. Real cohorts churn hardest in the first month and stabilise afterwards, so applying a blended rate forever substantially overstates average lifetime — and every LTV built on it inherits the error.

Should I measure customer churn or revenue churn?

Both, because they answer different questions. Customer churn counts logos and reflects product fit; revenue churn weights by value and reflects financial impact. If they diverge sharply, your churn is concentrated in one customer size band, which is itself the finding.

What is a good SaaS churn rate?

For business software, monthly churn around one to two per cent is considered healthy, with early-stage products often running considerably higher while finding fit. Consumer subscriptions tolerate much higher rates because acquisition is cheaper and faster.

Does churn include downgrades?

Customer churn does not — a downgraded customer is still a customer. Revenue churn does, because contraction reduces recurring revenue exactly as a cancellation does. This is one of the main reasons the two figures diverge.

How does churn timing affect the calculation?

Customers who both join and leave within the period are not in the starting base, so a period rate excludes them while a cohort rate does not. In a fast-growing business the two diverge substantially and the period rate flatters.

What is the fastest way to reduce churn?

Fix involuntary churn first, because it needs only billing work. Then address the first-month experience, since that is where real cohorts lose the most customers. Both are cheaper and faster than the product changes teams usually reach for.

Related Marketing Engines