Subscription churn rate: how to calculate it and what is good

Subscription churn rate is the share of your customers, or of your recurring revenue, that you lose in a given period. It sounds simple, and it is the most misquoted number in subscription businesses. Half the confusion comes from mixing time periods: 5% monthly churn is not “5% churn”, it is about 46% of your customers gone in a year.

This guide gives you the formulas, the edge cases that change the answer, checked worked examples, and a table of benchmarks where every figure names its publisher, year, sample and caveat. It also lists the popular churn numbers we could not trace to a source, so you can stop quoting them.

The short answer.

• Formula: customers lost in the period, from the customers you had at the start of it, divided by the customers you had at the start. Revenue churn does the same with recurring revenue.

• What is “good” depends on: whether the figure is monthly or annual, customer or revenue based, and how much each customer pays. Higher-priced plans churn less.

• Convert carefully: annual churn is 1 minus (1 minus monthly churn) to the power of 12. Multiplying monthly churn by 12 overstates it.

• Check the source: most published benchmarks are top quartile, vendor-reported or undated. Use them as context, not as a target.

What churn rate means: the four versions people confuse

“Churn rate” is used for at least four different measurements. They give different numbers for the same business, so say which one you mean every time.

Customer (logo) churn

Customer churn, also called logo churn, counts accounts. It is the number of customers you lost in the period divided by the number you had at the start. A customer who pays $20 a month and one who pays $20,000 a month count the same.

Gross and net revenue churn

Revenue churn, also called MRR churn, counts money. Gross revenue churn is the recurring revenue lost to cancellations and downgrades, divided by recurring revenue at the start of the period. Net revenue churn also subtracts expansion revenue from the customers who stayed, so it can go negative. Negative churn means the extra revenue from existing customers is larger than the revenue you lost.

Gross revenue retention (GRR) is 100% minus gross revenue churn, and it cannot exceed 100%. Net revenue retention (NRR) includes upsells, cross-sells and price increases, and it can. SaaS Capital measures both by comparing each customer’s revenue across the two points in time before adding anything up. Adding first lets expansion hide losses.

Voluntary and involuntary churn

Voluntary churn is a customer choosing to leave. Involuntary churn is a subscription ending because a payment failed, for example an expired card or a declined charge, without the customer deciding to go. They need different fixes, which is why they should be tracked separately. If you want the detail, read our guide to involuntary and voluntary churn.

Churn rate, retention rate and NRR

Retention is the other side of the same coin. A 3% monthly customer churn rate is a 97% monthly customer retention rate. NRR is a revenue measure that includes expansion, so it is not simply “100% minus churn”, and it is usually measured over twelve months.

How to calculate subscription churn rate

Customer churn, step by step

1. Choose the period (a month is common).

2. Count the customers you had at the start of the period.

3. Count how many of those customers left during the period.

4. Divide the second number by the first.

There is one edge case that trips people up. Customers who join and leave inside the same period were not in your starting count, so they should not be in the numerator. ChartMogul excludes them in its own formula for the same reason.

Example A: one month of logo churn. You start the month with 1,000 customers and 60 new customers sign up. During the month 35 customers cancel, and 5 of them were among the 60 who just joined.

• Customers lost from the starting base: 35 minus 5 = 30.

• Monthly logo churn: 30 ÷ 1,000 = 3.0%. If you wrongly include the 5 new customers it reads 3.5%.

• Customers at the end: 1,000 + 60 − 35 = 1,025.

• If the rate stayed constant, the average customer would last 1 ÷ 0.03 = 33.3 months.

Revenue churn, with downgrades and expansion

Example B: one month of revenue churn. Monthly recurring revenue at the start is $40,000. During the month, cancellations take away $1,200 and downgrades take away $600. Customers who stayed expand by $2,000. One returning customer comes back at $400.

• Gross MRR churn: (1,200 + 600) ÷ 40,000 = 4.5%. Gross retention for the month is 95.5%.

• Net MRR churn: (1,800 − 2,000) ÷ 40,000 = −0.5%. Net retention is 100.5%.

• The $400 reactivation is shown separately as reactivation revenue, not netted against churn.

This business is shrinking in logos and in gross revenue, yet its net revenue churn is negative. Both statements are true, which is why gross and net should always be reported side by side.

Monthly and annual churn are not interchangeable

Churn compounds, so annual churn is not twelve times the monthly figure. The correct conversion is:

Annual churn = 1 − (1 − monthly churn)^12

Check one row: 0.95 to the power of 12 is 0.540, so 5% monthly churn loses 46% of a cohort in a year, not 60%. Going the other way, 20% annual churn is about 1.8% a month, and 10% annual is about 0.9% a month.

Two cautions. The formula assumes a constant rate across the whole base, which real customers do not follow. And SaaS Capital warns that annualising a single month is fragile: one month with no losses annualises to a meaningless 100% retention. Compare the same month a year earlier, or use three-, six- and twelve-month trailing averages.

Cohort retention shows what averages hide

Example C: one cohort over five months. 200 customers join in January. At the end of each of the next five months, 170, 150, 138, 130 and 125 are still active.

Churn falls with age: young customers leave fastest. A single blended monthly rate would overstate churn for your older customers and understate it for your newest.

Logo and revenue retention can also diverge. If 20 of 100 customers leave but they held half your revenue, logo retention is 80% and revenue retention is 50%.

Renewal-based churn for annual contracts

Example D: annual contracts. 120 annual contracts come up for renewal this quarter. 18 do not renew, and 6 renew at a lower price.

• Renewal-based logo churn: 18 ÷ 120 = 15.0%, so logo retention at renewal is 85%.

• The 6 downgrades are revenue contraction, not logo churn.

The denominator here is contracts up for renewal, not your whole customer base. Both are legitimate. Always state which one you are using, because they produce different numbers.

Edge cases: how to count downgrades, reactivations, trials and pauses

Most disagreements about churn are really disagreements about edge cases. Pick one rule for each, write it down, and keep it. This table gives the rule we recommend and the reasoning behind it.

On pauses: Stripe lets you set a subscription to paused after failed payment retries, and it stops generating invoices until the subscription resumes. That is a billing state, not a churn definition, so decide your own.

What is a good churn rate? Benchmarks with their caveats

There is no single good churn rate. Before you compare yourself to anything, read the column that matters most in the table below: the caveat. Every figure here comes from a source we opened. None is a universal target.

How to read the table

• Price drives churn more than anything else. ChartMogul’s medians fall from 6.1% to 2.2% monthly as the price per account rises. Higher-paying customers are also different buyers, so this is a correlation, not a lever you can pull by raising prices.

• Dollar retention and customer churn are different numbers. A GRR of 91% implies about 9% annual gross revenue churn. That is not comparable with Recurly’s 3.22% annual figure, which probably reflects a different customer mix and definition.

• Medians and top quartiles are different things. A top-quartile figure is what the best quarter of companies achieve. Treating it as an average will make you feel worse than you should.

• Check the period. Some sites repeat Recurly’s older figures as if they were monthly. Recurly’s current page labels them annual.

• Ask who is excluded. Surveys of companies that were active for the whole year leave out the ones that failed.

Annual and monthly plans

Customers on annual plans have fewer renewal decisions and fewer card failures, so they tend to churn less during the term, and the churn arrives in a spike at renewal. One published example is Buffer’s, which Baremetrics reports: monthly plans churned about 7% a month and annual plans about 2.4% a month equivalent. That is a single company. Customers who choose annual plans are also more committed to start with, so you cannot say the plan type caused the difference.

Churn numbers you will see that we could not source

Some figures appear on dozens of pages with no traceable source. We checked, and could not find a primary source for any of them, so we have not used them.

• “5% annual churn is good for SaaS”, and similar single numbers. Pages that quote it attribute it to “experts”.

• “Enterprise under 1% monthly, mid-market 1 to 3%, SMB 2 to 5%”. It is repeated widely and has no primary source.

• “Netflix has 2% monthly churn”, and monthly rates for subscription boxes, telecoms, media and ecommerce quoted without a source.

• “Dunning recovers 50 to 70% of failed payments” and “smart retries recover 40 to 55%”. These come from vendors and third-party blogs, and the Stripe and Recurly documentation we read publishes no recovery percentage.

• “Gyms lose 30 to 50% of members a year”. We could not open the underlying membership reports.

If you read a churn benchmark that does not state its period, its sample and whether it is a median or a top quartile, treat it as an anecdote.

Churn analysis: finding out why customers leave

Churn rate tells you how much you lose. Churn analysis tells you where, when and why. It is the same data cut in a few specific ways, and each cut answers a different question.

A simple routine: fix your definitions first, using the edge cases table above. Calculate logo churn and gross revenue churn each month. Split both by cohort and plan. Tag every cancellation with a reason, and add a short exit survey so voluntary churn has an explanation. Track voluntary and involuntary churn separately. Then review the numbers quarterly and act on the largest segment first.

Why subscribers churn

Voluntary reasons

Recurly’s 2026 subscription report says 52% of consumers cancelled at least one subscription in the past year because they were not using it. The report gives no method, so treat it as a signal rather than a measurement. Price and lack of perceived value are the other reasons that come up repeatedly, but we have no verified data to put a share on them.

Involuntary causes

Failed payments end subscriptions that customers did not mean to end. The usual causes are expired cards, insufficient funds and issuer declines. Estimates of how much of total churn this represents vary and are vendor-reported. Stripe says 25% of lapsed subscriptions are purely due to payment failures, and elsewhere says up to 40% of total churn. Recurly’s own figures imply roughly a third for SaaS (1.06 of 3.22 points). A fair summary is “roughly a quarter to two-fifths, depending on who you ask”. Measure your own split, since it is the one part of churn you can address without changing your product or your price.

Churn prevention strategy: levers ranked by evidence

Fix involuntary churn first

Payment failures are the most mechanical part of churn, so they are the easiest to work on.

• Card updates. Stripe’s documentation says automatic card updates refresh saved cards when issuers replace them, and are widely supported in the United States for most American Express, Visa, Mastercard and Discover cards. Support varies by country, and you cannot tell which individual cards will be updated.

• Smart retries. Stripe’s documentation says smart retries typically improve recovery compared with fixed rules. By default Stripe retries up to 8 times within 2 weeks for new accounts, and a custom schedule is limited to 3 retries. It does not publish a recovery percentage. Recurly’s documentation describes a machine learning engine that uses decline codes, with a retry count typically between three and fifteen, and also publishes no recovery rate.

• Dunning communications. Emails that tell customers a payment failed and give them an easy way to update their card. See our guide to dunning management.

• Hard declines. Do not keep retrying a card the issuer has said will never work. Ask the customer for a new payment method instead.

Our guides on how to reduce involuntary churn and the dunning strategy playbook go deeper.

Annual plans and billing frequency

Moving customers to annual plans reduces how often they can leave and how often a card can fail. Treat the mechanism as real and the size of the effect as unproven, because of the self-selection described above.

Onboarding and the first 90 days

Example C shows why: churn is highest early, so anything that gets new customers to their first real result sooner works on the part of the curve where most of the loss happens. We have no verified figure for how much onboarding changes retention, so measure it with your own cohorts.

Cancellation flows, pause and skip

Recurly’s 2026 report says pause usage rose 337% year on year where offered, and that 3 in 4 customers who pause come back. Some vendors report “save rates” from cancellation flows, but they define a save differently, and some count a pause as a save. A save that only delays cancellation by two weeks does not count. Measure the 90-day retention of customers you saved.

Frequently asked questions

What is churn analysis?

Churn analysis is the process of breaking churn down by cohort, plan, reason and type, so you can see where customers leave, when and why. A single churn rate tells you how much you lose. Churn analysis tells you what to fix.

What is a good churn rate for a subscription business?

It depends on the period, on whether you measure customers or revenue, and on how much each customer pays. ChartMogul’s data puts a typical monthly churn rate at 3 to 4%, and 1 to 2% in the top quartile. SaaS Capital’s median gross revenue retention for bootstrapped B2B SaaS is 91% a year. Use these as context, and compare yourself to businesses with a similar price point.

Is 5% churn good?

It depends on the period. 5% monthly churn means you lose about 46% of a cohort in a year, which is high for most subscription businesses. 5% annual churn would be excellent for most.

How do you calculate subscription churn rate?

Divide the customers you lost during the period, out of those you had at the start, by the customers you had at the start. For revenue churn, use recurring revenue lost, including downgrades, divided by recurring revenue at the start.

How do you convert monthly churn to annual churn?

Annual churn is 1 minus (1 minus monthly churn) to the power of 12. For example, 2% monthly is about 21.5% annual, not 24%.

What is the difference between customer churn and revenue churn?

Customer churn counts accounts. Revenue churn counts the money those accounts pay. They diverge when your larger customers behave differently from your smaller ones.

What is net negative churn?

Net negative churn happens when the extra revenue from existing customers, through upgrades and added seats, is greater than the revenue you lose from cancellations and downgrades. Net revenue retention above 100% is the same thing.

What is the difference between voluntary and involuntary churn?

Voluntary churn is a customer choosing to cancel. Involuntary churn is a subscription ending because a payment failed. Involuntary churn is usually the easier of the two to reduce.

What is the churn rate for annual versus monthly subscriptions?

Annual plans usually churn less within the term, with a spike at renewal. We could not verify a trustworthy benchmark that compares them across companies, so measure both on your own data.

How much does reducing churn by 1% increase revenue?

It depends on your customer base and your plan prices, so the honest answer is to model it. The ceiling example shows why it matters: with 1,000 new subscribers a month and a constant 5% monthly churn, your base levels off at about 20,000. At 8% churn it levels off at 12,500.

Key takeaways

• Churn rate comes in four versions: customer, revenue, voluntary and involuntary. Always say which one you are measuring and over what period.

• Annual churn is 1 minus (1 minus monthly churn) to the power of 12, so 5% a month is about 46% a year, not 60%.

• Decide your rules for downgrades, reactivations, trials, pauses and non-renewals before you calculate anything, and keep them fixed.

• Published benchmarks are mostly top quartile, vendor-reported or dated, so compare yourself to businesses with a similar price point and read each source’s caveat.

• Involuntary churn is the most mechanical part to fix: card updates, retries, dunning and sensible handling of hard declines come before any change to product or price.

Churn is one of those numbers that rewards being boring: the same definition, the same period, every month. Pick your rules this week, write them down where your finance and product teams can both see them, and look at the trend rather than any single month. If failed payments turn out to be a big part of your picture, our subscription management platform is built around fixing exactly that.