SaaS Churn Rate Benchmarks (2026)

Search for a SaaS churn benchmark and you get a number without being told who was counted. That's the whole problem. The two kinds of panel that publish these figures, surveys of operators and billing data measured across a platform, disagree at the small end by nearly forty points, and each is right about the companies it holds.

High Alpha surveys more than 800 SaaS founders and executives, half of them the chief executive, and reports median gross revenue retention of 92% for companies under $1M ARR (High Alpha SaaS Benchmarks, 2025). ChartMogul computes from billing data across more than 2,100 SaaS businesses and puts the median at 53% for companies under $300k ARR, over the twelve months to March 2023 (ChartMogul SaaS Benchmarks, 2023). That edition is the most recent one to publish the size and price cuts this page uses, so two of its tables are three years old and say so in their notes. Converted to a monthly rate, that is the difference between losing 0.7% of your recurring revenue each month and losing 5.2% of it.

Neither number is wrong, and the gap between them is the part worth understanding. A founder who reads the survey figure and concludes they are three times worse than the market may only be reading a benchmark drawn from a different population. What is a good SaaS churn rate has no answer until you know which of those two groups you are being measured against.

Every figure below names the panel behind it.

Four different quantities travel under the word churn, and a published figure is unusable until you know which one it is. Logo churn counts customers. Revenue churn counts money. Gross revenue retention is the mirror of revenue churn and cannot exceed 100%. Net revenue retention adds expansion from customers you kept and so has no ceiling, which is how a company can report 110% net retention while still losing a fifth of its customers every year.

The distinction matters most at the two ends of the market. A business selling at $20 a month loses customers steadily and has almost no expansion to offset it, so its logo churn and its revenue churn sit close together. An enterprise vendor with seat-based growth can hold net retention above 100% while gross retention sits in the eighties. You can compute either against your own numbers with the churn rate calculator, or the revenue-weighted version with the MRR churn calculator.

Period is the trap

The second thing a figure needs is a period, and this is where most published comparisons quietly fail. Stripe's 38% is annual churn measured on monthly-billed subscriptions only (Stripe Churn Benchmarks, 2025). Compounded down, 38% a year is about 3.9% a month, an ordinary monthly figure that looks alarming when printed beside one. Recurly publishes 2.5% voluntary and 0.9% involuntary churn without stating anywhere whether the rate is monthly or annual (Recurly State of Subscriptions, 2025), which leaves it impossible to place on either scale.

The tables below set out the four measures and then four published SaaS figures beside the basis each publisher states. The figures span more than thirty-fold, and no two of them are directly comparable, so convert yours to one period before you set it beside any of them.

Almost every SaaS retention benchmark in circulation comes from one of two places, and the difference between them is not a detail.

Operator surveys

High Alpha and Benchmarkit both survey operators. High Alpha's 2025 edition has more than 800 respondents, half of them founders or chief executives, weighted 69% to the United States (High Alpha SaaS Benchmarks, 2025). Benchmarkit's is syndicated by partners to their portfolio companies and vendor audiences, and its retention questions are answered by 225 of its 583 participants (Benchmarkit B2B SaaS Metrics, 2025). Both are self-reported. Operators generally know their own metrics, but the number is typed into a form and never reconciled against a billing system, which is a different kind of evidence from a computed one.

Measured billing panels

ChartMogul computes from billing data across more than 2,100 SaaS businesses, with no self-report in the loop (ChartMogul SaaS Benchmarks, 2023). It carries a selection effect of its own in the opposite direction: companies not active for the full twelve-month window are excluded, so the panel describes survivors, and its members are businesses that chose a self-serve subscription analytics product.

What the two disagree about

At the small end they are almost forty points apart. As company size rises they converge, and they converge in a specific way. ChartMogul's top quartile runs 70% at the smallest band and 82% at $15-30M, which is roughly where High Alpha's medians sit across its whole range (ChartMogul SaaS Benchmarks, 2023). The survey panel reports something close to the measured panel's top quartile.

That is what you would expect if the operators who answer a benchmarking survey are the ones confident enough in their metrics to have them at hand. It isn't evidence that either publisher is careless, and it doesn't make the survey figure useless. It makes it a top-quartile target, not a median.

This is the cut where the two panels contradict each other outright, so it is worth being precise about what each one says.

In High Alpha's survey, gross retention runs 92, 92, 88, 90 and 88 across five ARR bands from under $1M to over $50M (High Alpha SaaS Benchmarks, 2025). It is flat, drifting down four points. The previous edition showed the same shape more steeply, 92% under $1M against 83% above $50M (High Alpha SaaS Benchmarks, 2024). OpenView's survey dips through the middle and recovers at the top, running 97, 95, 90, 90, 90 and 95 across six bands (OpenView SaaS Benchmarks, 2022).

The measured panel goes the other way. ChartMogul's median gross retention rises from 53% under $300k ARR to 69% at $15-30M, and the rise holds at every point in the distribution except the top decile, which is flat between 85% and 89% throughout (ChartMogul SaaS Benchmarks, 2023).

Reading the table

The spread down each column is larger than the spread across the bands. At $1-3M ARR the gap between the bottom quartile and the top decile is 38 points; the gap between the smallest and largest band at the median is 16. Company size is a weak predictor of gross retention next to where you sit among companies of your own size.

The two panels are reconcilable. Survey respondents at $1M ARR resemble the measured panel's top quartile, and that top decile really is flat across size. What isn't supportable is the claim, common in benchmark write-ups, that gross retention reliably improves as a company scales. In the one panel here that measures rather than asks, it improves; in the surveys it doesn't. Check which kind of panel your benchmark came from before you read a trend into it.

Customer size moves retention further than company size does, and it does so in both panels, which makes it the first cut to check.

Across ChartMogul's ARPA bands the median runs from 50% under $25 a month to 82% above $1,000, a 32-point spread against 16 points across its ARR bands on the same measure (ChartMogul SaaS Benchmarks, 2023). The effect appears at every point in the distribution and is largest at the bottom, where the weakest quartile improves 33 points from end to end.

Bessemer's guidance from an earlier era says the same thing in target form: 70-80% gross retention for SMB contracts under $12,000, 80-90% for mid-market, above 90% for enterprise contracts over $50,000 (Bessemer State of the Cloud, 2019). Those are offered as targets to aim for rather than measured outcomes, and that edition states no population or method anywhere, so they belong here as corroboration of direction and not as benchmarks.

The practical reading is that a business selling at $20 a month will not reach enterprise retention by improving onboarding. The ceiling is set by what the product costs and who buys it. Moving up the price ladder changes the ceiling; retention work moves you within it. Both matter, and they are not substitutes for one another.

One caveat on the table below: ChartMogul excludes companies under $300k ARR from every ARPA aggregate (ChartMogul SaaS Benchmarks, 2023), so these bands describe established businesses. The smallest companies appear in the ARR table above but not in this one.

Pricing model moves retention by roughly the same magnitude as a large move in company size, which makes it worth checking before concluding that a churn problem is a product problem.

Benchmarkit's 2025 edition puts gross retention at 92% for usage-based pricing against 88% for both subscription and hybrid models, the first year it calculated that split (Benchmarkit B2B SaaS Metrics, 2025). On the net side, its 2026 edition reports 108% for usage-based against 98% for seat-based (Benchmarkit SaaS and AI-Native Metrics, 2026).

Treat the gross figure with some caution. The same publisher's 2024 edition reported usage-based gross retention three points lower than other models (Benchmarkit B2B SaaS Metrics, 2024), the opposite sign one edition earlier. That reversal, and what it implies about reading any single edition of a benchmark survey, is worked through on the NRR and GRR benchmarks page, which covers the expansion side of pricing model in more depth.

Three questions settle which benchmark on this page applies to you.

What do you charge?

This usually dominates everything else. Find your average revenue per account in the customer-size table and read the band you are in. A business under $25 ARPA whose gross retention is 63% sits in the top quartile of its band; the same figure at $500-1,000 ARPA is below the bottom quartile. Company size rarely changes that reading.

Which panel resembles you?

If you are venture-backed B2B SaaS selling to mid-market or enterprise buyers, the operator surveys describe companies like yours, and 88-92% gross retention is the range to hold yourself against (High Alpha SaaS Benchmarks, 2025). If you are self-serve, prosumer or small-business SaaS billing monthly, the measured billing panel is the closer comparison, and the median for your size is a great deal lower than the survey figure.

Are the units the same?

Convert everything to one period before comparing. Annual and monthly figures for the same business differ by roughly an order of magnitude, and a substantial share of published SaaS churn figures don't state which they are. The retention rate calculator will put your own number on the same annual basis as the tables above, and the churn rate benchmarks page carries a full conversion table.

What good looks like

For established B2B SaaS, gross revenue retention in the high eighties with net retention above 100% is a defensible target, and it is where both surveys sit. For self-serve SaaS under $1M ARR, the measured median is between 53% and 62% and the top quartile between 70% and 76% (ChartMogul SaaS Benchmarks, 2023). Aim at your own band's top quartile, not at a number lifted from an enterprise survey. Whichever applies, the trend across four quarters tells you more than the level does, because the level is largely set by what you charge and who buys it.