Average SaaS churn rate: benchmarks by contract size and billing
Median GRR for private B2B SaaS is 84% to 91% a year, roughly 1% lost a month. What that means for a monthly e-commerce app and how to measure yours.
In short: The two largest recent surveys of private B2B SaaS put median gross revenue retention at 91% (SaaS Capital, 1,000+ companies) and 84% (Aleph x Benchmarkit, 226 companies), which works out to roughly 0.8% to 1.4% of revenue lost per month. Month-to-month contracts and contract values under $12,000 a year sit at the low end of every cut, which is where most Shopify and TikTok Shop apps live. Before you compare, check whether the number you are quoting is logo churn, revenue churn or gross revenue retention, because they are three different numbers from the same month.
Logo churn, revenue churn and gross revenue retention
Most churn benchmark arguments come down to two people quoting different metrics. Here are the three, on one example. A Shopify app starts the month with 400 paying merchants and $40,000 in MRR.
| Metric | Formula | In the example |
|---|---|---|
| Logo churn | Merchants lost in the period, divided by merchants at the start | 48 uninstalled, so 48 / 400 = 12% |
| Gross revenue churn | MRR lost to cancellations plus downgrades, divided by starting MRR | $3,000 from the 48 plus $1,000 of downgrades, so $4,000 / $40,000 = 10% |
| Gross revenue retention (GRR) | Starting MRR minus churn and downgrades, divided by starting MRR | $36,000 / $40,000 = 90% |
GRR is the mirror of gross revenue churn, measured over a year for a cohort. It never counts upgrades, which is what separates it from net revenue retention. I cover NRR and its benchmarks in the guide on net revenue retention.
The gap between 12% logo churn and 10% revenue churn is not a rounding error. The merchants who leave an e-commerce app are mostly on the smallest plan, so you lose more logos than dollars, and a logo churn number on its own will always look worse than a revenue number.
What is the average SaaS churn rate?
Published numbers come as retention rather than churn, and the samples differ, so here are the sources and the cuts rather than one headline.
SaaS Capital's 2025 retention benchmarks, its 14th annual survey with more than 1,000 private B2B SaaS companies responding, report a median GRR of 91% and median NRR of 101% across the whole sample. Its cut by annual contract value, which excludes companies under $1M ARR, puts the under-$12k band at 90% GRR, every band from $12k to $250k at 90% or 91%, and only the above-$250k band higher at 95%. By contract length, month-to-month contracts show 89% GRR and 100% NRR, annual contracts 90% and 101%, multi-year 94% and 103%. The report's own conclusion is that GRR of at least 90% is table stakes.
Aleph's summary of the 2026 Aleph x Benchmarkit report, built on full-year 2025 data from 342 SaaS and AI-native companies, 226 of which reported GRR, puts the median at 84%, down from 88% the year before, with the top quartile at 91% and the bottom quartile at 76%. Product-led companies sit at 79% and sales-led at 88%. Companies with contract values under $5,000 sit at 80%.
Lighter Capital's 2025 B2B SaaS startup benchmarks are the only set I could verify that is built from connected business data rather than a survey. Across 155 private B2B SaaS startups in the US, Canada and Australia, the median annualised customer churn is 16.25% and the median annualised revenue churn is 12.5%. Lighter Capital annualises monthly churn with the formula 1 minus (1 minus monthly churn) to the power of 12. The page does not say whether revenue churn is gross or net of expansion, so treat it as roughly comparable to gross revenue churn.
| Source | Sample | Median gross revenue retention | Low-end segment |
|---|---|---|---|
| SaaS Capital 2025 | 1,000+ private B2B SaaS, survey | 91% | Month-to-month 89%, under $12k ACV 90% |
| Aleph x Benchmarkit 2026 | 226 companies reporting, full-year 2025 | 84% | Product-led 79%, under $5k ACV 80% |
| Lighter Capital 2025 | 155 startups, connected data, 2020 to 2024 | 87.5% (from 12.5% revenue churn) | Not broken out |
Two samples seven points apart on the same metric shows how much the answer depends on who was asked. SaaS Capital's ACV cut excludes companies under $1M ARR, and Aleph reports product-led companies nine points below sales-led, so the mix of each sample moves its median. For a self-serve app sold to small merchants on monthly billing, the relevant medians sit between 79% and 90% GRR, and 90% is the ceiling of normal for the segment rather than the floor.
What does an annual retention figure mean per month?
The surveys report annual GRR. Most e-commerce apps bill monthly and watch a monthly number. The conversion is arithmetic, not a benchmark: monthly revenue loss is 1 minus the twelfth root of annual GRR. If retention were spread evenly through the year, which it is not, the medians come out like this.
| Annual GRR | Implied monthly gross revenue loss |
|---|---|
| 95% | 0.43% |
| 91% | 0.78% |
| 90% | 0.87% |
| 89% | 0.97% |
| 84% | 1.44% |
| 80% | 1.84% |
| 79% | 1.95% |
The same arithmetic on Lighter Capital's 16.25% median annual customer churn gives about 1.5% of customers lost per month. The caveat is that churn is front-loaded. A merchant who installed this month is far more likely to leave than one who has paid for a year, so a growing app sees a monthly number above the implied figure and a mature app sees one below it.
Why do month-to-month e-commerce apps sit at the low end?
Three reasons, and only one of them is about your product.
The contract has no walls. SaaS Capital's data shows retention rising with both contract value and contract length. A $49 a month app has no implementation, no account manager and an uninstall button in the Shopify admin. The decision to leave costs the merchant nothing, so more of them make it.
Merchants churn for reasons that are theirs. Since April 2026, Shopify's updated uninstall reasons require a merchant to pick one of eight reasons before removing an app, and two of them, "Store is closing or pausing" and "Testing multiple apps", have nothing to do with whether your app worked. Stores close, stores go seasonal, and a merchant who installed four review apps on Tuesday will uninstall three by Friday.
Most of the loss is in the first month. Userpilot's 2025 product metrics benchmark puts average month 1 retention at 46.9% across 83 B2B companies, from first-party data in Userpilot's own analytics dashboards. The sample is Userpilot's customers, not e-commerce apps, and I would expect the shape to hold for e-commerce apps too. On that data, fewer than half of accounts are still active after the first month. If your churn number includes trials and first-month installs, it is measuring onboarding, and I cover that separately in the guide on reducing Shopify app uninstalls.
None of this is an excuse. It is the reason the benchmark for your segment is 80% to 90% rather than 95%.
How do you measure your own churn rate without fooling yourself?
I spent close to two years in customer success at Social Snowball before moving into product design, and the churn number I trust is the one with its rules written down. These are the rules I would write.
- Count paying merchants only. The cohort is merchants who had paid at least one invoice at the start of the period. Trials and free plans get their own metric, trial to paid conversion. Mixing them in makes every benchmark comparison meaningless.
- Fix the churn date. For logo churn, use the day the subscription was cancelled. For revenue churn, use the end of the last paid period. Pick one definition per metric and do not change it when the number looks bad.
- Treat a payment hold as a hold. A store whose card failed is on hold, not gone. Count it as churn only when the subscription is cancelled, and track the held ones separately because they are the easiest merchants to keep.
- Treat a reinstall inside 30 days as a non-event. Merchants uninstall, change their mind and reinstall. That is not a churn and a new customer.
- Keep downgrades in the revenue number and out of the logo number. A downgrade is contraction. It belongs in GRR and it is invisible to logo churn, which is why you need both.
- Report all-cause and preventable side by side. Tag every churn with its reason from the Shopify uninstall survey or your own cancel flow. "Store is closing" stays in the headline number, because the revenue is gone. The preventable subset is the number product and customer success can move.
- Use cohorts, not a blended rate. A blended monthly rate falls automatically when you add installs, because new merchants have not had time to leave. Look at each install month's retention at month 1, 3, 6 and 12 instead.
- Compound when you annualise. 1 minus (1 minus monthly) to the 12th, never monthly times 12. Two percent a month is 21.5% a year, not 24%.
| Mistake | What it does to the number | Fix |
|---|---|---|
| Including trials | Churn looks two or three times worse than peers | Separate trial conversion |
| Blended rate while growing fast | Churn looks better every month you grow | Cohort retention curves |
| Ignoring downgrades | GRR overstated | Include contraction |
| Multiplying monthly by 12 | Annual churn overstated | Compound it |
| Stripping out "store closing" | Revenue hole understated | Tag it, keep it in all-cause |
Once the number is honest, get ahead of it. The usage patterns that show up before a merchant cancels are in the guide on churn signals.
Where to start
- Write down which of the three metrics each of your dashboards shows, and rename any chart that is ambiguous.
- Rebuild the headline number on paying merchants only, with the churn date rule fixed, and recompute the last six months.
- Add a cohort retention table by install month with month 1, 3, 6 and 12 columns.
- Tag every churn in the last quarter with a reason, and split the total into all-cause and preventable.
- Convert your annual GRR to the implied monthly figure and compare it with the 0.8% to 1.9% range for your segment.
Frequently asked questions
What is a good monthly churn rate for a SaaS app?
Converted from the annual medians, the surveys imply roughly 0.8% to 1% of revenue lost per month for the typical private B2B SaaS company, and 1.4% to 1.9% for product-led or sub-$5k contract segments. For a month-to-month e-commerce app measured on paying merchants, anything under 2% a month on revenue is within the normal range for the segment.
Should churn be measured on customers or revenue?
Both, and reported separately. Logo churn tells you how many merchants you are losing and is the right input for support and onboarding decisions. Revenue churn and GRR tell you the size of the hole and are what investors and your own forecasts use. For an e-commerce app the two usually diverge because the smallest plans churn the most.
Why is my churn rate higher than the published benchmarks?
Usually one of three things. The number includes trials or first-month installs, which measures onboarding rather than churn. The app is month-to-month and under $12k a year, which is the lowest band in every survey. Or it is a blended rate on a base that has stopped growing, so the early churn is no longer hidden by new installs. Check those before you assume the product is the problem.
Does gross revenue retention include downgrades?
Yes. GRR subtracts both cancelled revenue and downgraded revenue from the starting cohort's MRR, and ignores any upgrades. That is why it can never exceed 100% and why it is the stricter of the two retention metrics.