SaaS Churn Rate Benchmarks 2026: What's a Good Rate by Segment
There is no single good SaaS churn rate. ChartMogul's median monthly logo churn is 6.5% below $300K ARR and 3.1% at $8–15M. Recurly's SaaS median is 3.22% annual. Benchmarkit GRR sits at 84%.
Customer churn is the share of paying accounts that cancel in a period. Revenue churn is the share of recurring revenue those cancellations — and downgrades — take with them. Net revenue retention (NRR) is the same cohort twelve months later after expansion is added back. In 2026 the published “good” numbers still disagree by a factor of four, because they are measuring different companies on different clocks.
Growth Centr publishes evergreen, research-backed analysis on unit economics for founders and operators who need a peer check they can take into a board meeting, not a blended average that hides the segment they actually sell to.
ChartMogul, on anonymized billing data from 2,500+ private SaaS businesses, puts median monthly logo churn at 6.5% below $300K ARR and 3.1% at $8–15M ARR. Recurly's July 2026 network puts median annual SaaS churn at 3.22%, with a top quartile at 1.78% or below. Benchmarkit's 2026 report (CY-2025 actuals, 342 B2B SaaS and AI-native companies) puts median gross revenue retention at 84% — 16 points of annual gross revenue churn — and median NRR at 102%. All three are real. None is your number until you match sample, period, and formula.
Key Takeaways
- Logo churn counts accounts. Revenue churn counts dollars. GRR is 1 minus gross revenue churn and cannot exceed 100%. NRR adds expansion and can. Report logo and GRR together or expansion will hide a leak.
- ChartMogul median monthly logo churn: 6.5% (<$300K ARR), 3.7% ($1–3M), 3.1% ($8–15M). Compounded, 3.1% monthly is ~31% annual — not Recurly's 3.22% annual. Different books, different clocks.
- Recurly (July 2026): SaaS median annual churn 3.22% (voluntary 2.16%, involuntary 1.06%). Top quartile 1.78%. Involuntary churn at $250+ ARPC is 0.18% — a billing problem, not a product problem.
- Benchmarkit CY-2025: median GRR 84% (from 88%), top quartile 91% (from 95%). Median NRR 102%; usage-based 108%, seat-based 98%. Expansion already supplies 40% of net-new ARR at the median.
- A “good” rate is the one for your ACV, contract length, and billing mix. SMB monthly is not enterprise annual. Test the definition in 30 days before you set a board target.
How to Calculate Churn (And Stop Mixing the Formulas)
Most teams argue about the number because they are computing different ones.
Logo (customer) churn. Paying accounts lost in the period divided by paying accounts at the start. ChartMogul subtracts same-period join-and-churn and churn-and-reactivate so new trials do not inflate the rate. Exclude free and trial. Example: 100 at start, 12 cancel, one reactivates, one had joined that month → (12 − 2) ÷ 100 = 10%.
Gross revenue (MRR) churn. Recurring revenue lost to cancellations and contractions, divided by starting MRR. No expansion. This is the dollar leak. GRR = 1 − gross revenue churn, usually reported annually, capped at 100%.
Net revenue churn / NRR. Starting ARR + expansion − contraction − churned ARR, divided by starting ARR. New logos never enter. Above 100% means the installed base grew without a single new customer. Aleph's write-up of the same Benchmarkit book is the clean formula: NRR = (Starting ARR + Expansion − Contraction − Churn) ÷ Starting ARR.
Monthly is not annual times twelve. Recurly is explicit: 2% monthly compounds to about 22% annual, not 24%. Formula: 1 − (1 − monthly)^12. A 5% monthly logo rate compounds to ~46% of the starting base gone in a year. Mixing a monthly ChartMogul print with an annual Recurly print is how “we are 10× worse than benchmark” decks get written.
Voluntary vs involuntary. Recurly splits the annual rate: the customer cancelled, or a card failed. Involuntary is recoverable with dunning. Voluntary is a product, onboarding, or ICP problem. Treat them as two queues.
Benchmark Table: Monthly Logo Churn by ARR
ChartMogul's customer-churn page is the public monthly logo table. Use it if you are on monthly billing and your book looks like private SaaS on a billing platform — not if you are a $50M ARR enterprise with multi-year contracts.
| ARR band | Median monthly logo | Top quartile | Top decile | Compounded annual (median) |
|---|---|---|---|---|
| <$300K | 6.5% | 3.2% | 1.5% | ~55% |
| $300K–$1M | 4.1% | 2.5% | 1.4% | ~39% |
| $1–3M | 3.7% | 2.2% | 1.3% | ~36% |
| $3–8M | 3.8% | 2.3% | 1.3% | ~37% |
| $8–15M | 3.1% | 2.0% | 1.5% | ~31% |
| $15–30M | 4.1% | 1.7% | 1.3% | ~39% |
ARPA is the other cut that actually predicts churn. ChartMogul median monthly logo: 6.1% below $25 ARPA, 3.1% at $100–$250, 2.2% at $500–$1K, 1.8% above $1K. Price and switching cost, not a motivational poster, is why enterprise looks “better.”
Companies with NRR ≥100% run a median 1.7% monthly logo churn on the same ChartMogul book. Companies with NRR below 60% run 7.3% — double. High new-business growth also prints higher churn: >100% annual growth sits at a 5.3% median monthly logo rate versus 3.1–3.5% in the 25–75% growth bands. Early virality is not retention.
The Other Clock: Annual Billing-Network and Board-Survey Numbers
Recurly's July 2026 network is an annual subscriber-churn book across subscription verticals, not a monthly SaaS-only sample.
| Vertical (Recurly, July 2026) | Total annual | Voluntary | Involuntary |
|---|---|---|---|
| SaaS | 3.22% | 2.16% | 1.06% |
| Business & professional services | 3.44% | 2.27% | 1.18% |
| Travel, hospitality & entertainment | 3.91% | 2.63% | 1.28% |
| Digital media & entertainment | 4.14% | 2.55% | 1.59% |
| Ecommerce | 4.25% | 2.87% | 1.38% |
| Education | 4.99% | 3.30% | 1.69% |
SaaS is the stickiest vertical on that network. Recurly's own rule of thumb: below 2% annual is strong; 2–4% is the well-run zone; above 5% annual is worth an investigation regardless of vertical. Top-quartile SaaS is 1.78% or below. At $250+ ARPC, involuntary churn collapses to 0.18% — failed payments stop being the story once the invoice is large and the payment method is a real card or ACH.
That 3.22% annual figure is not a rebuttal of ChartMogul's 3.1% monthly. Recurly's SaaS cohort is more established, more annual-plan, and reported on a twelve-month clock. ChartMogul's median company below $300K ARR is still finding who stays. Put them on one slide without a label and you will fire the wrong team.
Finance-grade private B2B is the third clock. Benchmarkit CY-2025, with Aleph: median GRR 84% (down from 88%), top quartile 91% (down from 95%). That is a market-level slide, not one company's onboarding miss — top performers moved too. Median NRR 102% (n = 230), 25th percentile 92%, 75th around 108–110%. Usage-based pricing prints 108% median NRR; seat-based prints 98%, below the break-even line, as AI-driven headcount cuts shrink seats. Companies growing >50% sit at 111% NRR; companies growing <10% sit at 92%. Sub-$5M ARR median NRR is 94%; the $25K–$50K ACV band leads at 105%.
Read those next to B2B SaaS CAC benchmarks for 2026: expansion CAC is $0.80 of S&M per $1 of ARR versus $1.63 for a new logo, and expansion already supplies 40% of net-new ARR at the median. A 16-point GRR hole is a CAC problem wearing a retention badge. You cannot demand-gen your way out of a leaking bucket — you will pay the new-name tax twice.
Why Published 2026 Numbers Disagree
Three honest datasets, three jobs.
Billing platforms (ChartMogul) measure monthly logo and MRR movement on the companies that send them invoices. The book is heavy with early-stage and self-serve products. Median monthly logo churn of 3–6.5% is what that population does. ChartMogul's own NRR table on the same sample has median NRR still below 100% until the $15–30M band (93.4% median, 105.3% top quartile).
Billing networks (Recurly) measure annual subscriber churn across SaaS, media, ecommerce, and education. A 3.22% SaaS annual median is the number for businesses whose subscribers look like Recurly's network — more annual plans, higher ARPC, fewer seed-stage experiments.
Board surveys (Benchmarkit / Aleph) measure annual GRR and NRR on 342 private B2B SaaS and AI-native companies that filled in a finance questionnaire. Median 84% GRR and 102% NRR is the later-stage private-company print. It will not match your ChartMogul dashboard if you are at $800K ARR on monthly plans.
ChartMogul's 2025 AI retention cut (3,500 categorized companies, $250K+ ARR) is the fourth warning label: B2B SaaS median NRR 82% (upper quartile 97%), B2C 49%, AI-native 48% NRR and 40% GRR. AI products above $250/month look like B2B (70% GRR, 85% NRR). Below $50/month they do not (23% GRR, 32% NRR). Easy to buy is easy to cancel. That is a pricing and workflow-embed problem, not a CS-headcount problem — the same structural split as usage versus seat pricing.
Product-led motions inherit the high-volume, lower-ARPA row of the ChartMogul table. That is not a failure of PLG. It is the math of monthly self-serve. Pair it with NRR and GRR, not with an enterprise annual logo target.
What Is Actually a Good Churn Rate in 2026
There is no useful single average. Use the row that matches how you bill and who pays.
| Your motion | Logo churn that is not a fire drill | Revenue / retention pair |
|---|---|---|
| Early self-serve / PLG, <$300K ARR, monthly | ~5–7% monthly logo is median, not a crisis. Trend it down. | NRR often <100%. Get GRR moving before you scale paid. |
| $1–10M ARR, mixed monthly/annual, SMB–mid-market | ~2–4% monthly logo. ChartMogul median is 3.7–3.8%. | Aim for GRR climbing and NRR crossing 100%. |
| Mid-market, $10K–$50K ACV, mostly annual | Sub-2% monthly logo; Recurly-style annual in the 3% band is plausible. | Benchmarkit: $25K–$50K ACV leads NRR at 105%. |
| Enterprise / high ARPA, $50K+ ACV, annual or multi-year | Sub-1% monthly logo. Recurly $250+ ARPC involuntary is 0.18%. | GRR 90%+ and NRR 110%+ is the board-quality band. |
| Usage-based B2B | Logo can look noisy; dollars matter more. | Median NRR 108% in Benchmarkit; 75th percentile much higher. Do not manage to logo alone. |
| Seat-based B2B in 2026 | Logo may be fine while seats shrink. | Median NRR 98% — below break-even. Watch contraction, not just cancels. |
Bessemer-style folklore still holds as a directional overlay, not a law: 100% NRR is the floor, 110% is better, 120%+ is best-in-class — for the later-stage, high-ACV companies that folklore was written for. SMB and early PLG beating 100% NRR is a genuine win. Enterprise sitting at 100% is average.
If involuntary is more than about a third of logo churn, fix billing before you hire more CS. Recurly's SaaS split is 2.16% voluntary / 1.06% involuntary. A 16× dunning ROI is a collections program with a KPI, not a feature flag.
A 30-Day Operator Test
Do not set a “we will hit 1% monthly” OKR until the definition matches the book.
Days 1–7 — Name the four rates. Write logo, gross revenue churn, GRR, and NRR on one page, each with period (month vs year), who is in the denominator (paying only), and whether expansion is in. Recompute last quarter. If finance, CS, and the billing tool cannot produce the same logo number in one sitting, stop targeting and reconcile.
Days 8–14 — Split the leak. Tag last-90-day churn as voluntary, involuntary (failed payment), contraction (downgrade, seat loss), and “never activated.” If involuntary is the cheap win, turn on retries and updater before you rewrite onboarding. If never-activated dominates month-1, that is a product-led time-to-value problem, not a renewal-call problem.
Days 15–21 — Benchmark the right row. Pick ChartMogul if you are monthly and sub-$15M ARR. Pick Recurly if you are annual-plan SaaS comparing subscriber churn. Pick Benchmarkit GRR/NRR if you are reporting to a board on private B2B. Do not average them. Put your ACV and pricing model next to the number.
Days 22–30 — Move one lever, then re-read. One of: dunning, month-1 onboarding, ICP (stop selling the accounts that always churn), or expansion quota. Recompute GRR and NRR on the same cohort. If logo improved and GRR did not, you kept cheap accounts and lost dollars — a CAC problem. If you need the playbook after the diagnosis, use predictive churn reduction, not another blended target.
Methodology
This is a 2026 planning brief, not a survey we ran. Monthly logo, ARPA, and NRR-banded churn figures are from ChartMogul's public customer-churn benchmarks (2,500+ private SaaS businesses) and ChartMogul's 2025 AI retention analysis (3,500 categorized companies, $250K+ ARR). Annual subscriber churn, voluntary/involuntary split, and ARPC cuts are from Recurly's churn-rate benchmarks page, figures updated with July 2026 network data. Annual GRR, NRR, pricing-model split, expansion share of net-new ARR, and expansion-versus-new-logo CAC are from Benchmarkit's 2026 B2B SaaS and AI-Native Performance Benchmarks (1 June 2026, with Aleph; 342 companies; NRR n = 230; CY-2025 actuals), cross-checked against Aleph's public NRR write-up. Compounded annual equivalents of monthly rates use 1 − (1 − monthly)^12. No statistic appears here unless it was on a page we fetched. ChartMogul is monthly billing-tool data; Recurly is annual network subscriber churn; Benchmarkit is a finance survey. They are not interchangeable.
Read Next
- B2B SaaS CAC Benchmarks 2026: What It Costs to Acquire a Customer
- B2B SaaS Churn Reduction Tactics Using Predictive AI for Subscription Businesses
- What Is Demand Generation vs Lead Generation?
- SaaS Pricing Models Guide with AI Optimization Tactics
- What Is Product-Led Growth (PLG)?
FAQs
1. What is a good SaaS churn rate in 2026?
The useful answer is a range for your segment. ChartMogul median monthly logo churn is 6.5% below $300K ARR and 3.1% at $8–15M. Recurly's SaaS median is 3.22% annual, top quartile 1.78%. Benchmarkit median GRR is 84% (16% annual gross revenue churn) with 102% NRR. Match sample and period before you call any of those “the” benchmark.
2. How do you calculate SaaS churn rate?
Logo churn = paying customers lost in the period ÷ paying customers at the start. Gross revenue churn = MRR lost to cancel and contraction ÷ starting MRR. NRR adds expansion back and excludes new logos. Do not annualize by multiplying monthly by 12; compound it.
3. What is the difference between logo churn, GRR, and NRR?
Logo churn counts accounts. GRR counts dollars kept, ignoring expansion, and cannot exceed 100%. NRR counts dollars kept plus expansion, and can exceed 100%. A company can print low logo churn, ugly GRR (large accounts left), and pretty NRR (expansion covering the hole). You need at least GRR and NRR on the same slide.
4. Why is my churn so much higher than Recurly's 3.22%?
You are probably looking at a monthly rate, an early-stage or SMB book, or both. 3.1% monthly compounds to ~31% annual. Recurly's 3.22% is an annual subscriber rate on a different network. Compare ChartMogul's ARR band if you bill monthly; compare Recurly if you are annual-plan SaaS; compare Benchmarkit GRR if you are presenting to a board.
5. Does involuntary churn count?
Yes. A failed card is still a lost account and lost MRR. Recurly's SaaS split is 2.16% voluntary and 1.06% involuntary annual. At $250+ ARPC involuntary drops to 0.18%. Audit dunning before you treat the whole rate as a product verdict.
Disclaimer: This content is provided for informational purposes only and does not constitute financial, investment, or operating advice. Churn, GRR, and NRR figures reflect publicly reported research as of August 2026, from studies with different formulas, periods (monthly vs annual), and sample frames. ChartMogul numbers are billing-platform medians; Recurly numbers are subscription-network annual rates; Benchmarkit numbers are finance-survey medians. Treat them as directional peer checks, not board targets without your own cohort data.
