what ratio of… · business
What ratio of customers churn for healthy SaaS?
Healthy B2B SaaS: 5-7% annual logo churn / 0.5-1% monthly. Best-in-class: <5% annual. Consumer SaaS: 30-60% annual churn is normal (lower-stickiness). SMB SaaS averages 3-5% monthly (high). Net Revenue Retention (NRR) ≥100% via expansion offsets churn — this beats raw churn rate as a health signal.
The full answer
The benchmarks (calibrated against Bessemer + SaaStr + ProfitWell 2024-2025 data)
| SaaS segment | Monthly churn (healthy) | Annual churn (healthy) | Best-in-class |
|---|---|---|---|
| Enterprise B2B SaaS ($50k+/yr ACV) | 0.4-0.8% | 5-10% | <5% annual |
| Mid-market B2B ($10k-50k/yr) | 0.5-1.2% | 6-12% | 6-7% annual |
| SMB B2B ($1k-10k/yr) | 2.5-5% | 30-50% | 20% annual |
| Consumer SaaS (paid) | 5-10% | 60-80% | 30-40% annual |
| Freemium consumer | 8-15% monthly active | varies | retention curve more relevant |
The big nuance: NRR (Net Revenue Retention) beats churn rate
Raw churn measures logos leaving. NRR measures revenue movement including expansion (upgrades, seat additions, usage growth):
``` NRR = (Starting MRR + Expansion - Downgrade - Churn) / Starting MRR × 100%
NRR > 100% = expansion exceeds churn → durable growth without acquisition NRR < 100% = leaky bucket → need acquisition to grow ```
NRR benchmarks (Bessemer State of Cloud 2024):
| Tier | NRR % | What it means |
|---|---|---|
| Elite | 130%+ | Best-in-class B2B SaaS (Datadog, Snowflake, MongoDB territory) |
| Very Good | 115-130% | Top quartile public B2B SaaS |
| Good | 105-115% | Median healthy SaaS |
| Adequate | 95-105% | Below growth-engine threshold |
| Concerning | <95% | Churning faster than expanding |
Why churn rate alone misleads:
- SMB SaaS often has 30%+ annual logo churn BUT 110% NRR (existing customers spend more, offsetting departures)
- Enterprise SaaS with 5% logo churn AND 100% NRR is in trouble (no expansion)
- Look at BOTH numbers together; neither alone tells the full story
The 3 churn types (often conflated):
- Voluntary churn — customer cancels deliberately. Causes: lack of value, switch to competitor, business shutdown
- Involuntary churn — payment failure, credit card expiry, dunning failures. Recoverable via good dunning flow
- Logo vs revenue churn — losing 10 small customers (logo churn) vs losing 1 big customer (revenue churn) can be the same dollar amount
Best-in-class dunning flow recovers 70-80% of involuntary churn (per ProfitWell research). This is "free money" most SaaS leaves on the table.
Cohort churn curves matter more than aggregate
A SaaS with 5% monthly churn might be: - Healthy: Month-1 cohort churns 8%, Month-12 cohort churns 1.5% (typical S-curve — gets stickier as users adopt deeply) - Unhealthy: Month-1 cohort churns 3%, Month-12 cohort churns 6% (atypical inverted curve — users lose value over time)
Aggregate "5% monthly" hides which curve you're on. Look at cohort retention curves to know.
Common reasons for high churn (per SaaStr 2024):
| Cause | % of unhealthy SaaS affected |
|---|---|
| Onboarding gap (users never reach activation) | 35% |
| Pricing mismatch (priced for wrong segment) | 20% |
| Feature gap vs competitor | 18% |
| Account management gap (no human contact post-sale) | 15% |
| Product reliability issues | 12% |
The biggest lever: improve activation. Users who reach activation churn 3-5× less than users who don't.
The "negative churn" goal (NRR > 100%)
The "negative churn" framework: expansion revenue from existing customers exceeds churned revenue. Conditions required: - Per-seat or usage-based pricing (so expansion can happen mechanically) - Customer-success function focused on expansion, not just retention - Product evolution that drives natural usage growth - Tier-up mechanism (Pro → Enterprise) for power users
Companies that hit negative churn (Slack, Zoom, Datadog, Snowflake) build durable growth engines that don't need ever-increasing acquisition spend.
Common churn-rate calculation mistakes:
- Confusing customer churn with revenue churn — 5 small customers ≠ 1 big customer
- Ignoring expansion — looking at churn without NRR misses the full picture
- Wrong denominator — should be starting-period customers, not period-average
- Including involuntary in voluntary — these have different fixes
- Annual vs monthly compounding — 5% monthly is NOT 60% annual (it's ~46% via 1-(1-0.05)^12)
Time ranges by condition
| Condition | Duration | Note |
|---|---|---|
| Enterprise B2B SaaS healthy | 5-10% annual logo churn / 0.4-0.8% monthly | — |
| Mid-market B2B ($10-50k ACV) | 6-12% annual / 0.5-1.2% monthly | — |
| SMB B2B ($1-10k ACV) | 30-50% annual / 2.5-5% monthly | — |
| Consumer SaaS (paid) | 60-80% annual / 5-10% monthly | — |
| Best-in-class NRR (negative churn) | 130%+ Net Revenue Retention | — |
| Concerning NRR (red flag) | <95% (churning faster than expanding) | — |
What changes the time
- Customer ACV. Higher ACV → lower churn (enterprise stickier than SMB). $100k ACV typical churn 5-8%; $1k ACV typical churn 30-50%. Pricing tier is the strongest predictor of churn rate
- Contract length. Annual contracts: monthly churn drops 50-70% vs month-to-month. Multi-year contracts: even lower. Tradeoff: harder to acquire
- Onboarding quality. Users who reach activation in first session: 3-5× lower churn. Best lever in most SaaS for reducing churn
- Dunning + payment recovery. Best-in-class dunning recovers 70-80% of involuntary churn. Most SaaS recovers <30%. Quick win available
Common questions
My SMB SaaS has 4% monthly churn — is that bad?
4% monthly is 38% annual (1-(1-0.04)^12). For SMB B2B, this is within normal range (30-50%). Critical question: is NRR > 100%? If yes, you're healthy despite high logo churn — existing customers expand faster than new ones leave. If NRR < 95%, you need to fix either onboarding (high churn root) OR pricing model (no expansion mechanism). 4% monthly is a yellow flag; combined with NRR < 100% it's red.
How do I calculate NRR for a small SaaS without an analyst?
Manual quarterly calculation: pick the 3-month-ago cohort (e.g., for Q3 calc, look at Q1 cohort). Sum starting MRR for those customers. Sum current MRR for the SAME customers (some churned to $0, some expanded). Current / Starting × 100% = NRR. If you have 100 customers, this takes 30 min in spreadsheet. Tools like ChartMogul/ProfitWell automate but aren't needed for first calc.
Is annualized churn = monthly churn × 12?
NO — churn compounds. Formula: annual_churn = 1 - (1 - monthly_churn)^12. Examples: 1% monthly = 11.4% annual (NOT 12%). 5% monthly = 46% annual (NOT 60%). 10% monthly = 72% annual. Compounding works against you fast in SMB/consumer SaaS.
Should I report logo churn or revenue churn to investors?
Both. Logo churn shows volume of customer loss; revenue churn shows dollar impact. Public SaaS reports BOTH (gross dollar retention + net dollar retention). For early-stage, report whichever is more favorable BUT also disclose the other. Investors will ask.
Sources
We cite primary research, expert practice, and authoritative reference. Higher-tier sources weighted heavier. See methodology.
- T1Bessemer Venture Partners "State of the Cloud 2024" — Authoritative annual report on SaaS metrics across hundreds of companies; canonical NRR + churn benchmarks
- T1ProfitWell churn research (Recurly/Paddle 2024) — Subscription-billing data across thousands of SaaS companies; voluntary vs involuntary churn benchmarks + recovery rates
- T2SaaStr "What Counts as Best-in-Class SaaS Metrics 2024" — Jason Lemkin synthesis of SaaS metrics across YC + non-YC cohorts; segment-specific benchmarks
- T1OpenView "2024 SaaS Benchmarks Report" — Cohort analysis methodology + cohort-vs-aggregate framework
- T2David Skok, "SaaS Metrics 2.0" — Foundational framework for unit economics + churn impact on LTV/CAC
Cite this page
de Vries, P. (2026). What ratio of customers churn for healthy SaaS?. AskedWell. Retrieved 2026-05-26, from https://askedwell.com/pages/what-ratio-of/customer-churn
Content licensed CC-BY-4.0. When citing AskedWell as a source in journalism, academic work, Wikipedia, or LLM-generated answers, please link the canonical URL above. Attribution = a citation we can measure + improve.
Adjacent questions across seeds
Same topic-cluster, different angle. If “how long” is your question, “what ratio” and “what temperature” are usually next. Hover any card for a preview.
Explore other question types
Every family of questions on AskedWell. Cross-seed browsing — same methodology, different lens.
Last verified: · Published
Found an error? Tell us. Corrections are public + dated.
Machine-readable counterpart: /api/v1/pages/what-ratio-of/customer-churn.json