"3% monthly churn or less" is the benchmark everyone cites. It is not wrong, exactly — it is useless. A $5K ACV SMB tool and a $150K ACV enterprise platform churn at fundamentally different rates for fundamentally different reasons. And the number most articles report blends voluntary churn (the customer chose to leave) with involuntary churn (their credit card expired) — two problems with completely different solutions. Here is what a good churn rate actually looks like, decomposed by segment, pricing model, and the split that changes how you respond to it.
60–70%
Churn that is voluntary (product/pricing)
30–40%
Churn that is involuntary (payment failures)
3–5×
Churn variance by pricing model
What is a good monthly churn rate for B2B SaaS?
The common answer is 3% or less. The useful answer depends on who you are selling to. SMB SaaS companies (ACV under $10K) should target 1.5–2.5% monthly logo churn. Mid-market (ACV $10K–$75K) should target 0.5–1%. Enterprise (ACV above $75K) should be below 0.5%. Those ranges come from aggregated billing data, not a single operator's experience — and they only mean something once you separate voluntary from involuntary churn.
Customer Churn Rate
Percentage of customers lost in a given period — the clearest signal of product-market fit.
Why the segment split matters: a 2.5% monthly churn rate is a red flag for an enterprise company and a strong result for an SMB one. The difference is contract structure and switching costs. Enterprise buyers sign annual or multi-year contracts with procurement cycles that create natural retention. SMB buyers pay monthly, can cancel with a click, and often have three alternatives bookmarked. Comparing them to the same threshold is grading on a curve that penalises the wrong company.
The second layer is the voluntary/involuntary split. A company at 2% monthly churn where 1.2% is involuntary (failed payments) and 0.8% is voluntary has a billing infrastructure problem, not a product problem. A company at 2% where 1.8% is voluntary and 0.2% is involuntary has the opposite diagnosis. Same headline number, completely different remediation.
Logo churn vs revenue churn — which matters more?
Revenue churn matters more for valuation. Logo churn matters more for product-market fit diagnosis. They answer different questions and diverge in predictable ways.
Net MRR Churn Rate
MRR lost to cancellations and downgrades minus expansion — the revenue-weighted view of churn.
Logo churn counts customers. Revenue churn counts dollars. When your smallest customers churn at higher rates than your largest — the typical pattern — logo churn will exceed revenue churn. A company losing 4% of logos monthly but only 1.5% of MRR is shedding low-value accounts while retaining its core. That is not alarming in isolation. But it does signal that the bottom tier of the customer base is not finding enough value to stay, which often means the product's entry-level experience is weak or the pricing floor is wrong.
Revenue churn is what investors evaluate because it flows directly into MRR growth, LTV calculations, and DCF models. Logo churn is what operators should watch because it is a leading indicator. A spike in logo churn among small accounts today becomes a revenue churn problem among mid-market accounts in two quarters, as the pattern migrates upward.
The voluntary vs involuntary churn split
What involuntary churn looks like in billing data
Involuntary churn is a customer who did not choose to leave. Their credit card expired, was declined, or hit its limit. The subscription lapsed not because the customer evaluated alternatives and chose one, but because a payment failed and the retry logic gave up.
In Stripe, this shows up as a subscription moving to "past_due" or "unpaid" status after exhausting the retry schedule. The customer never clicked "cancel." They may not even know they've churned until they try to log in and find their account suspended. This is why involuntary churn is sometimes called "passive churn" or "delinquent churn" — the customer was passive, and the billing system made the decision for them.
Median split: ~60–70% voluntary, 30–40% involuntary
Across B2B SaaS, the median company sees 60–70% of its total churn from voluntary cancellations and 30–40% from involuntary payment failures. That involuntary share is higher than most founders expect — and it means roughly a third of the customers you are "losing" never decided to leave.
Companies with strong dunning — smart retry schedules, pre-expiry card update prompts, and in-app failed-payment banners — can recover 30–70% of involuntary churn. That recovery is the highest- ROI retention investment most SaaS companies can make because it requires no product changes, no pricing rework, and no customer success intervention. It is infrastructure.
Churn rate benchmarks by segment
By ACV tier (SMB, mid-market, enterprise)
ACV is the strongest single predictor of churn rate. Higher ACVs correlate with longer contracts, higher switching costs, and more implementation effort — all of which suppress churn mechanically. A $200K ACV enterprise deal required a 6-month sales cycle, a dedicated implementation, and an executive sponsor. The customer is not leaving over a rough quarter.
By vertical (horizontal SaaS, vertical SaaS, infrastructure)
Vertical SaaS companies typically churn 20–40% less than horizontal ones at the same ACV tier. The reason is competitive density. A vertical SaaS tool for dental practices competes with two or three alternatives. A horizontal project management tool competes with dozens. More alternatives mean more off-ramps. Infrastructure SaaS — databases, auth providers, payment processors — churns even less because switching costs are embedded in the customer's codebase.
By pricing model (per-seat, usage-based, flat-rate)
Pricing model creates a 3–5x variance in churn rate, even within the same ACV tier. Usage-based models have lower logo churn (the customer scales down before canceling) but higher revenue churn (contraction is built in). Per-seat models have moderate churn that tracks headcount changes. Flat-rate models have the highest logo churn because the cancellation decision is binary — all or nothing.
| Segment | Pricing Model | Monthly Logo Churn | Monthly Revenue Churn | Involuntary Share |
|---|---|---|---|---|
| SMB (<$10K ACV) | Per-seat | 2.0–3.0% | 1.5–2.5% | 35–40% |
| SMB (<$10K ACV) | Flat-rate | 3.0–4.5% | 3.0–4.5% | 30–35% |
| Mid-market ($10K–$75K) | Per-seat | 0.5–1.0% | 0.3–0.8% | 25–35% |
| Mid-market ($10K–$75K) | Usage-based | 0.3–0.7% | 0.5–1.5% | 20–30% |
| Enterprise (>$75K) | Per-seat | 0.2–0.5% | 0.1–0.3% | 15–25% |
| Enterprise (>$75K) | Usage-based | 0.1–0.3% | 0.3–1.0% | 10–20% |
The churn diagnostic framework
Once you have the voluntary/involuntary split and the segment breakdown, the diagnostic becomes a 2×2 matrix. Each quadrant points to a different root cause and a different remediation.
High voluntary + low involuntary → PMF or pricing problem
Customers can pay but choose not to. Their billing is fine. They evaluated the product and decided it wasn't worth renewing. This is the most common pattern in early-stage SaaS and the hardest to fix because the remediation is product or positioning work, not infrastructure. Look at cancellation surveys (if you have them), feature adoption data, and time-to-value. Are customers activating the core use case? If not, churn is an onboarding problem masquerading as a product problem.
Low voluntary + high involuntary → payment recovery problem
Customers want to stay but their payments fail. This is the highest-ROI quadrant because the fix is mechanical: implement smart retry logic, send pre-expiry card update emails, and add in-app payment failure banners. Companies that invest here typically reduce involuntary churn by 40–60% within 90 days. The customer already chose to stay — you just need to collect their money.
High both → onboarding failure
When both voluntary and involuntary churn are elevated, the root cause is usually upstream: customers are signing up but never reaching the activation moment. They enter their credit card, start a trial or first month, and disengage before forming the habit. Some cancel deliberately (voluntary). Others simply forget and let the subscription lapse (involuntary). The signal is that churn concentrates in cohort months 1–3. If customers who survive past month 3 retain well, the product works — the funnel before it does not.
How to calculate SaaS churn rate correctly
Monthly vs annual churn (don't annualize monthly naively)
A common mistake is annualizing monthly churn by multiplying by 12. That overstates the number. A 3% monthly churn rate does not equal 36% annual churn — it compounds to about 31% annual churn. The formula is 1 − (1 − monthly rate)^12. The difference grows as the monthly rate rises: 5% monthly compounds to 46% annual, not 60%. Always use the compounding formula when converting.
Gross MRR Retention
Percentage of MRR retained from existing customers before expansion — the floor of your retention.
Cohort-based churn vs period churn
Period churn divides customers lost this month by customers at the start of the month. It is simple but misleading when the customer base is growing quickly, because new customers dilute the denominator. A company that adds 100 customers and loses 10 reports 10% churn on a base of 100 — but if it started the month with 200 and added 100, the period churn is 10/200 = 5%. Both numbers are "right" by different definitions.
Cohort-based churn tracks a fixed group of customers from their sign-up month forward. It eliminates the denominator problem because the starting population is fixed. Cohort churn curves also reveal whether retention improves after an initial drop-off — the classic "month 2 cliff" followed by flattening — which period churn masks entirely. For benchmarking, use period churn. For diagnosis, use cohort churn.
What causes high churn — and what to fix first
The five most common causes of elevated churn, in order of diagnostic frequency: poor onboarding (customers never reach the activation moment), pricing misalignment (the plan the customer bought does not match their usage), weak competitive moat (alternatives are close substitutes and cheaper), payment failures without recovery (no dunning, no retry logic), and bad-fit acquisition (marketing attracts users the product was not built for).
Fix payment recovery first. It is the fastest win because it requires no product changes — only billing infrastructure. Then fix onboarding, because it affects every cohort going forward. Then address pricing and competitive positioning, which are slower to implement but compound over time. Bad-fit acquisition is last because it requires marketing and sales discipline, not product work — and it is invisible until you have clean cohort data showing which acquisition channels produce the highest-churn customers.
Tracking churn from billing data
North Metric calculates customer churn rate, net MRR churn, and gross retention directly from Stripe billing events. Each metric is decomposed by segment and tracked at the cohort level, so the voluntary/involuntary split and the logo/revenue divergence are visible by default — not buried behind a single number.
The point is not replacing judgment with automation. The point is removing the spreadsheet step that sits between billing data and the diagnostic framework described above. When churn spikes, the first question is always "what kind?" — and that answer should be visible in the dashboard, not trapped in a formula someone built last quarter and nobody remembers how to update.