Retention
SaaS churn: why 2% and 5% a month are different businesses
By Assaf Schwartz · · 6 min read
Churn is the only input in a SaaS model that is simultaneously small enough to ignore in any given month and large enough to determine the maximum size of company you will ever build. A point of monthly churn is a rounding error in March and a different business over five years.
Three different churns
"Churn" is used for at least three distinct measurements, and conversations go wrong when two people mean different ones.
- Logo churn counts accounts lost as a share of accounts held. It is the right measure for support load, onboarding capacity and product-fit questions.
- Revenue churn counts recurring dollars lost. It is the right measure for financial modelling, because dollars pay the payroll and logos do not.
- Net revenue retention takes revenue churn and adds expansion — upsells, seat growth, usage increases — back in. Above 100% the existing base grows on its own.
Logo and revenue churn diverge whenever customers are unequal in size, which is always. A company can lose 5% of its logos and 1% of its revenue if every departure is small, or lose 2% of logos and 9% of revenue if a single enterprise contract walks. Reporting only logo churn in a business with a long tail of small accounts systematically hides the damage.
What compounding does
Monthly churn compounds. A cohort's survival after n months is (1 − churn)ⁿ, and over 60 months the differences become extreme.
| Monthly churn | Cohort left after 5 yrs | Avg. lifetime | LTV at $420 ARPA |
|---|---|---|---|
| 1.0% | 54.7% | 100 months | $32,760 |
| 2.0% | 29.8% | 50 months | $16,380 |
| 3.0% | 16.1% | 33 months | $10,920 |
| 5.0% | 4.6% | 20 months | $6,552 |
| 8.0% | 0.7% | 12.5 months | $4,095 |
Read the first and last rows together. At 1% monthly churn, more than half of a cohort is still paying you after five years. At 8%, essentially none of it is — you have replaced your entire customer base roughly five times over, and paid full acquisition cost each time.
Churn sets your ceiling
This is the most useful thing to know about churn and the least frequently said. If you add a roughly constant amount of new recurring revenue each month, your MRR does not grow forever. It converges on a ceiling:
steady-state MRR = new MRR per month ÷ monthly churn rate
At that point every dollar you add is exactly offset by a dollar churning out, and growth stops regardless of how well sales performs.
| Monthly churn | MRR ceiling | ARR ceiling |
|---|---|---|
| 1.0% | $1,500,000 | $18.0M |
| 2.0% | $750,000 | $9.0M |
| 3.0% | $500,000 | $6.0M |
| 5.0% | $300,000 | $3.6M |
| 8.0% | $187,500 | $2.3M |
A team adding $15,000 of new MRR a month at 5% churn will asymptote around $3.6M of ARR no matter how long it runs. The same team at 2% churn tops out at $9M. Nothing about the sales organisation changed — the retention rate alone moved the ceiling by two and a half times.
Net revenue retention, and why it can exceed 100%
Expansion revenue is churn's counterweight. If existing customers spend more over time — more seats, more usage, higher tiers — that growth is subtracted from gross churn to give net revenue retention.
NRR above 100% means the installed base grows without a single new logo. It is the single strongest signal in a SaaS business, because it makes growth compounding rather than additive, and it makes the ceiling above disappear entirely. Best-in-class B2B businesses run 110–130%; the very best usage-based businesses exceed that.
Benchmarks by segment
Acceptable churn depends almost entirely on who you sell to. Comparing an SMB tool to an enterprise platform on churn alone is meaningless.
| Segment | Typical monthly | Annual equivalent | Notes |
|---|---|---|---|
| SMB / self-serve | 3–7% | 30–58% | High volume, low CAC, churn is structural |
| Mid-market | 1–2% | 11–22% | Annual contracts damp month-to-month movement |
| Enterprise | 0.3–1% | 4–11% | Multi-year deals; churn arrives in lumps |
Note the last column of that caption. Annualising churn by multiplying by twelve overstates it — 5% monthly is 46% annually, not 60% — and the error grows with the rate.
What actually moves churn
In rough order of impact, and inversely to how often they get attention:
- Who you sell to. Most churn is decided before onboarding, at qualification. Customers acquired outside your best-fit segment churn at multiples of the rate, and no amount of customer success recovers that.
- Time to first value. The gap between signing and getting something useful out of the product predicts retention better than almost any feature.
- Contract length and billing period. Annual contracts do not improve the product, but they convert twelve monthly churn decisions into one.
- Depth of adoption. Accounts using several parts of the product, with several active users, churn far less than single-seat single-feature accounts.
- Involuntary churn. Failed cards and expired payment methods are a meaningful share of SMB churn and are the cheapest to fix — dunning and card-updater flows are pure recovered revenue.
Churn also sets the payback maths described in CAC payback vs. LTV:CAC — the two metrics share the same denominator, which is why retention work moves acquisition efficiency without touching a marketing budget.
Written by
Assaf Schwartz
Assaf Schwartz builds and maintains SimulateFin — the projection engine, the guides and the site around them. The methodology is published in full precisely so it can be argued with: corrections, disagreements and missing metrics are welcome by email.
Keep reading
Retention
NRR vs. GRR
Net revenue retention can exceed 100% while the customer base quietly empties. Gross revenue retention cannot …
Cash & capital
Runway, done properly
The back-of-envelope runway number assumes burn stays flat. It never does. A month-by-month projection with gr…
Capital efficiency
Burn multiple
Net burn divided by net new ARR. It is the one efficiency measure that punishes buying growth, which is exactl…
Growth vs. profit
Rule of 40
Growth rate plus profit margin should clear 40. The rule is simple; the trade-off it encodes is not. Here is h…
Educational content, not financial advice. Figures here are illustrative and exclude taxes, financing and one-off items. Model your own numbers in the simulator and check them with a qualified accountant before acting.

