The rate of customers or users who stop using your product over a given period.
In practice
Two churns worth telling apart: logo churn (number of customers lost) and revenue churn (amount lost) — a customer who downgrades without leaving counts in the second, not the first. Another useful split: voluntary churn (the customer decides to leave) vs. involuntary churn (a failed payment) — the latter is often fixed with plain billing mechanics. The strongest health signal a SaaS team can aim for is "negative churn": expansion revenue from existing customers (upsell) outpacing what churn takes away.
The formula
Churn rate = customers lost during a period ÷ customers at the start of that period
- Customers lost
Those who were customers at the start and aren't at the end. Don't count the ones who signed up and left within the same period — that's a different problem (activation), and it distorts the rate.
- Start of period
The denominator is the opening count, not the average or the closing one. Using the closing count during fast growth makes churn look lower than it is.
- Revenue churn
Same formula with revenue instead of customers: MRR lost (cancellations + downgrades) ÷ MRR at the start. Net revenue churn also subtracts expansion from existing customers, and can go negative.
Churn and retention are two faces of one number: 5% monthly churn is 95% monthly retention. Retention is usually read by cohort over time (D7, D30, month 6); churn is usually read as a rate per period. Same customers, two views.
Worked example
One month, three churns
- Start of month: 400 customers, €20,000 MRR. During the month: 12 cancel (€600 MRR), 8 downgrade (−€400 MRR), 15 upgrade (+€900 MRR).
- Logo churn: 12 ÷ 400 = 3%.
- Gross revenue churn: (600 + 400) ÷ 20,000 = 5%.
- Net revenue churn: (600 + 400 − 900) ÷ 20,000 = 0.5%. Had upgrades brought €1,200 instead of €900, net churn would be −1%: negative churn, a revenue base growing on its own.
Three true numbers for one month, from 3% to 0.5%. Whoever reports "our churn is X" without saying which one is either simplifying — or choosing the flattering one.
Orders of magnitude
- Commonly cited ranges for B2B SaaS: around 1-2% monthly logo churn is considered healthy for high-ticket B2B SaaS. Low tickets run much higher: ChartMogul puts the median at about 6% a month under $25 of monthly revenue per account, against about 2% above $500. Enterprise products aim for single-digit annual churn, because their customers are fewer and each loss is large. B2C subscriptions routinely run far higher and live off re-acquisition.
- Small monthly numbers compound: 3% a month is not 36% a year but 1 − 0.97¹² ≈ 31% — a third of the customer base gone every year. Read monthly churn in years before deciding it's fine.
- The best SaaS businesses report negative net revenue churn — existing customers grow faster than the lost ones shrink the base. It is rare, and it is the single strongest signal that a product has become indispensable.
How to improve it
- Find the cause before the fix. Exit interviews, a one-question cancellation survey, and a look at what churned customers did in their first two weeks usually name the top reason within a month. Guessing produces features nobody was leaving over.
- Separate voluntary from involuntary churn. Failed payments, expired cards and billing errors can be a quarter or more of churn in subscription businesses, and they are fixed with payment retries and card updaters, not with product work.
- Fix the first two weeks. Most customers who churn at month three decided at week two: they never reached the moment where the product paid off. Onboarding is a retention lever wearing an activation costume.
- Build a reason to come back — a weekly report, an alert, a shared artefact — so the product re-enters the customer's week on its own instead of waiting to be remembered.
In the Tour
Do you know your main cause of churn?
- Yes, backed by data or interviews20 pts
- A hunch, not confirmed7 pts
- No idea0 pts
Two Retention questions touch churn. "Do you track a retention rate?" measures whether you can see it; "Do you know your main cause of churn?" measures whether you understand it — 20 points for a cause backed by data or interviews, 7 for a hunch, 0 for none. A team that tracks the rate but can't name the cause typically lands in the 10-15 band: the dashboard exists, the diagnosis doesn't.
Questions people ask
How do I convert monthly churn to annual churn?
Not by multiplying by twelve. Annual churn = 1 − (1 − monthly churn)¹². At 2% monthly that is 1 − 0.98¹² ≈ 21.5%, not 24%; at 5% it is ≈ 46%, not 60%. The gap grows with the rate, and the multiplied version always overstates.
What is negative churn?
Net revenue churn below zero: over a period, the extra revenue from existing customers who upgraded or bought more exceeds the revenue lost to cancellations and downgrades. The customer base can shrink in logos while growing in revenue. It requires a pricing model with room to grow — seats, usage, tiers — and it is the reason "expansion playbook" is one of the Tour's Revenue questions.
Is a churned customer who comes back still churn?
Count the churn when it happens and the return as a reactivation, separately. Netting them out hides both stories: why they left, and how you won them back — the second being a channel worth knowing. Most subscription analytics tools report reactivations on their own line for exactly this reason.