Churn is the revenue you lose from existing customers. Gross MRR churn counts only the losses. Net MRR churn subtracts what existing customers added through upgrades - and it can go negative, which is the single most valuable property a subscription business can have. Founders mix the two up constantly, which is why one company's "5% churn" is a crisis and another's is a rounding error. This guide separates them.
Gross MRR churn
Gross MRR churn rate = (churned MRR + contraction MRR) ÷ MRR at start of month
- Churned MRR: revenue from customers who cancelled entirely.
- Contraction MRR: revenue lost from customers who downgraded, removed seats, or got a discount.
Gross churn is always zero or positive. It answers "how leaky is the bucket?" and it is the number to watch when you're diagnosing product or pricing problems, because expansion can't hide it.
Example: you start the month at $10,000 MRR. Two customers on $200 plans cancel ($400) and one customer drops from $500 to $300 ($200). Gross MRR churn = $600 ÷ $10,000 = 6%.
Net MRR churn
Net MRR churn rate = (churned + contraction − expansion) ÷ MRR at start of month
- Expansion MRR: revenue added by existing customers - upgrades, extra seats, add-ons. (Reactivations of previously churned customers are usually counted under new MRR, not expansion, but some teams include them here; pick one and be consistent.)
Same month, but three customers upgraded for a combined $800 of expansion. Net MRR churn = ($600 − $800) ÷ $10,000 = −2%. Your existing customer base grew by 2% without a single new logo.
That's negative net churn (also called net revenue retention above 100%): existing customers are worth more each month than the month before. A business with negative net churn grows even if acquisition stops. It's why usage-based and seat-based pricing is so prized, and why investors ask for net revenue retention before almost anything else.
Side by side
| Gross MRR churn | Net MRR churn | |
|---|---|---|
| Includes cancellations | Yes | Yes |
| Includes downgrades | Yes | Yes |
| Subtracts upgrades / expansion | No | Yes |
| Can be negative | No | Yes |
| Best for | Diagnosing leaks | Judging overall health |
| Typical "good" (SMB SaaS) | < 3% / month | ≤ 0% (negative is excellent) |
Report both. Net churn alone lets a fast-expanding company hide a real cancellation problem; gross churn alone makes a healthy seat-based business look worse than it is.
Customer churn vs MRR churn
Both of the above are revenue churn. Customer churn (logo churn) is simply cancelled customers ÷ customers at start of month, and it can diverge sharply from revenue churn: losing five $9 customers is 5% of a 100-customer base but might be under 1% of MRR, while losing one $2,000 account is 1% of customers and 20% of revenue. Track logo churn to understand who leaves; track MRR churn to understand what it costs.
Benchmarks
Rules of thumb, not laws:
- SMB / indie SaaS, monthly billing: 3–7% gross monthly churn is common early; below 3% is good; below 2% is excellent.
- Mid-market: 1–2% gross monthly.
- Enterprise, annual contracts: measured annually; under 10% a year is healthy, and net churn is expected to be negative.
- Consumer subscriptions: often 5–10%+ monthly; the business model relies on cheap acquisition.
Early on, percentages are noisy: at $2,000 MRR one cancellation can be 5–10%. Use a three-month average before concluding anything.
Why the distinction matters for growth
MRR growth is new MRR plus expansion minus churn and contraction. As the base grows, a constant churn rate becomes a larger absolute number, and at some point it swallows all your new business: 5% gross churn on $50k MRR is $2,500 a month you have to re-sell just to stand still. Reducing gross churn by two points, or building expansion so net churn goes negative, raises your growth rate permanently - unlike a launch, which raises it once.
Practical levers, roughly in order of impact for small SaaS:
- Annual plans - churn is decided once a year instead of twelve times.
- Onboarding to first value - most cancellations happen in the first 30 days.
- Seat- or usage-based pricing - creates expansion without a sales team.
- Dunning - failed cards are involuntary churn; retry and email before giving up.
- Exit surveys - you can't fix what you don't hear.
Churn and verified MRR
HitMRR reads active subscriptions from Stripe every 30 minutes, so the MRR history on a startup page reflects churn the moment a subscription stops being active - not when a founder gets around to updating a number. Downgrades show as contraction in the same snapshot. If you want your growth curve to be believed, the churn has to be in it; connect Stripe and it is.