Here is how to calculate churn rate: divide the customers you lost during the month by the customers you had at its start, not its end. MRR churn is the same idea in money: recurring revenue lost divided by MRR at the start. MRR itself is the sum of what every active customer pays per month, with annual plans divided by 12 and one-time fees left out. To calculate MRR and churn from real data, start from MRR at the beginning of the month and sort every event of the month into new, expansion, contraction and churned revenue. Below is one month of a small SaaS worked through line by line.
MRR and churn rate: what they are and why they matter
MRR (monthly recurring revenue) tells you how big the business is right now, without the noise of when invoices happened to be paid. Churn tells you how fast that business leaks. Together they answer the questions a founder asks every month: are we growing, and is the growth real or just replacing customers who left?
They are also the first two numbers anyone asks for when they consider buying or investing in the company.
A worked example: a booking app for driving schools
DriveSlot helps driving schools schedule lessons with instructors and cars. It has three ways to pay:
Plan
Price
MRR per customer, UAH
Basic, monthly
900 UAH a month
900
Pro, monthly
1,400 UAH a month
1,400
Basic, annual
9,000 UAH a year
9,000 / 12 = 750
Step 1. Calculate MRR at the start of the month
On 1 March DriveSlot has 70 Basic monthly, 10 Pro monthly and 12 Basic annual customers: 92 in total.
70×900+10×1400+12×750=63000+14000+9000=86000 UAH
Note the annual customers: they paid 9,000 once, but contribute 750 a month. Counting 9,000 in the month of payment would make MRR jump and fall for no real reason. Why annual plans are still worth offering, and at what discount: annual plan discount.
Step 2. Sort the month's events
The March log from the payment system and the admin panel:
Event
Type
MRR change, UAH
6 new schools on Basic monthly
New
+5,400
1 new school on Pro monthly
New
+1,400
2 new schools on Basic annual (paid 9,000 each)
New
+1,500
3 schools upgrade from Basic to Pro
Expansion
+1,500
1 school downgrades from Pro to Basic
Contraction
−500
4 Basic schools cancel
Churn
−3,600
1 Pro school's card fails, not recovered by month end
Churn
−1,400
1 annual school renews for another year
No change
0
A one-time 2,000 onboarding fee from the new Pro school
Not MRR
0
Two traps are in this list on purpose. A renewal of an annual plan doesn't change MRR: that school was already counted at 750 a month. And a one-time fee is revenue, but not recurring revenue.
Step 3. MRR at the end of the month
Component
UAH
Starting MRR
86,000
+ New MRR
8,300
+ Expansion MRR
1,500
− Contraction MRR
500
− Churned MRR
5,000
Ending MRR
90,300
Net new MRR is 90,300 − 86,000 = 4,300 UAH.
Check it by counting customers on each plan on 31 March: 70 Basic monthly (70 + 6 − 3 − 4 + 1), 12 Pro monthly (10 + 1 + 3 − 1 − 1) and 14 annual (12 + 2). That gives 70 × 900 + 12 × 1,400 + 14 × 750 = 90,300. The two methods agree, which is the best test that no event was missed.
Step 4. Calculate churn rate and MRR churn
Customer churn. Five schools left (four cancellations and one failed payment); 92 were there at the start.
customer churn=925≈5.4%
Gross revenue churn (gross MRR churn). Revenue lost to cancellations and downgrades, divided by starting MRR:
860005000+500≈6.4%
Some teams leave downgrades out and get 5,000 / 86,000 ≈ 5.8%. Either is fine; pick one definition and keep it.
Net revenue churn. Subtract expansion from the losses:
860005000+500−1500≈4.7%
When expansion from existing customers exceeds what you lose, net revenue churn becomes negative: the customer base grows in value even with no new sales. That's the goal, and price rises and upgrades to bigger plans are the usual way there.
What a 5.4% monthly churn rate means
A monthly number sounds small. Over a year it compounds:
1−(1−0.054)12≈49%
At this rate DriveSlot loses about half of any group of customers within a year, and the average customer stays about 1 / 0.054 ≈ 18 months. Every month, the first 5 or so new schools only replace the ones that left. The same loss seen from the other side is retention: 94.6% of customers stay each month; the retention rate formula shows how to track it for groups of customers over time.
At 5.4% monthly churn about half of a group of customers is gone after 12 months, against 69–78% still paying at the 2–3% a month that is good for small-business SaaS
That makes the reasons for leaving worth more attention than the next marketing idea. In DriveSlot's case one of the five was a failed card: an automatic retry and a reminder email could have saved it without any product change.
Common mistakes
Dividing by end-of-month customers. 5 / 96 ≈ 5.2% looks better and is wrong: the new customers weren't there to churn.
Counting annual payments in full. Divide by 12, always.
Counting one-time fees, setup work or refunds as MRR. They are real money, but not recurring.
Mixing customer churn and revenue churn. Losing five small customers and losing one big one look the same in customer churn and very different in revenue churn. And if one big customer holds a large share of MRR, check your customer concentration.
Ignoring failed payments. Involuntary churn is often the easiest to fix.
Key takeaways
MRR = what every active customer pays per month; annual plans ÷ 12; one-time fees excluded.
Ending MRR = starting MRR + new + expansion − contraction − churned. Check it by recounting customers per plan.
Customer churn = customers lost ÷ customers at the start; revenue churn = MRR lost ÷ MRR at the start.
Turn monthly churn into a yearly number to feel its size, and look at why customers leave.
FAQ
What is the churn rate formula?
Churn rate = customers lost during the period ÷ customers at the start of the period × 100%. For MRR churn, put MRR lost to cancellations (and, if you count them, downgrades) over MRR at the start. In the example: 5 ÷ 92 ≈ 5.4% customer churn and 5,500 ÷ 86,000 ≈ 6.4% gross MRR churn.
What is a good monthly churn rate for a small B2B SaaS?
It varies by market, but for products sold to small businesses, monthly customer churn of 2–3% is good and above 5% is a sign to investigate why customers leave. Small businesses close more often than large ones, so some churn is unavoidable.
Should free trial users be counted in MRR or churn?
No. Count a customer in MRR from the first payment. Trial users who don't convert are a conversion problem, not churn.
How do I count a customer who joins and leaves in the same month?
They weren't in the starting count, so they don't belong in the churn rate for that month. Track them separately; many of them point to onboarding problems.
Why do buyers care so much about MRR and churn?
MRR shows the size of the business and churn shows how durable it is. Together they drive most of what a small SaaS is worth; see how much a small SaaS business is worth.
Learn it hands-on
The lesson MRR and churn works through a month of a product for car repair shops, with numeric exercises and a practice task checked by AI. The next one, Why shops leave, is about the reasons behind the churn number. Both are part of the course From Idea to Profitable Product: Building Without Investors.
