Why the Churn Rate Formula Gives Three Different Answers

What is the churn rate formula?

Analysts write the churn rate formula as customers lost during a period divided by customers at the start of that period, times 100, and the denominator is where one month of identical data splits into three different answers:

churn rate = (customers lost during period / customers at start of period) x 100

Work it on one SaaS month. The company starts March with 1,000 customers, adds 200 and loses 40. 40 divided by 1,000 is 0.04, and 0.04 times 100 gives a churn rate of 4.00%. That is the arithmetic finished. The rest of this page is about the 200 who joined mid-month, the annual contracts that could not have churned in March, and the fact that the 40 leavers were not average-sized accounts.

Why does the same month produce three different churn rates?

Three conventions for the denominator produce 3.70%, 4.00% and 6.00% from the same March. The simple method divides by the opening count and ignores that the base grew. The mid-period method averages the opening and closing counts, since a customer who joined on March 28 had three days to leave and the opening count never saw them. The revenue-weighted method drops customer counts and divides lost MRR by opening MRR.

March closed with 1,000 + 200 - 40 = 1,160 customers, so the average base is (1,000 + 1,160) / 2 = 1,080. The company opened March with $100,000 in MRR, and the 40 departing accounts carried $6,000 of it.

A second mid-period convention adds half the new customers to the opening count instead of averaging the endpoints: 1,000 + 100 = 1,100, giving 3.64%. Four correct calculations, one month, a spread from 3.64% to 6.00%. Publish the method next to the percentage or the percentage cannot be read.

What is the difference between logo churn and revenue churn?

Logo churn counts customers and revenue churn counts money, so they diverge whenever your departing accounts are not the size of your average account. In March the 40 lost customers were 4.00% of the base but carried 6.00% of the revenue, which means the average leaver paid $150 a month against a book average of $100. That 1.5x ratio is the whole gap between the two numbers.

Read them as a pair. Logo churn above revenue churn says small accounts are leaving and your enterprise book is holding. Revenue churn above logo churn, the March case, says the accounts you can least afford to lose are the ones going, and a headline of 4.00% understates the damage by half. Track both against monthly recurring revenue so account size stays visible.

How do annual contracts break a monthly churn calculation?

An annual customer cannot churn in eleven months out of twelve, so putting all of them in a monthly denominator drags the rate toward zero. Churn is only possible at a renewal decision, and a customer twenty days into a twelve-month term has no such decision in March. The fix is an at-risk denominator: count only the customers who could have left during the period.

Split the 1,000 opening customers into 700 on monthly billing and 300 on annual terms, of whom 25 renewed in March. The at-risk base is 700 + 25 = 725, and the same 40 losses give 40 / 725 x 100 = 5.52% against the 4.00% the blended denominator reported. Report monthly and annual cohorts separately, or the annual book flatters every monthly number you publish.

What is the difference between gross and net revenue churn?

Gross revenue churn counts lost MRR and downgrades, and net revenue churn subtracts expansion from surviving customers. Gross measures the leak. Net measures whether growth from your existing book outruns that leak, which decides whether the business grows without new logos.

March lost $6,000 in MRR to departing customers and $1,500 to downgrades from customers who stayed, against $9,000 in expansion from upgrades and added seats.

gross revenue churn = (6,000 + 1,500) / 100,000 x 100 = 7.50%
net revenue churn   = (7,500 - 9,000) / 100,000 x 100 = -1.50%

Gross churn of 7.50% and net churn of -1.50% describe the same month. Quote gross when you are diagnosing a retention problem, since expansion hides it, and quote net when you are forecasting, since expansion pays the bills. Both belong beside net revenue retention, the same arithmetic expressed as what you kept: 101,500 / 100,000 = 101.5%.

Why can negative churn happen?

Negative churn happens when expansion revenue from existing customers exceeds everything lost to cancellations and downgrades in the same period. March produced it: $9,000 of expansion against $7,500 of gross losses leaves the book $1,500 larger before a single new customer signed. Seat-based pricing and usage tiers create it.

Negative churn is only possible on the revenue side. Logo churn cannot go below zero, because a customer who leaves cannot leave twice and an upgrade does not create a new logo. A team quoting negative churn is quoting a revenue figure, so ask which one before reading it against your own.

Why can't you multiply monthly churn by 12?

Multiplying a monthly churn rate by 12 overstates annual churn, because each month's losses come out of a base the previous month already shrank. Churn compounds on the survivors, so the correct move is to compound the retention rate and subtract:

annual churn = (1 - (1 - monthly churn)^12) x 100

Take March's 4.00%. Monthly retention is 0.96, and 0.96^12 = 0.6127, so 61.27% of the base survives the year and annual churn is 38.73%. Multiplying 4.00% by 12 returns 48%, overstating the loss by 9.27 percentage points, or 93 customers on a 1,000-customer base.

The error runs the other way too. Dividing an annual churn rate of 30% by 12 gives 2.50%, while the correct calculation is 1 - 0.70^(1/12) = 2.93%. Anyone budgeting against the divided figure plans for less churn than the annual number they started from. Pair the compounded rate with your retention rate and your LTV to CAC ratio, since lifetime value is 1 divided by the churn rate and inherits every error in it.

How do you pin the churn rate formula to one denominator?

Write the denominator rule down before anyone builds the dashboard, then hold it fixed for at least four quarters. A churn number is only comparable to another churn number calculated the same way, and switching from a simple to a mid-period denominator moved the March figure from 4.00% to 3.70% with nothing happening to the business. Store the rule beside the metric, not in a footnote.

Flowsery pulls revenue by source from Stripe, Paddle, Polar and Lemon Squeezy, so the MRR you divide comes from the billing system instead of a spreadsheet copy, and it does not sample data. Once the rate is defined, use product analytics and session replay to see what the churned accounts did in the weeks before they left, because the formula counts how many went and never why.

Frequently asked questions

What is the churn rate formula?

Churn rate equals customers lost during a period divided by customers at the start of that period, times 100. 40 lost from an opening base of 1,000 is 4.00%. State the denominator with the result, because a mid-period average of 1,080 returns 3.70% from the same 40 losses.

Should churn be measured on customers or on revenue?

Measure both and publish both. Logo churn says how many accounts left, revenue churn says what they were worth, and the two only match when departing accounts are average-sized. In the March example the 40 leavers were 4.00% of customers and 6.00% of revenue.

How do I convert monthly churn to annual churn?

Compound the retention rate instead of multiplying the churn rate. Annual churn is 1 minus monthly retention raised to the twelfth power: at 4.00% monthly, 0.96^12 is 0.6127, so annual churn is 38.73%. Multiplying 4.00% by 12 returns 48%, which is 9.27 percentage points too high.

Can churn rate be negative?

Net revenue churn goes negative when expansion from existing customers exceeds cancellations and downgrades in the same period. March lost $7,500 gross and gained $9,000 in upgrades, giving net revenue churn of -1.50%. Logo churn cannot go below zero, since an upgrade does not create a customer.

How should annual contracts be handled in monthly churn?

Keep customers out of the monthly denominator unless their renewal falls in that month. An annual customer has one churn decision per year, so counting them in all twelve months drags every monthly rate down. Restricting the March base to 700 monthly customers plus 25 annual renewals raised the rate from 4.00% to 5.52% on the same 40 losses.

What is the difference between gross and net revenue churn?

Gross revenue churn adds lost MRR and downgrades, then divides by opening MRR. Net revenue churn subtracts expansion from that same numerator first. March was 7.50% gross and -1.50% net, and a team quoting only the net figure shows no retention problem while $7,500 a month walks out.

Why do two mid-period denominator methods give different churn rates?

One method averages the opening and closing customer counts, giving (1,000 + 1,160) / 2 = 1,080 and a rate of 3.70%. The other adds half of the new customers to the opening count instead, giving 1,000 + 100 = 1,100 and a rate of 3.64%. Both treat customers who joined mid-month as only partially exposed to churn, they just split that exposure differently. Neither is wrong, but a dashboard has to pick one and label it.

Why does net revenue retention move in the opposite direction from net revenue churn?

Net revenue churn and net revenue retention are the same March arithmetic read from different sides. Net churn is -1.50%, the $1,500 the book grew after $9,000 of expansion covered $7,500 of gross losses. Net revenue retention states the same result as what survived and grew: $101,500 against the $100,000 opening MRR is 101.5%. A negative net churn number and a retention number above 100% describe one identical outcome.

Why does the churn rate formula matter for calculating LTV?

Lifetime value is calculated as 1 divided by the churn rate, so any denominator choice that changes the churn rate changes LTV by the same proportion. The March figures move from 4.00% to 3.70% to 6.00% depending on the method, so an LTV built on the simple method already disagrees with one built on the revenue-weighted method before any other input changes. Pin the churn denominator before quoting LTV, or two teams end up computing different numbers under the same name.

How long should a company keep the same churn rate denominator?

Hold the denominator rule fixed for at least four quarters once you pick it. A churn number is comparable only to another churn number calculated the same way, and switching conventions moved March's own rate from 4.00% to 3.70% with no change to the business. Write the rule down before the dashboard gets built so nobody swaps it mid-comparison.