You know roughly how many clients you have and roughly how many left last month, and you have a vague sense that too many are leaving. That is not a number you can act on. Until you calculate churn the same way every month, you cannot tell a bad month from a trend, and you cannot tell whether the thing you changed in March did anything.
This article is about the measurement: the formula, the four ways it goes wrong, and what the number tells you once you have it. Why clients leave and what to do about it are separate questions with separate answers — why online coaching clients quit covers the causes and how to retain online coaching clients covers the fixes.
The formula
Client churn rate is the share of your roster that left during a period.
Churn rate = clients lost during the period ÷ clients at the start of the period
Start of the period, not the end, and not the average. The denominator is the group that had the opportunity to leave.
Say you started October with 42 clients and five cancelled during the month:
5 ÷ 42 = 0.119 = 11.9% monthly churn
Use a month unless your roster is small — more on that below. Use the same window every time. A number you calculate monthly and one you calculate quarterly are not comparable, and switching between them is the most common way coaches convince themselves churn is falling.
Almost every "average coaching churn rate" figure you will find published is a vendor's marketing number with no stated method behind it — no sample size, no definition of a churned client, no date. Do not measure yourself against those. Measure yourself against your own number from three months ago.
What to do about clients who joined mid-period
If someone signed up on the 6th and cancelled on the 28th, do they count?
They are a real loss, but they were never in the starting roster, so putting them in the numerator while the denominator excludes them inflates the rate. Both conventions are defensible. What is not defensible is switching between them.
Take that October roster again. Five clients cancelled, and one of them had joined on the 6th of the same month:
| Method | Working | Result |
|---|---|---|
| All departures over starting roster | 5 ÷ 42 | 11.9% |
| Only starting-roster departures | 4 ÷ 42 | 9.5% |
Two and a half percentage points apart, on the same month, from the same data. Pick the first one — counting every departure never lets you flatter yourself, and same-month cancellations are worth seeing rather than hiding. Then write the definition down somewhere, because in four months you will not remember which one you used.
Clients who joined mid-period and stayed do not go in the denominator at all. They join the count at the start of next month.
Why a small roster makes the number jump
On a 12-client roster, one cancellation is 8.3% churn. Two is 16.7%. There is no such thing as 4% churn on twelve clients — the smallest non-zero number available to you is 8.3%.
So a month where two people leave and a month where one does looks like churn doubling. It is not. It is one person, and one person's decision to move cities is not a signal about your coaching.
The fix is to widen the window rather than the roster. Add up the losses across three months and divide by the average roster size, then divide by three to get back to a monthly figure:
4 clients lost over 3 months ÷ 3 = 1.33 lost per month
1.33 ÷ 12 average clients = 11.1% per month
Below about 20 clients, report the rolling three-month figure and nothing else. Above about 40, one client is 2.4% and the monthly number starts carrying real information. This is one of the reasons churn is worth watching alongside capacity rather than on its own — how many clients an online coach can handle covers where the ceiling actually sits.
Client churn and revenue churn are different numbers
If every client pays the same, they are the same number. If you run tiers, they can diverge badly, and the client number will be the reassuring one.
Take a 42-client roster on three tiers:
| Tier | Clients | Monthly fee | Monthly revenue |
|---|---|---|---|
| Foundation | 12 | $180 | $2,160 |
| Core | 25 | $320 | $8,000 |
| Premium | 5 | $600 | $3,000 |
| Total | 42 | $13,160 |
Now two different months, each losing five clients — 11.9% client churn both times:
| Who left | Revenue lost | Revenue churn | |
|---|---|---|---|
| Month A | 5 Foundation | $900 | 6.8% |
| Month B | 1 Foundation, 1 Core, 3 Premium | $2,300 | 17.5% |
Month A: $900 ÷ $13,160 = 6.8%. Month B: $2,300 ÷ $13,160 = 17.5%. Same headline, and one of those months costs you two and a half times as much as the other.
Calculate both. Client churn tells you about your coaching and your onboarding; revenue churn tells you about your business. When they separate, the direction matters: revenue churn running above client churn means you are losing your best-paying clients, which is a different problem from losing your cheapest, and it usually points at the top tier under-delivering rather than at the price. If you have not read how to price online coaching packages, the tier structure that creates this gap is worth understanding before you try to close it.
Your platform's payments reporting should give you the inputs without a spreadsheet — Fitsly's payments and packages area shows monthly recurring revenue, active subscriber counts per package and average lifetime value, which are the numbers you would otherwise be adding up by hand each month.
Turn churn into average client lifetime
This is where the number starts earning its keep. If churn is steady, average client lifetime is its reciprocal:
Average lifetime (months) = 1 ÷ monthly churn rate
At 11.9% monthly churn:
1 ÷ 0.119 = 8.4 months
Which is the same as 42 ÷ 5. The average client stays 8.4 months.
Multiply that by what the average client pays and you get lifetime revenue. Average revenue per client here is $13,160 ÷ 42 = $313 a month, so:
8.4 months × $313 = $2,629 in lifetime revenue per client
Two things follow. First, you now have a ceiling on acquisition spend. Spending $400 to sign a client is 15% of lifetime revenue and about 1.3 months of their fee — fine, if your delivery costs leave room for it. Second, that ceiling moves with churn, hard. Let churn drift to 20% a month and average lifetime drops to five months, lifetime revenue drops to $1,565, and the same $400 is now 26% of everything that client will ever pay you. Nothing about your marketing changed. The maths under it did.
The same reciprocal explains why. At 11.9% a month you retain 88.1%, and 0.881 compounded over twelve months is 21.9%. So roughly 78% of your roster turns over in a year, not the 143% that 11.9 × 12 would suggest. Always compound.
Voluntary and involuntary churn
Some clients decide to leave. Others just stop paying because a card expired, was reissued after fraud, or hit a limit. The second group is involuntary churn, and it is admin, not coaching.
Sort every departure into one bucket or the other before you draw any conclusion, because the fixes share nothing. Voluntary churn is a product problem. Involuntary churn is a dunning problem — card update prompts, retry schedules, an email that goes out on the failure rather than a week later.
The volume is easy to underestimate. Cards expire on roughly a three-year cycle, so about one in 36 of your clients' cards expires in any given month by simple arithmetic — on a 42-client roster that is more than one a month before you count reissues and declines. Every one of those becomes a cancellation if nobody chases it.
Treat a recovered failed payment as a client you did not lose rather than an invoice you finally collected — what to do about late payments covers the process side.
A blended number hides where the losses are
Churn is not spread evenly across a roster. It is concentrated early — the first two months, in most coaching businesses, before the client has seen enough change to believe the process works.
One monthly percentage averages a client who joined last week with one who has been with you two years, and those two are not remotely the same risk. Group signups by the month they joined and track each group down the page:
| Signup cohort | Signed up | End of month 1 | Month 2 | Month 3 | Month 6 |
|---|---|---|---|---|---|
| January | 10 | 8 | 6 | 5 | 4 |
| Loss that month | 2 (20%) | 2 (25%) | 1 (17%) |
Twenty per cent gone in month one, another 25% of what remained in month two, then it flattens. The blended figure for that business might read 11.9%, and 11.9% is not a number that tells you to fix onboarding. The cohort table is.
Run this for three or four cohorts and the shape becomes obvious: either you have an onboarding problem in the first eight weeks, or you have a long-term value problem where clients drift off at month seven or eight. They need completely different responses, and the blended number cannot tell you which one you have.
What to actually do with the number
Calculate it monthly, on the same definition, and record it. Three data points is a trend; one is trivia.
Then use it for these three things and stop there:
Churn is one of a handful of numbers worth tracking rather than the only one; the rest of the dashboard is in the KPIs every online coach should track.
