You want a number, and every coach who answers gives you a different one. That is not evasion — the number depends almost entirely on what you have promised each client, and two coaches with the same roster size can be doing jobs that share nothing but the job title.
So the useful version of this question is not "how many clients can a coach handle" but "how many can I handle, at the service level I sell, with the hours I actually work". That is a calculation, and it takes about ten minutes.
Start with hours, not clients
Take your working week and split it in two. Some of your hours are attributable to a specific client — writing their program, reading their check-in, replying to their messages. The rest are not: sales calls, content, your own admin, bookkeeping, the software you are learning.
For a full-time independent coach working a 40-hour week, the client-facing half usually lands somewhere near 25 hours. The other 15 disappear into everything that keeps the business existing, and if you have not allowed for it, you have not been counting the hours that decide whether you can take another client.
Your ceiling is 1,500 divided by what one client costs you in a week. Everything else in this article is about getting that second number right, because coaches routinely underestimate it by half.
Use your own figure if you know it. If you coach part-time around a job, the divisor might be 10 hours; if you have no sales pipeline to feed because you are full, it might be 30. The method does not change.
What one client actually costs in a week
Break the per-client time into four buckets and estimate each one honestly, using the way you work now rather than the way you would work on a good week.
Programming. A bespoke four-week block takes most coaches 30 to 45 minutes to write from scratch, plus mid-block adjustments. Spread across four weeks, that is roughly 10 minutes a week. If you build from templates and adjust, it is closer to 3. This is the bucket with the biggest gap between coaches, and the one where writing programs faster for multiple clients moves the number most.
Check-in review. Reading a week of training logs, weights, photos and answers, then writing a reply that is worth reading, takes 15 to 20 minutes if you do it properly. A skim and a "great work, keep going" takes two. Both are called a check-in, which is why capacity claims are so hard to compare.
Messaging. Ten to fifteen minutes a week for a client who messages normally. The cost is not the typing — it is that the messages arrive in twenty fragments across six days, and each one pulls you back into that client's context.
Admin. Ten minutes a week, minimum, and this is the bucket nobody counts.
Add those up for standard full-service coaching and you get 35 to 55 minutes per client per week. At 1,500 minutes that is 27 to 43 clients — which is why the honest answer to the headline question sits where it does, and why a coach claiming 80 full-service clients is either working 70 hours or not delivering what you are delivering.
The admin nobody counts
The ten-minute admin figure looks small until you list what is in it. Per client, per month, you are doing some mix of: onboarding a new starter, chasing a failed card, rescheduling a check-in that was missed, updating a program because of an injury or a holiday, answering a question about the app, writing notes so you remember the conversation next week, and reading back through three weeks of history because they have asked about something you half-remember.
None of it is coaching. All of it is time, and it scales linearly with the roster while your programming time does not.
There is a second cost that never appears in a time audit: context switching. Holding 40 clients in your head means every check-in starts with two or three minutes of reconstruction — what were we doing, what did they say last time, why is this weight lower. At 15 clients you remember. At 40 you look it up. That reconstruction cost is real, it grows with roster size, and it is the main reason coaches feel busier at 40 clients than the arithmetic says they should.
The lever here is not working faster. It is making the reconstruction unnecessary — history, metrics and past answers in one place per client rather than spread across a spreadsheet, a chat thread and your memory. Fitsly puts every check-in submission in a single review queue, with body metrics and progress photos flowing onto the client's profile, so the check-in and form tools are the same place you look for last month's answers.
Capacity by service level
Here is the same arithmetic run across five service levels. The time column is per client per week, all four buckets included. The roster column is 1,500 minutes divided by that time, rounded down to leave room for reality.
| Service level | What each client gets | Time per client per week | Realistic roster |
|---|---|---|---|
| Template group | Shared program, group chat, monthly form | 5–8 min | 100+ |
| Hybrid | Template adjusted per client, fortnightly check-in, group chat | 15–20 min | 60–80 |
| Standard full service | Individual program, weekly written check-in, direct messaging | 35–45 min | 30–45 |
| Full service plus nutrition | The above, plus macro adjustments and food-log review | 50–65 min | 20–30 |
| Bespoke plus weekly call | The above, plus a 30-minute video call every week | 90–110 min | 10–15 |
Two things to read out of that table.
The first is that the range across the top and bottom rows is roughly ten to one. Service level, not software and not efficiency, is what sets your ceiling. A coach who moves from weekly calls to weekly written check-ins has not become better organised; they have sold a different product.
The second is that the group row is not really limited by per-client minutes. At 100 clients in a shared program, what caps you is the volume of messages in one channel and the fact that you can no longer name everyone. Group capacity is bounded by attention, not arithmetic, which is a different constraint and a different set of problems.
What breaks first
The roster does not fail all at once. It degrades in a predictable order, and knowing the order lets you catch it early.
Check-in quality goes first. It is the most compressible part of the job — nobody complains about a shorter check-in immediately, and the ten minutes you save is available today. So the written response gets shorter, then generic, then it becomes three lines you could have sent anyone. The client notices about six weeks later, usually as a slow drop in effort rather than a complaint. Running check-ins that hold up at volume is mostly about having a fixed structure so the quality floor does not depend on how tired you are.
Message response time goes second. Same-day becomes next-day becomes "I'll get to it tonight". This one clients do notice, because your reply speed is the part of the service they experience every week.
Programming goes third, and later than coaches expect, because it is the easiest part to systematise. What actually happens is not that programs get worse but that they get less individual — last block copied forward with the weights nudged.
Admin goes last and hurts most. Failed payments sit for a fortnight, new clients wait four days for onboarding, someone's program runs out and nobody notices. These are the failures that cost you a client outright rather than slowly.
If you want the roster to stay whole, watch the first two. They are the leading indicators, and they move weeks before churn does.
The number you can hold and the number you can sell into
There are two capacity numbers, and confusing them is how coaches end up permanently at 90% full and permanently anxious.
The first is delivery capacity: how many clients you can coach at your standard. The second is sustainable capacity: how many you can coach while still doing the sales, content and onboarding that replace the ones who leave.
Churn makes that gap concrete. At a full roster of 35 with 6% monthly churn, you lose two clients a month, so you must sign two a month to stand still. Sales calls, follow-ups and onboarding for two new clients is a realistic four to six hours a month — hours that have to come out of the same week. If you are delivering at 35 with nothing left over, your roster is not stable at 35. It is stable at about 30, with five slots' worth of time reserved for keeping it at 30.
That is not an argument for staying small. It is an argument for either building the pipeline into the capacity plan or accepting that you will sell in bursts, coach at reduced quality during those bursts, and lose people because of it.
It also has a pricing consequence. If your capacity is 30 rather than 35, the price that funds your year is higher than the one you calculated, which is why capacity and price have to be worked out together rather than in sequence — pricing your packages properly starts from this number.
Why the last five clients cost the most
Going from 30 to 35 clients does not add one-sixth to your workload. It adds considerably more, for three reasons.
You lose the buffer. At 30 you had slack, and slack is what absorbed the week a client got injured, the week you were sick, and the Monday that vanished into a billing problem. At 35 there is no slack, so those weeks now come out of check-in quality instead.
Variance goes up. Clients are not average. One in ten is in some kind of crisis at any given time — a competition, a bereavement, a diagnosis, a holiday that derails everything — and that client costs three times the normal load for a month. On a roster of 35 you have three or four of them at once, permanently.
The tail gets longer. Every roster has clients who never message and clients who message constantly. Adding five more does not add five average clients; statistically, one of them is high-maintenance, and high-maintenance is the load that has no ceiling.
The practical version: if the last five clients feel like they cost double, that is not a failure of organisation. That is what the last five clients cost. Price them accordingly or do not take them.
Most coaches find the ceiling by going past it
Nobody works out their capacity in advance. They take clients until something gives, and then they know — usually about two months after the point where they should have stopped, because the damage lags the cause.
You can see it coming if you know what to look for:
- Check-ins get done on Sunday night in one batch, at speed, because there was no other time.
- You start reading a client's message, decide to answer it properly later, and later does not arrive.
- You cannot remember what you told someone without opening their file.
- You feel relief when a client cancels.
- New enquiries produce dread rather than interest.
The last two are the reliable ones. Relief at a cancellation is a capacity signal, not a personality flaw, and it shows up before the numbers do — which is the whole argument for watching for burnout as a business metric rather than a mood.
The other thing worth saying plainly: your ceiling moves. It goes up as your systems get better and as your client base gets more homogeneous, and it goes down when you take on clients outside your niche, when life gets busier, or when you add a deliverable and do not remove one. Recalculate it once or twice a year rather than treating the number you found at 28 as permanent.
Deciding what to do at the ceiling
Once you know the number, there are only four honest moves: raise prices and hold fewer clients, reduce the per-client time, sell a lower-touch tier to the clients who do not need high-touch, or bring in help. Working faster is not on the list, because you have already tried it.
Each of those changes what you sell rather than how hard you work, and they suit different coaches — that is the subject of scaling an online coaching business, which picks up exactly where this article stops.
One thing to do before any of them: publish your cap and mean it. A coach with a stated limit of 35 and a waiting list has a business. A coach who takes everyone who asks has a queue of people receiving a service that gets slightly worse each month, and boundaries around availability are what make the cap survive contact with a good prospect.
