Ask an owner what payroll costs and you will get a percentage. Ask which classes lose money and the room goes quiet. The two are related: the percentage is an average across the whole schedule, and an average is exactly the number that lets a handful of losing classes hide inside a business that looks fine.
One invented studio runs through this piece: one room of yoga and mat Pilates, seventy classes a week (about 300 a month), $60,000 a month of revenue, and a flat rate per class that averages $60. Payroll from the books is $24,000 a month, four times the rent, and it has read 40% of revenue, give or take a point, for two years. Nine of its seventy weekly classes lose money anyway.
The full cost breakdown of a studio covers the other lines. This piece stays with payroll.
Instructor pay in one minute
- The percentage is a check, not a control. It says that something changed, never which class.
- Manage by the class. Payroll plus rent over the classes you ran is what a class costs; divide by what a visit earns for the headcount every slot must clear.
- Count the regulars before you cut. A losing class can be the only reason several members pay at all.
| The number | The sum | In the example | What it tells you |
|---|---|---|---|
| Payroll share | payroll ÷ revenue | $24,000 ÷ $60,000 = 40% | that something changed, not where |
| Cost per class | (payroll + rent) ÷ classes run | $90,000 ÷ 900 = $100 | what every class must cover |
| Break-even headcount | cost per class ÷ revenue per visit | $100 ÷ $20 = 5 people | the line to sort slots against |
| A slot's own sum | its pay + $40 shared, against what its people pay | Thu 7am: $75 + $40 = $115, break-even 6 | whether a slot near the line is over it |
| The retention check | only-class regulars × their dues | 4 × $170 = $680 a month | whether a cut saves money or loses members |
What percentage of revenue should a studio spend on payroll?
There is no percentage that is right for every studio, and this piece will not give you one. Your own is a check: if it moves several points without a decision behind it, something changed, and you should go and find it.
What it cannot do is say what to change, because it is worked out across every class at once. Under flat pay, the example's 40% would not move if attendance drained out of six more classes into the busy ones, since neither payroll nor dues would change. And it moves for reasons unrelated to pay: a price rise lowers it without anyone being paid less, and a slow August raises it without anyone being paid more.
The unit that changes decisions is the class: what it costs to run, and what it earns.
What does one class cost to run?
Payroll plus rent, divided by the classes you actually ran. Take both from the books for the last three closed months, so one odd week does not set the number. In the example, $72,000 of payroll and $18,000 of rent over 900 classes is a cost per class of $100.
Divide that by what a visit earns, revenue over attended visits, and you have the break-even headcount: $180,000 over 9,000 visits is $20, so a class needs five people to cover its $100.
What to do this week:
- Take three closed months of payroll and rent from the books, with payroll as the books have it: desk hours, payroll taxes and sub cover included.
- Divide by the classes that actually ran, not the ones scheduled. That is your cost per class.
- Divide it by revenue per attended visit, not per booking, and round up. That is your break-even headcount.
If you run a second location, give it a sum of its own: different rent, different room, different break-even.
How much should you pay fitness instructors?
Enough to keep the teachers your members come for, and no more than your classes can carry. Nobody outside your studio can tell you the first part. The second is arithmetic.
A class can carry pay up to what it earns at its usual headcount, less its $40 share of rent and non-teaching payroll. Five people earn $100 and can carry $60, which is exactly why five is the break-even at a $60 rate. Ten can carry $160. Every extra person in the room lifts the ceiling by $20.
So a rate is really a schedule question: the same payroll over fewer thin classes pays each teacher more. The only other lever is what a visit is worth, and a price rise has its own arithmetic.
The sum is the same for yoga, mat Pilates, reformer, barre or cycling; the room is what changes. Twelve reformers can never carry what a mat floor of thirty can at the same price per visit, which is why a reformer room is judged one session at a time and a cycling studio reads every ride against its own line.
Should instructors be paid per class or per head?
It depends on who should carry the risk of an empty room. Flat puts it on the studio, per head on the instructor, and a base plus a per-head bonus splits it. All three can pay the same at your average class; they differ only when the room is quiet or full.
Flat is predictable: the instructor can plan around the hour and the studio absorbs the swings, which is fairest to someone teaching a slot whose attendance they do not control. The 6am in January is not a performance.
Per head rewards the teachers who build a following. Applied bluntly, it also pays somebody less for teaching the same class well on a wet Tuesday, and quietly pulls your best teachers toward the slots that were already full.
A base plus a per-head bonus makes a quiet week survivable and shares the upside of a full one. Start the bonus at the break-even, as here, and the studio pays extra only for heads that already pay for themselves. Yoga room or Pilates floor, test your own rates at four headcounts before anyone's pay changes.
Whether an instructor is an employee or a contractor is a legal question that varies by state and turns on how the work actually happens, not on what an agreement says. Settle it with your accountant before you tune the rate.
Which classes are losing money?
The ones whose attended average sits under your break-even for months, not weeks. Sort every slot by attended average over the same three closed months and draw one line at the break-even. Whatever sits under it for the whole window is the shortlist.
It is a shortlist, not a verdict: one line treats every class as if it cost and earned the same per head, so redo the sum by hand for each slot on it and each one just above it.
That can move a slot either way. Thursday 7:00am averages 5.4 and clears the line, but it is taught by the most senior instructor at $75, so its own cost is $115, its own break-even six, and by hand it is under. Tuesday noon goes the other way, as the next section shows. Value each visit at what that person pays, too: an intro offer, a drop-in and a partner-app visit are not worth the same.
What to do this week:
- Sort every slot by attended average over the three closed months your cost uses.
- List the slots under the break-even for the whole window, plus any within one person above it.
- Redo each by hand: its instructor's pay plus its $40 share, against what its own people pay.
- Mark each: move or merge, check the regulars first, or give it time. A slot under three months old has not found its people yet.
Should you cut a class that loses money?
Not until you know who it is holding. Count the regulars whose only class is that slot. If there are several, it is a retention line, to be judged against what those members pay rather than what the class earns at an average visit.
Tuesday noon averages 3.5 people, so at $20 a visit it earns $70 against its $100 and loses $30 a session. Of its six regulars, four took no other class in three months, and all four pay $170 a month.
A cut also saves less than the $100: the rent stays, and usually so does the desk. What goes is the instructor's $60, about $260 a month, against $680 a month of dues. Those members will not cancel the week the class goes; they try other times, nothing fits, and they fade in a way nobody connects to the schedule. That is the slow fade a save note is written for, and it is cheaper not to cause it.
If a slot's regulars are spread across the timetable instead, move, merge or reformat it. The fill-rate piece covers that order, and the revenue leak calculator estimates what quiet churn costs a studio in a year.
And if one instructor's following carries a class, that concentration is its own risk. When that teacher leaves, the slot can go with them unless somebody has planned the handover.
Keeping the number honest
Payroll is also the number most likely to be wrong when an owner works it out herself: she sees the rate card every week and the rest of payroll hardly ever. The example's $60 average times 300 classes is $18,000 a month, or 30% of revenue. The books say $24,000.
None of that $6,000 is teaching and all of it is payroll. Leave it out and the cost per class falls to $80, the break-even to four, and five of the nine losing slots vanish from the list. An estimate that flatters payroll flatters every class.
So when your number and your accountant's differ, the books win. Decide once what counts, including whether your own pay is in, keep it that way so the trend means something, and read it monthly beside revenue per class and the other studio KPIs.
Where this lives
A spreadsheet can do every sum here, but it will not keep itself current as the months roll forward. That part is what Xyzios keeps. It works with Mindbody and is an approved Mariana Tek integration, and your booking platform stays the system of record. With QuickBooks connected, the Programming board shows an estimated cost per class, payroll plus rent over the classes run in the last three closed months, and one estimated break-even headcount per location, which each slot's attended average is compared against. The Finance board reads your P&L from the same books.
The Team view is built from the schedule, not payroll: who is actively teaching, how concentrated attendance is on a few instructors, which slots are fragile. Xyzios does not show what each instructor is paid, the pay cost of any one class, or a break-even for each slot; those sums stay yours. If a schedule change does cost you a regular, her attendance slipping against her own pattern is what the at-risk list flags while she is still a member, and a short save note is drafted for those at high risk. If she cancels anyway, the Sales Desk call sheet lists her for the next thirty days, with the reason where your platform records one, for a person to call; Xyzios drafts and sends her nothing. Nothing reaches a member until you tap approve.
Whatever you use, do the sum once this month. The two-minute studio check shows which of your numbers needs you first.