What a Yoga Studio Actually Makes: Break-Even, Rent, and the Numbers That Decide It

A studio that is working tends to land between 8 and 18 percent EBITDA. The range is wide, and almost none of it is explained by how well you teach. It is explained by two decisions made before a single class is on the schedule: how much space you signed for, and at what rent.

This post is the reasoning behind our studio profitability calculator. The calculator is free, needs no account, and every constant in it is editable, because every constant in it is a modelled estimate rather than a quote. What follows uses its default scenarios so you can click through and change any assumption you disagree with.

One caveat up front, and it applies to every number below. These are modelled figures, not survey data. They describe a stabilized year, not a ramp. They are EBITDA, so no debt service, no build-out amortization, no tax, and no owner draw beyond the manager's salary. Nothing here is financial advice. Take your own numbers to an accountant who knows your market.

The fixed cost trap

In the model, between 85 and 91 percent of a studio's monthly cost does not move with attendance. Rent does not care whether four people come or forty. Neither does insurance, software, cleaning, accounting, or the front desk.

Teacher pay is the one people misclassify. It is set by your schedule, not by your attendance. Put a class on the calendar and you owe the teacher whether it fills or not. That makes it behave like a fixed cost even though it feels variable, and it is why cutting a chronically empty class saves real money while adding one more student to a full class costs almost nothing.

The consequence is high operating leverage in both directions. Every student above break-even is worth close to the full price of their visit. And a ten percent drop in visits does not cut ten percent of your cost. It cuts almost none of it, and the entire shortfall lands on the bottom line.

The lease decides more than the schedule

Here is the clearest illustration the model produces. Two studios, same market, same price, same schedule shape, same 340 members. The only difference is about 600 square feet of rentable space.

Identical revenue. The smaller studio earns $3,405 more per month, about $40,860 a year, and needs roughly a tenth of the cash reserve. The extra 600 square feet was never producing anything. It was only ever a bill.

That is the whole argument for measuring twice before you sign.

The rent line

Below 12 percent of revenue is healthy. Between 12 and 15 percent is workable and worth watching. Above 15 percent, the lease is running the business, and every operational decision after that is made in service of covering it.

To make it concrete: take the big-room studio and move the rent from $43 to $47 per square foot. Break-even moves from 297 members to 305, and EBITDA falls from $7,392 to $5,920. Four dollars per square foot cost you eight members and a fifth of your profit. Negotiating that four dollars off the asking rent is worth about as much as a long stretch of very good retention, and you do it once, in an afternoon.

Capacity vanity and the 1.5x rule

The most common expensive design error is building the room for the class you hope to teach on a January Saturday rather than the one you will actually teach on a Tuesday at 2pm.

The big-room scenario above runs at 22 percent seat utilization. A 60-mat room averaging 10 students is paying rent on 50 empty mats, 32 times a week.

A working rule: size the main room to roughly 1.5 times your realistic average class size, not to your peak. Peak demand is three or four slots a week. You pay for the room 168 hours a week. If your peak classes are genuinely turning people away, that is a scheduling problem with a scheduling solution, which is to add a class at the peak time rather than square footage for all time.

Seasonality, and the reserve nobody holds

Studio demand is strongly seasonal. The cost base is not.

January runs about 25 percent above trend. The summer trough runs 15 to 20 percent below it. December is usually the worst month of the year. Utilities move in the opposite direction, peaking in exactly the months revenue bottoms out.

So a studio that is comfortably profitable across a full year still has cash-negative months, usually two to four of them. The default scenario has three, and the model suggests holding about $26,052 against them.

This is the number most first-year studios do not have, and it is why a profitable studio can still fail. Being annually profitable does not pay an August rent bill.

The aggregator question, answered honestly

Third-party marketplaces net you less per visit than your own channels. In the default scenario, turning aggregator visits on at 20 percent of traffic drops revenue per visit from $25.44 to $22.25, a 12.5 percent dilution.

And EBITDA goes up, from $7,392 to $12,798.

Both of those are true at once, and the reason matters more than either number. That studio is running at 22 percent utilization. The seats those visitors fill were empty and earning nothing. Diluted revenue on a seat that was producing zero is still a gain.

Flip the utilization and the answer flips with it. In a studio whose peak classes already fill, the same visits displace full-price students and train your local market to reach you through a cheaper channel. The threshold worth watching is around 15 percent of total visits, but the real question is not the percentage. It is whether those visits are landing in your empty Tuesday afternoon or your full Saturday morning.

We wrote more on the strategic version of this in Discovery Platforms vs. Your Own Funnel.

Where software actually sits

Studio management software is not usually a studio's largest problem, and it is one of the very few line items you can change in a week rather than at lease renewal.

Swapping the default scenario from an enterprise-tier platform to a self-hosted one moves EBITDA from $7,392 to $7,831 and break-even from 297 members to 294. That is $439 a month, about $26,000 over five years, and three members of break-even.

Real, and worth a look once a year. Not business-saving, and anybody telling you otherwise has something to sell. Compare that to the $3,405 a month the room size was worth and you have the correct ordering of priorities.

Tandava is one of the self-hosted options in that list: open source, MIT licensed, and you pay for hosting rather than a license. That is the entire pitch, and we would rather you fix your lease first.

Two notes on the software figures. The calculator deliberately does not name vendors, because published pricing moves constantly and real invoices depend on plan, size, modules and negotiation. It offers market positions with editable estimates instead. When you want the named, sourced version with links you can check yourself, that is a separate piece: What Studio Software Actually Costs, alongside the feature comparison.

Break-even, the way to actually compute it

Reason about it per visit rather than per month.

1. Take every cost that does not move with attendance, which is most of them.
2. Subtract whatever your teacher trainings and workshops net, since that covers fixed cost without needing mats.
3. Divide by your contribution per visit, which is average revenue per visit minus the costs that scale with it, mainly card processing and revenue-linked marketing.

That gives break-even visits. Divide by classes per month for students per class. Multiply by your member share of visits and divide by visit frequency for break-even members.

The number worth writing on the wall is margin of safety: how far attendance can fall before the studio stops paying for itself. Under 10 percent, you are running the business from the bank balance rather than from the plan.

Five studios to click through

Each of these is a starting point, not a claim about your market. Change anything you disagree with.

The heated studio looks like the winner, and on these inputs it is: one room, no second lease, high visit frequency, and a price the format supports. It is also the most exposed if the format goes out of fashion locally.

The right-sized two-room looks like the loser, and that is an artifact of the comparison. It is modelled at 300 members against the boutique's 340. Level the member count, as we did at the top of this post, and it earns $3,405 a month more. Comparing two studios at different member counts tells you about the member counts, not about the studios.

The math is public

Every formula and every constant in the calculator lives in one readable file, src/lib/studioEcon.ts. Each number in it is either sourced or explicitly labelled an assumption, the modelling caveats are in the header, and the file is MIT licensed like the rest of Tandava.

That is not a claim that the model is right. A transparent model built on estimated inputs still produces estimated outputs. It is a claim that you can go and check what it believes, and change the parts that are wrong for your market either in the drawer on the page or by sending us a correction.

Run it on your own numbers.