A gym user attending a gym doesn't incur significant additional staffing costs. The work required to clean the gym may increase a bit but that'd be nowhere near the price of even the cheapest gym memberships.

> The marginal cost of each user rises too, as people that are unable to get on machine are more likely to churn.

That's not a cost, it's a factor for the overall churn rate and marginal cost for acquiring a new user. It'd limit the max profitability of a gym (i.e. getting to churn = growth equilibrium) but it's not an operational cost.

But that's not the right way to gauge "cost per member-usage". If you get a lot more usages, you will have to pay noticeably more for staff, maintenance, replacement, etc.

For comparison, how would you gauge the actual cost of sending a letter in the mail? Naively, you would say "well, you can't detect the cost of sending it in labor or fuel costs, so it's zero".

But that would be wrong. The right way is to say "how much would it cost to send a million more letters? The cost of one letter should be treated as one millionth of that" because, in the large, your marginal costs will scale that way as letter transmission goes up and down.

So yeah, the gym is, for all relevant purposes, bearing a small cost each time you use the gym.[1] It's just that, for the vast majority of members, all those usages cost them much less than the membership fee.

[1] I've heard figures of $1/visit but can't find at the moment.

They are on staff anyways. It costs them nothing extra.

What is true is that they can understaff (and have less equipment) than if they had 100% regular participation.