You asked what itgrosses. Here's wherethat money goes.

Buildings ask me a version of the same question, usually right before they sign something. What does this gross? What has it sold so far? It is the correct question, and the number on its own will still mislead you — because almost nobody who hears a gross figure hears it as gross. They hear profit.

I run micro markets and AI smart coolers across the Las Vegas Valley, and I am not going to put my own figures on a public page. Be wary of any operator who does: a blanket figure published for everyone is quoted out of somebody else's building. What I can do is hand you the published industry structure, walk down the lines every dollar of gross passes through before anyone may call it profit, and give you five questions worth more than any figure an operator volunteers.

Gross is the top line. There are seven lines under it

These shares come from operator-facing models published by equipment and software vendors — people selling to operators rather than to buildings, which is why they are candid. Treat them as the shape, not as anyone's books.

The product itself runs roughly 45–52% of sales across published unattended-retail and micro-market models, and it splits hard by category: drinks carry the strongest product margin, candy the thinnest. Before anyone has driven to your building, about half the top line is committed.

Service and restocking is the line buildings underestimate most. One published micro-market model puts labour and route service near 17% of sales at a single healthy site. That is not the driving — it is counting, packing, carrying, facing the shelf, cleaning the glass and dating the perishables, every week, good week or bad.

Card processing lands between about 3% and 6% depending on how it is priced, and interchange-plus pricing adds a fixed charge per transaction on top of the percentage. That bites harder in unattended retail than in a shop: nearly every purchase is a card and the basket is small, so the fixed fee lands on a one-item basket rather than a full trolley.

Software and connectivity is around 2% in the same model — the kiosk platform, the telemetry, the cellular connection that keeps the unit off your network. Shrink has a published range of 2–5% for micro markets, with a worked case settling near 3% after layout and camera changes. Supplies and waste add under 3%, and waste is the honest one: anything with a date on it either sells or is thrown away.

Whatever goes to the building sits in the same stack. Published models put a host commission anywhere from 0% to about 15% of sales, and operators structure the building's side differently — some pay a commission, some fund fixtures or enclosures, some neither. Whichever shape yours takes, it comes out of the same gross as everything above it. That is why the shape of the stack matters more than the headline figure.

Add those up and roughly three-quarters to four-fifths of gross is gone before the parts nobody lists: insurance, permits, licensing, taxes, and the money that bought the hardware. The same models land at an operating margin of about a fifth of sales at a well-placed site — and the equipment is paid for out of that fifth, not out of the top line.

The part that is genuinely local

Nevada requires a food establishment health permit for machines dispensing temperature-controlled foods and for micro markets, inspected at least annually. Licensing sits separately and does not follow the valley: the City of Las Vegas, North Las Vegas, Henderson and unincorporated Clark County are distinct licensing authorities, so an operator covering the valley is permitted and licensed in several places at once.

None of that scales down for one building. A single site carries its share of a compliance burden sized for the whole footprint — one of the real reasons operators run a minimum headcount rather than a sliding scale. Ours is around a hundred people on site, and below it we tell buildings to wait; the format comparison sets out the bands.

“It pays for itself in a year” is one line out of a model

This comes up often, from someone told what the equipment costs who then did sensible arithmetic with it. Published payback models for this class of equipment run roughly 12 to 30 months. The fast end assumes a strong site, a full year at mature sales rather than ramp-up, and little or nothing going to the building. Move any one of those three and payback moves by a year or more.

So it is fair to raise, and I would rather you raised it than sat on it. It is just not a conclusion — it is the top line of a model whose other seven lines have not gone anywhere. Hardware getting recouped is where the thing starts to work, not proof it was already working.

A building that understands the stack negotiates better and stays longer. A building that signs believing gross is profit finds the gap at the first renewal — with a year of goodwill already spent.

Five questions worth more than a gross figure

1. What did this site gross in its first ninety days, and what does it gross now?

Ramp-up is real and slow, because habit is what is actually being installed. A mature number held next to a new one is not a comparison.

2. Show me item-level sales, not a total

A total hides everything useful. It cannot tell you whether a large part of the shelf never moves — and on any unattended shelf a meaningful share of it doesn't, ours included. Item-level data is where you find out whether your people are being served or merely charged. What actually moves on our shelves, ranked, if you want a sense of it first.

3. How is the equipment paid for, and what happens to it if we end the term?

Ownership and exit decide more of the economics than the price of anything. Ask who owns the box, who carries a bad charge and how a person gets it fixed without filing anything, and who physically pulls it out of the wall at the end of a term. An operator who answers that cleanly in writing has answered it before.

4. What is on my meter, and what is on yours?

Space and power are genuinely the building's, and I won't pretend otherwise. Everything else above should be itemisable on request. An operator who can't separate those two groups for you hasn't separated them internally either.

5. What would make you walk away from my building?

The one I would ask if I only got one. An operator with a stated minimum, and a stated point at which a site stops making sense, has done the arithmetic above on their own business. One who tells you every site works has not — and you meet that again in month nine, when the shelf is thinner and the visits are further apart and nobody can say why.

The number I would watch instead

Gross tells you how much money moved and nothing about how much of your building is being served. A market two-thirds of your people use weekly and one a tenth use daily can post the same total, and they are not the same asset — the first survives a bad month and a renewal conversation. Ask for the count of distinct accounts using it, not the total it rang up. If an operator can give you one and not the other, that is itself an answer.

What none of this settles is your terms. Those get decided in a conversation about your specific building, by people who have seen your room and your headcount. Everything above is what makes you good at having it.

If you are earlier than that and still working out whether a market or a cooler fits your building at all, the headcount tool on the homepage runs the bands and will tell you to wait if waiting is the right answer. And if you already have a gross figure in front of you and want to know what to put against it, tell me about your building below — whether or not it ends up being us.