Micro market, smart cooler,or vending: which one fitsyour Las Vegas office?

Almost every version of this question gets answered the wrong way round. People ask which format is best. The useful question is which format your building can keep full — because a great fixture that runs half empty is worse than a modest one that never does.

I run micro markets and AI smart coolers across the Las Vegas Valley, so I have an obvious interest here. I'll try to be useful anyway, including about the cases where the answer is that you shouldn't do this yet.

What the three things actually are

A vending machine holds roughly forty selections behind glass and dispenses them mechanically. Everything about it is a consequence of that: the range is narrow, the packaging has to survive a drop, and there is a small but real chance the thing you paid for gets stuck on the way down. That last one matters more than operators like to admit — one jam is usually the end of a person's relationship with the machine.

An AI smart cooler is a single glass-door unit against one wall. You tap a card, the door unlocks, you take what you want and close it, and the card is charged for exactly what left the shelf. No scanning, no kiosk, no queue. Computer vision watches the shelf rather than the person — it tracks what moved, not who moved it — so putting something back is simply putting it back, and changing your mind carries no penalty.

A micro market is a small unattended store: open shelving, a cooler bank, chilled and shelf-stable food, and a self-checkout kiosk, usually set into millwork built for the room. It carries several times the range of a machine, including things you simply cannot sell through a spiral — a chilled meal, a bag of ice, a box of cereal bars, a bottle of Advil.

The three variables that actually decide it

Not the technology. These:

• Headcount on site, daily. Not employees on the payroll — bodies in the building on a normal Tuesday. Hybrid schedules have made these two numbers very different, and the second one is the only one that matters.
• Dwell time. A warehouse where people are on site nine hours with a thirty-minute break behaves nothing like an office where half the floor goes out for lunch. Long dwell and no easy exit is the strongest signal a format will get used.
• Whether people leave the building. If there's a good deli two minutes' walk away, you are competing with the deli, and you should plan for that rather than pretend it isn't there.

Where the lines actually fall

These are the bands I use, and I use them the same way whether or not it wins me the job.

Under about 100 people — not yet

This is the part most operators won't put in writing. Below roughly a hundred people on site, a market can't turn its inventory fast enough to stay fresh. Product ages on the shelf, the chilled range has to shrink to almost nothing, and within a few months you have exactly the thing you were trying to replace — a fixture people have learned not to trust.

The honest move at that headcount is to wait, or to find out whether other tenants in your building could be counted alongside you. A shared lobby market across three floors of a multi-tenant building often clears the bar when no single tenant does.

About 100 to 200 — smart cooler

This is cooler territory. One wall, tap to open, no build-out, no counter. If your headcount is climbing, this is also the moment to design the wall so a full market can drop into the same footprint later without redoing the space — which is trivial to plan for now and painful to retrofit later.

About 200 to 350 — micro market

At this headcount a market earns its floor space. You get the widest assortment, real food, and the thing that actually changes behaviour: people browse. A cooler is a decision; a market is a walk.

Above about 350 — a market, plus satellites

Past roughly three hundred and fifty, one point of service makes a queue at noon, and a queue at noon means people stop bothering. The answer is a full market anchoring the main break room and smart coolers on the floors furthest from it, so nobody has to plan a trip.

What doesn't change between the three

The service. Whichever format goes in, it sits on a fixed weekly route, the pick list is driven by live telemetry from your specific site rather than a route average, the planogram is photographed after every visit, and there's one named local contact rather than a ticket queue in another state.

That's worth saying plainly, because the format conversation tends to eat the whole meeting and the format is the part that matters least in month nine.

Installing it is the easy half. Every operator can do that. What decides whether people still use it next spring is what happens on the weeks after.

How to decide in about ten minutes

1. Count the people actually in the building on a normal Tuesday. Not the headcount on the org chart.
2. Ask whether they leave the building at lunch, and how far they go.
3. Look at where the traffic already goes — the corridor everyone walks, not the room you'd like to use.
4. Decide whether that spot has a wall and an outlet, or a corner and an alcove. That's the difference between a cooler and a market, physically.
5. If the number is under a hundred, ask whether the rest of the building can be counted with you.

If you want the shortcut, there's a headcount tool on the homepage that runs these bands and tells you what it would recommend — including telling you not to do it. And if you'd rather see the thing standing in your own room before deciding anything, that's what we do on the call: we work from your dimensions and photos and place it in your space in augmented reality, so you're looking at your actual room rather than a brochure.

And once you've picked the format, the next problem is usually internal: how to get a break-room market approved when you're the one selling it upward covers the four departments that can stall it and the one question each of them asks.