What actually sellsin a break-room market.Our categories, ranked.

We already published the list we buy against — every category on our active assortment, with the counts. This is the other half, and the more useful one: what people actually take off the shelf. The two lists are not the same, and the gap between them is most of what I have learned running these.

A note on what follows: you get the shape of what sells and the order it sells in, not our volumes. The pattern is the part that transfers to your building anyway.

The industry split, and ours

Start with the published benchmark, because ours sits against it oddly. Industry census data for convenience services puts the category mix at roughly snacks 33%, beverages 29%, coffee 24%, and everything else 14%.

Ours is not shaped like that. Drinks are the clear majority of what leaves our shelves, snacks the balance, non-food a rounding error.

Two honest reasons, neither of them that we are cleverer. We run no coffee service, so that entire quarter of the benchmark does not exist on our shelves and has to land somewhere. And there is the valley: the well-travelled old statistic that cold drinks are north of half of vending sales traces to data from around 2009, and the industry has moved past it. We look like the old number anyway, because a building where nobody wants to walk to their car in August behaves like a building from a different decade.

The point for anyone comparing operators: category mix is local. A national planogram is an average of buildings that are not yours — and if the mix you are shown matches the national average exactly, nobody has looked at yours.

The top of the list isn't soda

Ranked by units, the best-selling item across our staffed-building markets is Smartwater. Fourth is C2O coconut water.

First and fourth are both water — a category most operators treat as the thing you stock so the cooler does not look empty. The entire soft-drink range sits below them. We carry more non-soda drinks than soda because that is what sells, not because it photographs well. If you take one thing into a conversation with any operator, make it this: in a desert, premium water is not the indulgence line, it is the anchor.

Three of the top five are spicy

The snack half of that leaderboard is where it gets specific. Second, third and fifth are Chester's Flamin' Hot Fries, Cheetos Crunchy Flamin' Hot and Munchies Flamin' Hot.

Three products, one flavour profile. Not brand loyalty — a heat preference, and not peculiar to us. NAMA, the industry body for convenience services, has flagged heat as the flavour story of the category, citing Euromonitor data that nearly 15% of new salty snack launches in 2023 carried bold heat, with demand starting among younger consumers and crossing generations from there.

The category that sells almost nothing, and earns its slot anyway

Non-food — pain relief, cold medicine, wipes, ChapStick — is the rounding error in that mix. It does not pay for its shelf space in units and it never will. We carry it anyway: nobody chooses a market because it has Advil in it, but the person who needed Advil at three in the afternoon and found it remembers, and that is the line between a market a building relies on and a fixture people buy chips from. An operator who only optimises the leaderboard cuts that shelf in the first review, and the building feels it long before anyone can point at a number.

The basket is one item, and a bit

Across our sites, the average transaction is barely more than a single item. People come for one thing.

That reframes the assortment question. The category that matters is whatever brought somebody across the room — overwhelmingly a drink — and the second item is a placement decision, not a category one. Which is why the same product looks dead on one shelf and moves on another: not the wrong product, just three feet from the thing people came for.

Nobody walks to the break room for a bag of chips. They walk for a drink and decide about the chips on the way.

Workplace and residential do not buy the same way

This one genuinely surprised me. Our apartment-lobby markets run noticeably larger baskets than our workplace break rooms. Residents are shopping; staff are grabbing. It makes sense once you see it — somebody on a short break has one errand and a clock, somebody crossing their own lobby at nine in the evening has neither. Same equipment, same city, different behaviour, so a category can be right for one setting and wrong for the other. Frozen is the clearest example.

The frozen lesson, which we paid for

At one of our residential markets the best seller in the whole market was a single-serve ice cream bar — while on the same shelves, over the same period, the ice cream pints barely registered, the frozen dinners went nowhere and a premium frozen dessert sold one.

Frozen was not the problem. The frozen assortment was inverted — built around a take-home grocery shop in a place where people buy something to eat now. Single-serve novelty works; pints, dinners and anything perishable that assumes a weekly shop do not, and the cost of that is not a lost sale, it is product spoiling on a shelf. If you are shown a residential proposal with a freezer in it, ask whether the frozen set is built for eating now or for a trolley.

The long tail is mostly dead, everywhere

The uncomfortable structural truth of the format, true of every operator including us: variety is the selling point, so the instinct is always to add. But on any unattended shelf a large share of the assortment sells nothing in a given month while the fast movers run out — over-assorted and under-stocked at once, which is commoner than it sounds and looks fine in a photograph.

I am not printing our number for this, and be sceptical of any operator who volunteers a flattering one. Ask for it instead. What share of the shelf sold nothing last month, and what came out as a result? Every operator has that number. Very few have looked it up, and the ones who have will tell you what they cut.

What I would take from this

1. Judge a proposed assortment on its drinks. In this valley that is the majority of what will sell, and it is where a generic planogram gives itself away fastest.
2. Look for heat. If the spicy set is one token item, the plan is older than the published data, never mind the registers.
3. Ask what share of the shelf sold nothing last month — and what got removed because of it. The second half of that question is the one that separates answers.
4. Make them say who the building actually is. Staff grabbing on a break and residents browsing a lobby want different shelves; an operator who does not ask will hand you the same one either way.
5. Ask which day you would be serviced, and why that day. On-site attendance now concentrates Tuesday to Thursday, so a schedule built for a flat five-day week services you on the wrong day. A good answer is about your building; a weaker one is about their route.

The caveat, because it matters

These are our buildings, in one valley. The rankings are real rather than estimated — more than most category advice can say — but they are not a national benchmark and I would not use them as one. What travels is the method: rank by units, look hard at what sold nothing, split the answer by the kind of building. Any operator worth hiring can do that for your site, and should offer to.

If you want to know which format a room like yours can actually carry, the fit tool on the homepage takes headcount and a few details about the space and gives you a straight answer, including when the answer is not yet. If you would rather ask a person, the form below reaches me.