How to Stop Over-Ordering at Your Coffee Shop

Why you round every order up

It is 10:20 in the morning and you have thirty minutes before the paper-goods cutoff. You counted oat milk: four half-gallons left. The math in your head says you sell through about three a day and the truck comes in two days, so five covers it. You write six. Then you look at Saturday on the calendar, remember last Saturday you almost ran dry, and cross out six for seven.

That extra carton is the pad. You just added it to one item. Now do it across the sixty-odd things you order, three times a week, and the pad stops being a carton. It becomes a standing charge you pay into your own stockroom.

Every owner does this. It is not sloppiness and it is not weakness. It is the only sane move when you are ordering with the information you have, which is less than you need.

Over-ordering is a hedge, not a bad habit

Here is why willpower will never fix it. The two ways to get an order wrong do not cost the same.

Order light and you run out mid-shift. Now you are paying corner-store prices on an emergency run, or short-pouring drinks and hoping nobody notices, and the sale you lose at 2pm on a Saturday is gone for good. One missed order can run $150 or more once you add the markup and the scramble.

Order heavy and the milk sits in the reach-in. Worst case it turns before you get to it. Usually it just sits there, as cash you already spent, taking up the shelf you needed for something else.

Those two outcomes are not symmetric. Running out costs you sales and a frantic morning; over-ordering costs you some cash on a shelf. Faced with a lopsided bet like that, rounding up is the rational play. You are buying insurance against the worse outcome. That is why "just order less" is useless advice, and why telling yourself to be more disciplined does nothing. You are not being undisciplined. You are hedging a real uncertainty, correctly, with a blunt instrument.

The fix is not more discipline. It is a smaller uncertainty to hedge against.

The register is why you pad

Trace the uncertainty back and it starts where every cafe's does: your register shows you what sold, not what you used.

That gap is the whole problem. Square recorded "iced matcha latte, oat, $6.75." What left your shelf was a scoop of matcha, a pour of oat milk, a cup, a lid, and a straw. Square's own inventory "only tracks finished goods" (Square product page, July 2026), and its Recipes beta cannot apply recipes to modifiers at all (Square documentation, July 2026). So when a customer swaps oat for whole, the register total does not move, but your oat number should.

Without that translation, you do not actually know your usage. You know your sales, and you know roughly how sales turn into shelf, and "roughly" is the exact word the pad exists to cover. The blurrier your read on what you used, the bigger the pad you add to feel safe. Sharpen the read and the reason to pad starts to shrink.

Only one kind of over-order shows on the P&L

Now the part that makes over-ordering so easy to ignore: most of it is invisible.

Split an over-order into two kinds.

The first is perishable. You bought more milk, cream, or pastries than you sold before they turned, and the difference went in the trash. That one you can see, at least if you log it. It shows up as waste, and waste eventually shows up as a food cost number that runs high for reasons the benchmark cannot explain.

The second is dry goods. Cups, lids, sugar, tea, filters. You over-ordered two cases of cups and nothing spoiled, so nothing looks wrong. But you spent that cash, and it is sitting on a shelf instead of in your account, and it will keep sitting there until you sell through it weeks from now. On the P&L, cash tied up in a stockroom looks identical to cash spent well. It never files a complaint.

This is why over-ordering never gets fixed by staring at the P&L. The spoiled milk shows; the six months of cups do not. Both are the pad. Only one of them ever raises its hand.

Size the pad to the cost, item by item

So you do not kill the pad. A pad is safety stock, and safety stock is doing a real job. You size it, item by item, to what getting it wrong actually costs.

Two questions decide the size for any item.

How bad is running out? For a long-lead item you cannot re-order fast, running short is a real wound. Matcha that takes five to seven business days to arrive means a stockout is a week without your signature drink. Pad those generously. For anything the truck brings tomorrow, running out is a phone call, so the pad can stay thin.

How bad is holding it? A case of cups keeps for months and stores flat, so a fat pad costs you a little cash and a little shelf. A case of milk keeps for days, so the same fat pad ends in the trash. Pad the cups, keep the milk tight.

Run those two questions across your items and the answer stops being one nervous "round everything up." It turns deliberate: heavy pad on the cheap, shelf-stable, hard-to-reorder things; thin pad on the perishable, cheap-to-reorder things. This is what a par level is for. The par is where you stop, and the safety stock built into it is the pad, sized on purpose instead of on nerves. Compute it, keep your hand on the pen, and do not let a tool set it for you by feel.

Shrink the uncertainty and the pad shrinks itself

The pad is a product of two things: how much running out hurts, and how blind you are to what you actually used. You cannot change the first much. A stockout on a Saturday will always hurt. The second is where all the room is.

Everything below narrows the same uncertainty:

• Map every sale to its recipe, modifiers and all, so "roughly three cartons a day" becomes an ounce number that reconciles against your count. When the ordering playbook says to stop ordering a little extra, this is the missing half of how. You stop padding because you finally know the number you were padding around.
• Know your delivery cadence cold, so the pad only ever has to cover the real gap to the next truck, not a vague "few days." Build the schedule once and it holds.
• Count on a rhythm and let sales-based depletion fill the days between, so your read on the shelf is never more than a couple of days stale.

Do that and the pad does not vanish, it right-sizes. You still carry safety stock on the matcha and the cups. You stop carrying three phantom cases of oat milk you bought because you could not see straight.

Pick your single worst offender this week. It is usually a perishable you keep throwing out or a dry good you keep tripping over in the back. For that one item, write down what you actually used over the last two weeks, from your recipes and sales, not from memory. Set the pad to the cost of getting it wrong. Order to that. One item, and the standing charge starts coming down.