Restaurant Cost Control: A Weekly Routine for Café Owners
22 August 2026 · MidaOne
Cost control in most cafés happens twice a year: once when the bank balance gives somebody a fright, and once when the accountant asks a question nobody can answer. Both times it eats a weekend, produces three decisions, and then stops. Meanwhile the costs that actually move — a supplier's unit price, a new barista's pour, a rota that grew by an hour here and an hour there — move every single week. A short routine you genuinely repeat beats a deep review you do twice a year, and it takes less time than one afternoon of panic.
Why the twice-a-year review never catches anything
A cost problem in a café is almost never one big event. It is 40 fils on a litre of milk, a jug steamed to the brim for a single flat white, an extra shift added for a busy week that nobody removed afterwards, and a menu item whose ingredient cost climbed while its price sat still. Each one is too small to notice. Together they can take a couple of points off your margin in a quarter with no moment where anything visibly went wrong.
By the time a six-month review finds them, you have already paid for six months. The point of doing this weekly is not thoroughness — a weekly check is much shallower than an annual one. The point is that a change you find in week one has cost you a week.
The Monday hour
Pick a morning, keep it, and do the same five things in the same order. Most café owners land somewhere around an hour once the routine settles, and the first two or three weeks are always slower than the ones after.
| Check | What you're looking for | Roughly |
|---|---|---|
| Short stock count | A variance between what sold and what left the shelf | 20 min |
| Food cost for the week | Movement of more than a point or two from your own baseline | 5 min |
| Labour as a share of sales | Hours that drifted away from when trade actually happens | 10 min |
| Top and bottom ten items | What to push, what to reprice, what to remove | 10 min |
| This week's delivery notes | Unit prices that changed without an announcement | 15 min |
Nothing in that list is difficult. What makes it work is that it happens on a schedule rather than in response to bad news, so you build a run of figures you can compare against each other. One week's number tells you almost nothing. Six weeks of the same number is a picture of your café.
Start with a short stock count, not a full one
A full stocktake is why most cafés count once a quarter. Don't do that. Pick five to eight lines that are high-value, high-volume or both — milk, house beans, the syrup that goes in everything, your best-selling pastry — and count only those, at the same point in the week, before any delivery lands.
Those few lines will carry the large majority of whatever you are losing. Compare what you counted against what your sales say should have gone, and the difference is the number worth watching. Our guide to where café stock actually disappears works through what the gap usually means — and it is worth saying plainly that in most cafés it is over-pouring and unrecorded staff drinks long before it is anything worse.
Two ratios, and the baseline underneath both
Food cost is the share of sales that leaves again as ingredients: opening stock plus purchases minus closing stock, divided by net sales. Use VAT-exclusive sales — UAE menu prices are shown VAT-inclusive, so the 5% sitting inside the number on your board was never yours. The full method is in our guide to food cost percentage.
Labour is the same shape: wages for the period as a share of sales for the same period. Both ratios come with published benchmarks, and both benchmarks are close to useless for a specific café. A specialty shop buying better beans and a high-volume takeaway counter can be equally healthy at very different food costs. A café on a mall unit with long mandated trading hours will carry a labour ratio a neighbourhood shop never would.
So don't manage against a benchmark. Calculate both for four consecutive weeks, take that as your own baseline, and then treat any movement of more than a point or two as a question rather than a verdict. "Why is labour two points up on a week with the same takings?" is a question you can actually answer — usually with the rota, which our guide to staff scheduling covers in detail.
Read the extremes of your menu, not the middle
Every week, pull your ten best-selling items and your ten worst. The middle of the list rarely tells you anything; the two ends usually do.
- A bestseller with a thin margin is the most expensive item on your menu, because you are selling a lot of it. It is also the safest thing to reprice by a dirham, since people who buy it every day have already decided they want it.
- A slow seller with a good margin is a merchandising problem, not a menu problem — it needs a better position on the board or a mention at the till before it needs removing.
- A slow seller with a thin margin that also ties up a stock line nothing else uses is usually the item to cut. The saving is rarely the item itself; it's the ingredient that stops dying in the fridge.
- Anything that vanished this week compared to last is worth ten seconds of thought. Sometimes it's the weather. Sometimes somebody moved the board.
This is the weekly, cheap version of the full exercise in our guide to menu engineering. You don't need the full grid every Monday. You need to notice the week your second-best seller quietly stopped being that.
Finish on the delivery notes
Supplier prices almost never arrive as an announcement. They arrive as a slightly different number on one line of a long invoice, and they compound. Keep a running unit price for your top ten items — a single sheet is enough — and check this week's notes against it before you file them. Check what was delivered against what was invoiced while you're there, because the doorstep is the only moment you have any leverage.
One habit worth adding: when a price moves, decide that week whether it changes a menu price. Cafés absorb increases by default, quietly, one line at a time, and then wonder in six months where the margin went. The five POS reports worth reading weekly will show you the sales side of that same decision.
Where MidaOne fits
The reason this routine doesn't survive in most cafés isn't the arithmetic — it's the assembly. Sales sit in the till, purchases sit in a folder of delivery notes, and stock sits in a spreadsheet somebody updates when they remember, so producing one honest figure costs an evening rather than an hour. MidaOne keeps the till, stock, purchases and accounting in one record: stock moves as you sell, ingredients come off by recipe, and the sales figures come out VAT-exclusive from the same place, alongside a summary that already carries cost of goods, gross margin and waste. The routine above stops being a reconstruction and becomes a read. The café POS guide shows how the pieces connect.
Spend the Monday hour reading numbers instead of building them. Free for 14 days, no card.
Start your free trialFrequently asked questions
How often should a café review its costs?
Weekly for the short version and monthly for anything deeper. A weekly check is shallow by design — a few stock lines, two ratios and the ends of your menu — but it is short enough to connect a change to something that actually happened that week. A quarterly review finds the same problems after you have already paid for them.
What should I check first if my margins are slipping?
Count a small number of high-value stock lines and compare them against what your sales say you should have used. Stock variance is the fastest thing to measure and the most common cause in a small café, and it is usually a portioning or recording issue rather than anything deliberate. Supplier price creep is the second place to look.
What is a good food cost percentage for a UAE café?
The industry rule of thumb is a wide band that suits almost nobody exactly, and it is far more useful as a sanity check than a target. Work out your own figure for four consecutive weeks, treat that as your baseline, and manage against movement away from it. Your last quarter is a better benchmark than anyone else's average.
Do I need to count stock every week?
Not all of it. Counting five to eight high-value, high-volume lines weekly gives you a usable signal within a month, while a full stocktake so demanding that it happens quarterly gives you almost nothing you can act on. Count the short list often and the full list occasionally.
Should I cut costs or raise prices?
Usually a bit of both, and in that order — a cost problem you fix once keeps paying, while a price rise you make to cover an uncontrolled cost has to be repeated. That said, absorbing every supplier increase by default is a decision too, and it is the one most cafés make without noticing.
The owners who control costs well aren't the ones with the strictest rules. They are the ones who look at the same five things often enough that a bad week is still a week. Put an hour in the diary for next Monday, do the list once badly, and do it again the Monday after. By the fourth one you'll be reading a trend instead of a number.