Industries

Purchasing: the order day decides the margin, not the menu

A restaurant is judged at the rush and ruined on Tuesday morning, on the phone, in three minutes.

Published on 19 May 2026 — Algeria Agency

The article beside this one is built around the peak: the rush, Friday evening, Ramadan week. It is right — this trade is judged on its worst twenty minutes.

This article is about the week that turns behind them. It is regular, it repeats, and it contains two days on which most of what will be left at the end of the month is decided.

Those two days are the order day and the delivery day. Neither is visible to the customer, neither appears in reviews, and neither is treated as a management moment in most establishments.

We give no food cost ratio. The reason is in section 10 and it is not the usual caution: that ratio moves when you change your prices, so it varies without a single purchasing decision having changed.

The margin is decided on the order day

A dish has a selling price set once and a cost that is reset every week. The first is visible on the menu and discussed at length; the second is decided on the phone, in three minutes, by somebody with other things to do.

It is a remarkable imbalance of attention. A restaurateur will spend three evenings choosing the typeface for a menu and thirty seconds deciding whether to take the meat at this morning’s price.

The reason is not carelessness, it is structural: a selling price is a decision and a purchase looks like an execution. You think you are executing an order when you are taking a margin decision.

The useful reversal is to treat the order day as the week’s management appointment. That is not a metaphor: it means a fixed hour, a prepared list, and somebody with the figures in front of them.

Everything else in this article follows from that shift. Once ordering is an appointment rather than a reflex, delivery, counting and the second supplier follow naturally.

The purchasing week: four positions

A restaurant’s purchasing week turns through four positions and always returns to the same point. Count what is left, order, receive, check what was consumed.

The first is the one almost nobody keeps, and its absence explains most of the problems in the other three. Ordering without knowing what is left produces either a shortage mid-service or a surplus that will become a loss.

The second is a short moment and it has to be prepared. An order placed from memory between two services systematically contains too much of what is in mind and not enough of what was forgotten.

The third is the only one where you have a power of refusal, and section 3 is devoted entirely to it. The fourth closes the loop: what was consumed against what was sold.

Those four positions take under two hours a week in a medium establishment. It is not their length that makes them hard, it is that they have no hour — and in this trade a task with no hour is a task the service eats.

The order day has to be fixed

The order goes out on the same day, at the same hour, every week. The constraint looks bureaucratic and it produces four effects, none of them obvious.

The first is that your supplier can organise. A customer whose order arrives at a fixed hour is served before one who calls when they think of it, because the round is already composed when the second telephones.

The second is that the order gets prepared. A fixed day allows counting the day before, which is the only way to have a stock position that is not an impression.

The third is that you can compare. Two orders placed on the same weekday three weeks apart are comparable; two placed at random are not, because the number of days elapsed varies.

The fourth is the most useful under cash pressure: a fixed day makes payment dates predictable, and a predictable date can be negotiated. A supplier accepts an announced deferral far more readily than an observed delay.

Delivery: the only moment you can refuse

The delivery arrives at the worst possible moment, generally mid-preparation, and it is signed for by whoever is free. It is the weak point of the whole chain and it is corrected without spending a dinar.

Three things are checked on delivery and only one takes time. The quantity, the condition, and the price written on the note. The first two are obvious; the third is the one nobody ever looks at.

Delivery is the only moment when refusal is possible and easy. Goods refused in front of the driver go back; the same goods accepted and disputed the next day become a negotiation you will lose half the time.

So somebody has to be designated to receive, with the right to refuse. A commis who signs without daring to refuse damaged tomatoes is not at fault: nobody told them they could.

What a signature commits has to be clear to everybody. Signing a note says the quantity and the condition conform; it is not a courtesy formality, and a note signed without checking has exactly the value of a blank cheque.

The price that moved without anybody telling you

The price of most foodstuffs moves, and it does not move by announcement. It simply appears differently on the next delivery note, and it is almost never read.

The mechanism is not generally dishonest: the supplier absorbs a movement too and passes it on. But it produces a curious situation in which your food cost changes without anybody in your establishment learning about it.

The remedy costs thirty seconds per delivery: the price of three or four key items, taken off the note, compared with last week’s. No spreadsheet, no software — one column in a notebook.

Key items are the ones weighing most in your best-selling dishes, and there are generally four or five. Tracking thirty references does not happen; tracking five happens for years.

That record serves twice. It lets you discuss an increase when it arrives rather than three months later, and it tells you when a dish on your menu has stopped being profitable — information the till never gives.

What is counted on delivery, and never afterwards

A quantity is verified when it arrives and at no other moment. That is obvious and it is nevertheless got round every day, because deliveries arrive when nobody has time.

The method that holds is a reduction: you do not count everything, you count what is expensive and what counts quickly. Proteins, high-value goods, and anything sold by weight.

Low-value items are counted by case rather than by unit, which takes three seconds and catches the great majority of errors — a missing case is a frequent error, a missing unit inside a case much less so.

Actual weight deserves particular attention on goods sold by the kilo. That is where the most expensive and least visible discrepancies live, and a scale at goods-in generally pays for itself within weeks.

Finally, any discrepancy is written on the note before signing, not in a message sent that evening. An annotated note signed by both sides is the only document nobody disputes.

A restaurant’s stock: three zones, three rhythms

A restaurant has three stocks with nothing in common, and they are nevertheless treated as one.

Fresh is counted daily because it perishes in days, and it does not store: what is over-bought is already lost, whatever price was obtained. It is the one stock where a quantity discount is almost always a bad deal.

Dry goods and groceries are counted weekly, store without risk, and are the right place to take advantage of a price. They are also the only one of the three where buying ahead makes sense.

Drinks form a third stock, and it has a property the other two lack: it counts exactly, it does not transform, and any gap between what came in and what was sold is a discrepancy in the strict sense. That is why any check always starts there.

Treating the three zones at one rhythm systematically produces the same result: too much fresh, not enough dry, and no idea what is happening with the drinks.

Loss: four causes, and one is fixed at purchase

What is bought and not sold is lost in four ways, and confusing them leads to correcting the wrong one.

Spoilage comes from buying too much or too early, and it is the only one of the four settled at the moment of ordering. It is also the easiest to measure, because it leaves a physical trace.

Over-production comes from the kitchen: too much prepared for a service that did not happen. It is corrected by service forecasting rather than by purchasing, and it is seasonal and predictable.

Uncontrolled portioning is the most expensive and the most invisible: ten grams more per plate, five hundred plates a week. It is visible neither in the kitchen nor at the till and appears only in the gap between stock consumed and sales.

The fourth is theft, and it should be named without being made the main subject. It is real, it is a minority in most establishments, and it is the first cause suspected — which wastes time on the other three, which together are much larger.

The single supplier and the second one

Most restaurants work with one supplier per product family, and it is a reasonable choice with a precise cost.

The cost is twofold: you have no point of comparison on prices, and you have no answer on the day that supplier does not deliver. The second problem is worse than the first and it always happens on a Friday.

The second supplier does not have to be an equal competitor. A secondary supplier receiving ten to twenty per cent of your orders stays reachable, knows your establishment, and will agree to help out — which an unknown supplier will not.

It also provides the only price reference you have. Two delivery notes from the same week, for the same goods, are a negotiating instrument nothing replaces.

The trade-off has to be accepted: two suppliers means two deliveries, two invoices, two relationships. That is a real administrative cost, and it is why a second supplier on three families beats one on twelve.

What changes when the menu changes

A menu change is treated as a culinary and commercial decision. It is also a purchasing decision, and that half is almost always taken afterwards.

A dish removed frees references that serve nothing else. If nobody notices, they keep being ordered for weeks out of habit and end as loss.

A dish added creates new references, often in small quantities, often at a different supplier. It is the moment a menu gains variety and purchasing gains complexity, and the second effect is never quantified.

The useful rule is to list the references before changing the menu rather than after: what disappears, what appears, and what is shared between several dishes. The third column is the deciding one, because a reference used by three dishes costs far less than one used by a single dish.

It is also the criterion that should separate two candidate dishes otherwise equal. The one reusing what you already stock is structurally more profitable than the one opening a purchasing line, and that gap does not appear on a recipe card.

What is counted, and why the ratio moves without you

Three things are counted weekly, in the same notebook as the orders. The amount purchased, the amount sold, and the gap between theoretical and counted stock on drinks.

The third is the most reliable of all, for the reason given in section 7: a drink does not transform. A gap there is a gap, and it points at a cause that can be looked for — whereas a gap in the kitchen can come from four places.

The figure we will not give is a benchmark food cost ratio, and the reason is not that it varies between establishments. It is that it varies without anything having changed in your purchasing.

That ratio is a relation between what you buy and what you sell. It falls if you raise your prices, it rises if you run a promotion, and it moves with every change in sales mix — a month selling more drinks and less meat lowers it without a single purchasing decision having been taken.

In other words it mixes a purchasing signal with a pricing signal in one number, and it is used to judge the first. Follow instead the price of your five key references and the gap on drinks: those two move only for reasons that concern you.

What we do, and what we refuse to do

What we do is a notebook and two appointments: the tracking column for the five key references, the goods-in sheet with its three checks, and the fixed order day and hour, written with a name against them.

We do not negotiate with your suppliers and we will not introduce you to one. A supplier relationship in this trade is built over years and over favours, and an intermediary adds nothing to it but another intermediary.

We will not sell stock management software to a forty-cover establishment. A notebook, a scale at goods-in and five tracked prices do most of the work, and a tool requiring data entry after service will not be filled in.

We will give no target food cost ratio, for the reason written in the previous section: that number moves when you change your prices or your sales mix, and a target set on it is met by raising the menu rather than by buying better.

Finally, the two most profitable actions are free and can be taken this week: fix the day and hour of the order, and record the price of five references on every delivery note. Thirty seconds per delivery, and you will know before anybody else when a dish has stopped being profitable.

Frequently asked questions

Why a fixed order day?

For four reasons, none of them obvious. Your supplier can organise and serves you before whoever calls when they think of it; you can count the day before, which is the only way to have a real stock position; two orders on the same weekday three weeks apart are comparable; and payment dates become predictable, therefore negotiable — a supplier accepts an announced deferral far more readily than an observed delay.

What should be checked on delivery?

The quantity, the condition and the price on the note — the third is the one nobody ever looks at. And somebody has to be designated to receive with the right to refuse: goods refused in front of the driver go back, the same goods accepted and disputed the next day become a negotiation you will lose half the time.

How do we track prices without software?

One column in a notebook and five references. Choose the ones weighing most in your best sellers, record their price from every delivery note, compare with last week. Thirty seconds. Tracking thirty references does not happen; tracking five happens for years.

Should we buy in quantity for a better price?

On dry goods and groceries, yes: they store without risk and are the right place to take advantage of a price. On fresh, almost never — what is over-bought is already lost whatever price was obtained, because that stock perishes in days and does not store.

Do we need a second supplier?

On two or three families, yes. It does not have to be an equal competitor: ten to twenty per cent of your orders is enough for them to stay reachable and to help out on a Friday, which an unknown supplier will not. And two delivery notes from the same week are a negotiating instrument nothing replaces.

What food cost ratio should we aim for?

None, and not because it varies between establishments: because it varies without anything having changed in your purchasing. It falls if you raise prices, rises during a promotion, and moves with every change in sales mix. It mixes a purchasing signal with a pricing signal and is used to judge the first. Follow your five key prices and the gap on drinks.

Where we come in

A supplier who knows which day you ring handles your order differently from one who discovers you afresh each week.

We open tracking on the five references that weigh most on margin.
We prepare the goods-in sheet whoever unloads can actually fill in.
We leave the negotiating to you: it rests on a relationship, not a table.

Forty covers justify no stock software: a notebook kept daily beats a system nobody updates.

See the Restaurant & Hospitality page Start with the free audit

Let us talk about your project

A free audit, no commitment: we look at your online presence and tell you what is holding it back.

Request the free audit Write to us