Menu engineering: which dishes to push, which to drop (and how to decide with numbers)

Every restaurateur has a dish they're attached to and a dish that "everyone orders anyway". Menu engineering exists to find out whether they're the same dish. The idea is simple: for every item on the menu you look at two numbers — how often it sells and how much it leaves in the till — and from there you decide what to push, what to fix and what to drop. No gut feeling, no "I reckon": numbers you already have in the till and in the kitchen.

First things first: what a dish actually leaves you

A dish's margin is not its menu price. It's selling price minus the cost of the ingredients — the contribution margin, in euros. A plate of hand-rolled pasta at 12 € with 2.80 € of ingredients leaves 9.20 €; the rib-eye at 26 € with 13 € of beef leaves 13. The steak "costs" more than twice as much on the menu, but for your bottom line the two dishes are far closer than they look.

Two warnings before you start counting:

• Think in euros, not just percentages. A 35% food cost on a 30 € dish leaves 19.50 €; a 25% food cost on an 8 € dish leaves 6 €. The "better" percentage is not the dish that pays your rent.
• Weigh the ingredients for real. Recipe costing works with actual gram weights, trim and waste included — not from memory.

To avoid doing it by hand, there's a free food-cost calculator: enter ingredients and weights and you get the plate cost, the food-cost percentage and the margin in euros, no sign-up needed.

The matrix: popularity × margin

Classic menu engineering crosses those two numbers in a four-quadrant matrix. For each dish you ask: does it sell more or less than the average of its category? Does it leave more or less margin than average? Four families come out.

Stars: popular and profitable

Stars sell a lot and leave a lot. The slow-braised beef cheeks that go out twenty times a night with 11 € of margin each: that's a star. What to do: no experiments. Don't change the recipe, don't discount it, give it the best spot on the menu and make sure the kitchen never gets it wrong.

Plough horses: popular, but thin margin

Plough horses sell constantly but leave little. In a gastropub it's the fish and chips; in a pizzeria, the margherita. What to do: work on the margin without killing the popularity — nudge the price slightly (50 cents on a dish that sells forty times a night shows up in the till, not on the guest), rethink the portion or one expensive ingredient nobody notices, or add a richer variant alongside it with a better margin.

Puzzles: profitable, but rarely ordered

Puzzles carry an excellent margin but barely sell. The venison and ale pie at 19 € the kitchen does beautifully and four people order: why? Sometimes the name says nothing, sometimes it's buried at the bottom of the page, sometimes the description doesn't explain what makes it special. What to do: give it visibility — a better position, a description that tells the ingredient's story (how to write one), a word from the server. If a month of genuine visibility doesn't move it, it wasn't a puzzle: it was a dog in disguise.

Dogs: neither orders nor margin

Dogs don't sell and don't earn. They take up space on the menu, lines in the storeroom and attention in the kitchen. What to do: drop them. The menu almost always wins twice — less choice fatigue for the guest, less complexity for the kitchen.

Where to place the dishes that earn the most

Position on the menu is not neutral. Studies on menus indicate that the first items of each section get the most attention and that the eye keeps returning to the start and end of a list. Three practical rules:

• Open every section with a star or a puzzle, not with the cheapest dish. The first line is your shop window.
• Don't bury the margin. The dish that earns you the most doesn't belong halfway down a list of twelve: put it at the top, or highlight it (a box, "chef's choice", house special).
• Fewer items, more weight. Twelve well-chosen dishes get read; thirty get skimmed. Every item you remove gives air to the ones that stay.

And price has a psychology of its own — from the currency symbol to the anchoring effect of a premium dish: we cover it in the menu pricing guide.

When to retire a dish (without regrets)

A dish comes off when, for two or three months running, it sits below its category average both on orders and on margin — and a serious relaunch attempt (position, description, price) hasn't moved it. The real exceptions are few: the signature dish that brings people in even if it earns little, or the dish that serves a specific guest (the vegan option, the children's plate). Everything else is attachment, and attachment doesn't pay your suppliers.

The digital advantage: every change is an instant test

On a printed menu, menu engineering is a yearly exercise: reprinting costs money, so it gets postponed. With a digital menu the cycle shrinks to days. Move the pie to the top of its section, rewrite the description, nudge a price — the change is live instantly, no reprint, and the QR on the table stays the same. If after two weeks the puzzle hasn't taken off, you try another lever or drop it: the experiment costs nothing.

With Menudetto you make those edits in chat or by voice ("move the venison pie to the top of the mains and put the rib-eye at 27 euros") and the public menu updates at once, in every language. Starting from scratch? The digital menu guide shows how to set one up in an evening.

In short

Menu engineering is two numbers per dish — orders and margin — and four boxes: stars to protect, plough horses to make more profitable, puzzles to put in the window, dogs to drop. Start from the food cost (work it out here), place the right dishes in the right spots, and test: with a digital menu every trial goes live instantly. If you want to try it on a real menu, you can create yours free and start with the three dishes you know best.