Menu engineering, properly

Menu engineering is the practice of deciding which dishes deserve their place, and it is one of the few restaurant exercises where a couple of hours with real numbers reliably changes the P&L. It goes wrong for one specific reason: most operators rank dishes by food cost percentage, and that is not the number that pays rent.

Contribution margin, not cost percentage

A dish with an excellent cost percentage on a low selling price can contribute less actual cash than a dish with a mediocre percentage on a high one. Your landlord, your staff and your suppliers are all paid in currency, not in ratios.

So the number to rank on is contribution margin: selling price minus the dish's own food cost, expressed in money. Cost percentage is still useful for spotting drift over time; it is a poor basis for deciding what stays on the menu.

The two axes

Plot every item on contribution margin against how often it sells, split each axis at your menu's own average, and you get four groups. The names vary between textbooks; what they tell you does not.

• ✓High margin, high popularity. Your winners. Protect them: do not move them, do not change them quietly, and make sure they are the easiest thing on the menu to find.
• ✓High margin, low popularity. The biggest opportunity on most menus. These do not need reformulating — they need better placement, a better name, a photograph, or a server mentioning them.
• ✓Low margin, high popularity. Your traffic drivers, and the group to handle most carefully. Small price increases and small cost reductions both work here precisely because volume is high. Removing them is usually a mistake.
• ✓Low margin, low popularity. Candidates to cut. Take one honest look for a reason to keep each — a dietary need nothing else covers, an ingredient that earns its place elsewhere — and then remove the rest.

Doing it without lying to yourself

• ✓Cost recipes properly, including the things people skip: garnish, oil, sauce, waste and the portion actually served rather than the portion written down.
• ✓Use a real trading period, and use the same one for both axes.
• ✓Analyse by daypart if your menu differs across the day — a lunch item and a dinner item are not competing for the same guest.
• ✓Re-run it after changing anything. Menu engineering is a habit, not a project, and margins move as suppliers do.
• ✓Watch what a change does to the items around it. Removing a low performer sometimes moves those guests to a competitor rather than to your winner.

Where the numbers come from

The exercise stalls when assembling the inputs takes longer than the analysis. You need per-item sales counts and per-item food cost from the same period, and in most restaurants those live in two systems, or one of them lives in someone's head.

Opero holds recipe costing on the same spine as the orders, so an item's cost and how often it sold come from one place. Product mix by item is reporting rather than a stock count, which is what makes re-running this quarterly realistic instead of aspirational. Recipe costing and inventory sit on the Growth plan and above.