Which dishes to remove from the menu to gain profitability: the high food cost myth against what actually drains the till

Which dishes should you cut from the menu to gain profitability?

Out go the dishes whose dollar contribution margin, multiplied by monthly turns, fails to reach 1 % of what your top seller delivers, whatever their food cost says.

The arithmetic is short and it unsettles almost everyone: a 34 % dish leaving 9.80 USD of margin over 180 monthly sales contributes 1,764 USD, while a 21 % dish leaving 6.40 USD over twelve turns contributes 76.80 USD. Twenty-three to one, in favour of the plate everybody calls expensive. Sector reference sits from 28 % to 35 % of price, according to the National Restaurant Association Restaurant Operations Report 2025, and a range like that portrays an entire category. It will never tell you which line to delete from your own menu on Monday morning.

Magnitude rules, percentage tags along

Nobody takes a percentage to the bank. Rent, payroll, utilities and the loan instalment get covered with dollars of contribution margin, while food cost stays where it does work: supplier negotiation and portioning drift. Put it in numbers, which is where the argument usually ends. A 14 USD dish at 22 % puts in 3.08 USD of raw material and gives back 10.92; a 46 USD dish at 33 % puts in 15.18 and gives back 30.82. The textbook's bad plate leaves nearly triple the money. When a venue billing 80,000 USD a month in food slides from 68 % to 64 % average margin after pulling its expensive dishes, 3,200 USD evaporate every month and fixed costs sit exactly where they always were.

What each menu price range includes?

Three price ranges, three different jobs: that is how a menu that works is put together in September 2026.

At the bottom, 9 to 16 USD, raw material cost runs 24 % to 30 % and the mission is filling weekday tables, with 7 to 11 USD of unit margin that only justify themselves on high turns. The backbone of the till is the middle range: 17 to 29 USD, cost running 28 % to 34 %, margins of 12 to 20 USD per plate. And up top, 30 to 55 USD, you absorb 33 % to 38 % cost, above the theoretical ceiling, and still keep 20 to 34 USD per unit. That top tier is the first thing trimmed out of fear of the percentage, and it is precisely what holds the average check up.

The four factors that move price and margin

Volatile raw material heads the list of what moves a price: eggs got so expensive in 2025 that Waffle House charged a 0.50 USD surcharge per egg, according to NPR, and a shock like that shifts a dish's cost by 8 to 15 points. Execution complexity weighs next, with its inventory SKUs, its perishable mise en place and its training minutes. Buying power is the third lever: change your order scale and 3 to 7 points move with it. Closing the list comes local market pricing, where 93 % of quick-service restaurants raised tariffs during 2024, per the Oysterlink compilation. Four levers, quantifiable in an afternoon, and ignore a single one and your menu falls out of tune within a quarter.

The case of fourteen pastas at 21 % food cost

The ceviche was not the problem. It cost 33 %, everyone had warned him that anything over 30 % bleeds money, and with that idea lodged in his head the owner opened the August 2026 report on his 64-item menu in Bogotá. That plate carried 11,400 USD a year into margin, ranked second in units, and was the reason guests crossed town to eat there. The real hole sat three sections further down, in the pasta: fourteen references with 21 % raw material cost that together never passed 38 monthly units and forced six exclusive inputs into the walk-in. We pulled twelve of them. Cooler space came back, two daily hours of mise en place disappeared, and the sales line never noticed.

Turnover weighs everything else

Eight sales a month with a splendid margin still amount to a museum piece: it takes a printed line, cooler space and room in the head of the cook who must remember it. The Masterestaurant menu audit, the way Diego F. Parra teaches it, fits in three moves: multiply unit margin by units sold, sort high to low, find where 80 % of total contribution piles up. On menus of 50 to 70 dishes that 80 % comes out of 14 to 22 references. Below the 20th percentile everything is a candidate to go, with one exception I defend without apology: the dish that makes no money but brings the group, the decent vegetarian or the kids' plate, stays, because its value sits in the whole table. Only 10 % of restaurants do quality menu engineering, according to Oracle NetSuite; 60 % never do it.

What happens if you halve the menu?

Suppose you go from 64 dishes to 32 tomorrow. The first effect appears not in food cost but in inventory: active references drop 25 % to 40 %, and out goes the product that used to expire before reaching any plate.

In the kitchen you notice it a week later, when training a new cook falls from three weeks to ten days and the Saturday peak runs faster because fewer stations fight over the pass. Then the third effect, the one that frightens owners and hardly ever arrives: lost sales. If the cut followed contribution rather than percentage, guests migrate toward what stayed, which is what they were already ordering. The risk is not cutting too deep, it is cutting the wrong dish after reading the wrong column.

How to negotiate and optimize before deleting?

A repaired dish beats a dead one, so before you delete the line, work this order. Renegotiating raw material with consolidated volume usually moves cost 4 to 8 points, because folding three similar inputs into one changes the conversation with your supplier.

Second move: reset the portion to what the guest actually finishes, and recovering 25 grams of protein on a 28 USD plate hands back roughly 1.60 USD. Then shift that dish into the top third of its section, where the eyes land and the turns rise. Price comes last: 1.50 USD more, four weeks of measuring, with the precedent that 42 % of restaurants already pushed increases through in 2023 according to Toast and guests absorbed them. If contribution still sits at the bottom after that cycle, delete it on Monday.

Four differences that decide the menu

Magnitude first, percentage second. Contribution margin —selling price minus raw material cost per portion— is the only number that reaches the bank account, and everything else is blackboard accounting. Percentage belongs in front of a supplier and in a portioning audit, two legitimate uses with nothing to do with deciding which line gets deleted. Rank your menu by dollar margin, rank it again by food cost, and the two lists will barely resemble each other: the second one pushes you straight at the dishes holding up your check. Turnover weights everything else, which is why our threshold is a 70 % share of the category's average sales: under that line an item gets 60 days to climb or it goes.

Four differences that decide the menu — in practice

A splendid margin showing up eight times a month is not profitable, it is an expensive ornament taking a printed line and a cook's memory. What almost nobody anticipates is what follows the removal: sales do not evaporate, they migrate. Between 55 and 75 % of that demand shifts to the neighbouring reference, and the neighbour usually carries a better margin. Complexity carries a price even when no recipe prints it. Do the arithmetic dish by dish: how many inventory references exist ONLY because of it, how many mise en place minutes it eats in each service, how much waste it leaves on the days nobody orders it.

Four differences that decide the menu — key points

In a twelve-person kitchen with 75 % annual turnover, teaching one dish costs around 40 minutes per new cook; at nine hires a year and 9 USD an hour of loaded labour, that is 54 USD annually in teaching alone. Multiply it by the twenty nobody remembers ordering and you have spent a fortnight's wage. Removal is the last resort, never the first. Of ten items the analysis flags red, three genuinely leave, four get repriced, two get reformulated around a neighbouring dish's inputs —that is how the exclusive reference dies— and one comes back to life simply by moving on the printed page. According to Gregg Rapp, the menu engineering consultant US industry press quotes regularly, design and placement measurably shift what a guest orders, so redesigning runs cheaper than removing. Cutting first and thinking afterwards is what sinks average checks.