Per-dish profitability: the mistakes that erase it and the method that recovers it

What actually measures a dish's profitability?

A dish's profitability is measured by absolute contribution margin per order multiplied by monthly turnover, never by food cost percentage. Look at that 140-cover restaurant in Mexico City:

for example, a dish with a higher cost percentage that clears eleven dollars and sells ninety times a month puts far more in the till than another with a lower percentage, five dollars of margin and forty sales. The gap runs several times in favor of the dish any percentage dashboard would flag red. And when labor eats a very large share of revenue, the plate that looked healthy on the costing sheet stops looking healthy on the P&L. Those dollars of difference per dish are exactly what separates a healthy net profit from a thin one. Percentage is a RATIO; the till holds money.

Your signature entrée won't absorb the increase your appetizers will

Raise prices by family, never across the board: elasticity is not distributed evenly across menu sections. An increase on appetizers gets absorbed without friction, while the same increase on the signature entrée —which anchors how guests read every price in the room— pushes diners toward lower absolute-margin options and can lift average check while leaving LESS cash than before. Evidence on price aversion is blunt: many diners call dynamic pricing in restaurants price gouging, and a good share say they would order less often. Translate that to your menu: if you touch the anchor, manage the mix too, because a menu can collect more per cover and still lose total profitability.

A cluster of dishes carrying a large share of the orders

Many of that menu's dishes sat below the average margin while concentrating a large share of all orders: that pattern shows up almost every time costing is done per portion, with standardized recipes and measured waste. The menu isn't badly costed, it's badly READ. Kasavana and Smith's menu engineering, published in 1982 and still the best tool available, plots popularity against contribution margin and sorts every dish into four quadrants; hardly anyone runs it on their own data, because it demands weekly inventory, and without inventory costing is a guess dressed up as a number. I got this wrong for years by recommending you cut the bottom quadrant: when a weak-margin dish carries a large share of volume, cutting it sinks traffic before it fixes the till. You redesign the recipe, you don't delete the line.

Beverages fix what the kitchen cannot

Bar margin rewrites the arithmetic of any kitchen menu, and that is the lever most owners leave untouched. Wine typically runs between 35% and 45% of its selling price according to BackBar's industry guide, while bottled and draft beer sit well below that range. An entrée at a high cost percentage paired with a draft beer at a low one produces a combined margin per cover that no portion adjustment achieves on its own. Dessert follows the same logic: a good share of U.S. consumers have dessert on any given day, and it typically costs a small fraction of its menu price. If your beverage attachment rate sits below 60%, you have margin points asleep on the table.

How to read these numbers in YOUR operation?

Scale the thresholds to your size before touching the menu. Small operation, under 60 covers a day and a menu of 20 to 30 dishes:

work out absolute margin for your ten best sellers by hand, on a spreadsheet, and leave it there; at that volume, reworking two recipes beats installing a system. Mid-size operation, 100 to 180 covers and 40 to 70 dishes like the Mexico City case: you need weekly inventory and a point of sale that breaks out item-level sales, because a large share of orders sitting at low margin is invisible without data, and there the gap in net profit is real money. Group of three or more locations: measure by location BEFORE consolidating, because sales mix shifts by neighborhood and a central decision built on blended averages destroys margin in the site that was doing fine.

The scenario almost nobody runs before cutting a dish

Suppose you cut all nineteen weak dishes at once, persuaded by the food cost dashboard. That 28% of orders doesn't vanish: part of it migrates to high-margin dishes and proves you right, another part —the guests who came for that specific dish— simply stops coming, and if only 30% of those orders are lost, you give up roughly 8.4% of total volume. With labor fixed between 25% and 35% of revenue per the U.S. Bureau of Labor Statistics, that drop frees not one cent of payroll in the first quarter, so margin improves as a percentage while net profit falls in dollars. Diego F. Parra puts it this way in Masterestaurant menu reviews: redesign the recipe and adjust the portion first, measure two full cycles after that, and only then cut whatever still can't defend itself.

Which new items go on, and at what margin?

When it's time to add dishes, pick the families where demand is growing and input cost stays controllable. Datassential recorded in 2025 that 47% of U.S.

consumers ate globally influenced food in the past week, and plant-based menu penetration has kept climbing across all operators since 2012. Plant bases usually land at a low cost and hold price on specialty perception, which is the exact opposite of seafood, whose purchase volatility punishes anyone without a locked contract. Add one uncomfortable fact: a meaningful share of consumers avoid major allergens, so a menu without a clear alternative loses whole tables, not stray covers.

Where these benchmarks come from and how far they go?

The figures in this analysis come from public industry sources, and it is worth saying what they do NOT cover. Toast's 2024 pour costs and the BackBar guides come out of U.S.

operations, with their own purchase prices and tax structure, and so do the labor cost benchmarks from the U.S. Bureau of Labor Statistics aggregates the entire food services category, so a full-service restaurant with a kitchen brigade lives at the top of that band and a fast casual at the bottom. Consumption data from Datassential, Technomic, Circana and FARE are surveys of stated intent and habit, not audited sales, and they always overstate real behavior a little. Use them as a reference line, not a target: the only number that decides your menu is the contribution margin of your own recipes, measured against last week's inventory.

What separates a profitable menu from an expensive one?

The difference lives in the mix, not in the price. A menu can lift its average check 9% and still lose cash if the increase pushes guests toward the lowest absolute margin dishes;

demand elasticity is not evenly distributed across families, and an 8% rise on starters absorbs far better than the same 8% on the signature main, which anchors price perception for the entire room. Absolute contribution margin beats percentage every time. A dish with a high food cost percentage that contributes a solid margin per order and sells often deposits far more cash than a low-cost one that contributes little and sells rarely. Chasing the percentage, any owner would cut the first one. In hard numbers, that is the most expensive well-intentioned decision made in this industry.

What separates a profitable menu from an expensive one — in practice?

The physical menu remains the selling instrument and the QR is the complement, never the replacement. At Masterestaurant we ALWAYS recommend keeping both:

the printed menu controls service pace, menu narrative and suggestive selling, which is exactly where marginal profitability per dish gets built; the QR menu adds price updates without reprinting, accessibility, delivery and consultation analytics. Drop the printed menu and you hand experience control to a phone. Quarterly recosting is not bureaucracy, it is margin defense. With food prices rising year after year and specific ingredients swinging sharply in a single season, a menu costed ten months ago describes a restaurant that no longer exists. Diego F. Parra presses this point with Masterestaurant teams because erosion never announces itself: it arrives dish by dish, without a single bad day to give it away.