Gross Profit Per Dish: Why Percentage Is Not the Whole Story

I once ran a menu where the accountant was thrilled and the bank account wasn't. Every dish was tidy on percentage. The seasonal salads sat at a beautiful 22% food cost. And yet at the end of a busy month there was less cash in the till than there should have been. It took me an embarrassingly long time to see why: a 22% salad that sells for $9 puts $7 in the drawer. A 38% braise that sells for $28 puts $17 in the drawer. The salad wins on the metric everyone talks about and loses on the only one that clears the wage bill.

That's the whole point of this piece. Food cost percentage is a ratio — useful, but it hides the size of the prize. What you take to the bank is gross profit per dish in hard currency, multiplied by how many of that dish you actually sell. Chase the percentage in isolation and you will, sooner or later, optimise your way to a leaner-looking, worse-earning menu.

Contribution margin is dollars, not a ratio

Strip away the jargon and gross profit per dish is one subtraction:

Menu accountants call this the contribution margin: the cash each sale contributes toward covering everything that isn't food — rent, wages, gas, the card-machine fees, your own wages if you're lucky. It's expressed in money, not percent, and that distinction is the entire argument.

Take two mains. The pasta costs you $3.10 to plate and sells for $16. The lamb rump costs $9.40 and sells for $29. Run the percentages and the pasta looks like the star: 19% food cost against the lamb's 32%. Run the cash and the picture flips.

The lamb has the "worse" food cost and puts nearly seven dollars more in the drawer every time it leaves the pass. If a customer is deciding between the two and you've trained your floor to nudge them toward the pasta because it "costs less," you've just talked them out of $6.70 of gross profit. The percentage lied by omission.

Why chasing the lowest food cost % leaves money on the table

The low-percentage trap is seductive because the metric is easy to compute and easy to brag about. "We run a 26% kitchen" sounds like control. But percentage rewards cheapness, and cheapness has a ceiling. There's a floor under what you can charge for a bowl of pasta before it stops feeling like a restaurant meal. There's much more headroom on a well-executed protein dish that a guest reads as worth $30.

Push the logic to its conclusion and the lowest-percentage menu in town is a bowl of rice: pennies to cook, sell it for $6, gorgeous food cost. Nobody runs that restaurant because it can't pay a chef. The dishes that carry a venue are almost always the ones with a slightly frightening food cost percentage and a fat cash margin — the steak, the seafood platter, the sharing roast. Kill them because a spreadsheet flagged them red and you gut your own earnings.

This isn't an argument to ignore percentage. A dish creeping past 40% with no cash reward behind it is genuinely a problem — that's a plate that's both inefficient and thin. The argument is that percentage is a screen, not a verdict. When a dish flags high, you ask the second question: how much cash does it make, and how often does it sell? Only then do you decide anything.

Weight it by units sold, or you're guessing

Per-dish gross profit tells you what one plate earns. It doesn't tell you what the dish earns your business, because dishes don't sell in equal numbers. The number that matters is total contribution: gross profit per dish multiplied by units sold over a real period.

This is where menus surprise their owners. A dish with a middling margin that sells fifteen times a night can out-earn a showpiece with a huge margin that sells twice. Here's a real-shaped week from a bistro section — five mains, a month of covers:

The market fish has the best margin per plate and the second-worst total contribution, because it barely moves. The roast chicken — unglamorous, middle-of-the-pack margin — is quietly the most important dish on the section, carrying $5,670 a month. If you'd ranked this menu on food cost percentage alone, or even on per-plate margin alone, you'd have missed the chicken entirely. Multiply by volume and the truth arrives.

You cannot do this without sales counts. A POS export by item is ideal; a clipboard tally kept for one honest week will do if you have nothing else. The moment you pair margin with units, the menu stops being a list of dishes and becomes a ranked list of earners.

This is menu engineering, in plain clothes

If total contribution feels familiar, it's because it's half of classic menu engineering — the framework that sorts every dish into four boxes on two axes: how profitable it is (per-plate gross profit) and how popular it is (units sold). The names are a bit corny but they stick:

• Stars — high margin, high popularity. Protect them. Never fiddle with a working star.
• Plowhorses — low margin, high popularity. The roast chicken. Sells like mad, doesn't earn enough per plate. Your biggest opportunity, because a small margin gain multiplied by big volume is real money.
• Puzzles — high margin, low popularity. The market fish. Earns beautifully but nobody orders it. Fix the visibility, not the dish.
• Dogs — low margin, low popularity. The halloumi salad here. Neither earns nor sells. Candidate for the chop.

Menu engineering is just gross profit per dish and units sold, plotted against each other so the eye can find the levers. The reason it beats a food-cost report is that it never lets you forget the cash dimension. A percentage report ranks the chicken and the salad as broadly similar "fine" dishes. The four-box view puts the chicken in the position of maximum leverage and the salad on death row, which is exactly right.

Costing the plate honestly comes first

None of this works if the plate cost is wrong, and plate cost is where most kitchens quietly lie to themselves. Gross profit is price minus cost, so an under-stated cost inflates every margin figure you're about to make decisions on. Three things bite hardest:

Sub-recipes and yield

The braise that sells for $28 isn't just a piece of meat. It's a portion of a batch that shrank 30% in the oven, sitting on a purée that had its own butter and cream, finished with a jus reduced from stock you costed as free because you made it from bones. Cost the whole tree — every prep, at the yield you actually get — or the margin is fiction. A dish you think earns $19 might earn $15 once the trimmings and shrinkage are honest.

Today's prices, not opening-day prices

Butter, oil and protein have all moved hard in the last two years, and margins move with them. The National Restaurant Association's industry research has tracked food-cost pressure staying near the top of operators' concerns through this period — which means a cost card built on last year's invoices is overstating your gross profit on exactly the dishes you rely on. Update your five most expensive ingredients every delivery and your margins stay roughly honest.

Portion drift

A "generous hand" is a margin leak with a friendly face. Ten extra grams of protein across four hundred covers is real money gone, and it never shows up in the recipe card — only in the gap between what you think you earn and what the P&L says. Scales and portion cards on the line are the cheapest margin repair there is.

What to promote, reprice or cut

Once you have gross profit per dish and units sold sitting side by side, the moves almost pick themselves. Work the four boxes in order of payoff:

1. Promote your stars and puzzles. Stars earn and sell — put them where the eye lands, top-right of the section, no price crammed against the description. For puzzles like the market fish, the dish is fine; the visibility is the problem. Move it up the menu, have the floor mention it, give it a photo. You're not fixing the food, you're fixing the ordering.
2. Reprice or re-engineer the plowhorses. This is the big one, because volume multiplies every gain. The roast chicken earns $13.50 across 420 covers. Add a dollar of margin — trim the plate cost by tightening the garnish, or lift the price by $1 — and that's $420 a month from one nudge. On a plowhorse, small changes are large because of the volume behind them. Just don't blunt the thing that makes it sell.
3. Fix or cut the dogs. The halloumi salad earns $588 a month and takes up a line and a mise-en-place slot. Give it one honest attempt — a reprice, a reposition — and if it doesn't move, cut it. A dog isn't just weak; it costs you the menu space and prep a better dish could use.

The discipline is to make decisions on cash and volume together, never on percentage alone. Every time you're tempted to cut a "high food cost" dish, pull its gross profit and its sales count first. Half the time the number talks you out of a mistake.

"The dish I nearly cut for its 39% food cost turned out to be the single biggest earner on the board. I'd been reading the wrong column for three years."

That's not a hypothetical — it's the roast chicken, more or less, in three different kitchens I've worked. Percentage is a genuinely useful number for spotting inefficiency and for pricing a plate against a target. But it is one column on a wider report. The columns that decide whether you can pay your team are the cash gross profit on each plate and the number of times that plate leaves the pass. Read those together and the menu tells you exactly where the money is.