Menu Item Profitability: Which Dishes Actually Make Money at a Restaurant Group

The problem: The menu gets designed around what sells, but what sells and what earns are rarely the same dish, and nothing in the business compares them.

The solution: Cost every item on food, prep labor, and waste, then multiply by how often it actually sells, so the menu is built on contribution dollars rather than popularity.

The math

If a four-location group moved even 8 percent of its entrée orders, about 20,000 a year, from a dish contributing $9 to one contributing $16, that is roughly $140k of margin from exactly the same covers.

Every operator knows the best sellers. They are on the specials board, they are what the servers recommend, they are the photograph on the website, and they are the first thing anyone names when asked how the menu is doing. That knowledge is real and it comes free with the point of sale system.

What almost nobody has is the other list: what each dish contributes after everything it costs to put on the plate. Not food cost percentage, which is the number the industry runs on and which is only part of the story. Contribution in dollars, after the protein, after the prep hours it takes before service, after the trim and the spoilage and the plates that come back. That list exists in the business. It has just never been assembled, because assembling it by hand means recipe cards, a scale, a stopwatch, and a week nobody has.

Food cost percentage is the wrong lens

A dish at 24 percent food cost sounds better than a dish at 34 percent, and on a spec sheet it is. On a P&L it might not be.

Percentage is a ratio, and the bank takes dollars. A $34 entrée at 34 percent food cost leaves $22.44 on the plate. A $16 dish at 24 percent leaves $12.16. The lower percentage looks disciplined and contributes ten dollars less every time it goes out the door. Multiply by a few thousand covers a month and the ranking by percentage and the ranking by dollars are describing different restaurants.

Then there is the part percentage does not include at all. Food cost counts what goes on the plate. It does not count the two hours a cook spent breaking down the protein at ten in the morning, the stock that simmered all day, the item that has to be made fresh daily and thrown out at close, or the dish that gets sent back often enough that the comp rate is a real line item. Those are all costs of that specific dish, and they land in labor and waste totals where no individual item can be blamed for them.

The three costs that never reach the item

Three things separate a plate cost from a dish's actual contribution, and none of them are in the recipe card:

Prep labor. A dish that takes twelve minutes of a prep cook's morning per portion carries real money before a guest ever orders it. At a loaded $22 an hour, that is $4.40 gone. Another dish that comes together in three minutes on the line carries almost none. The menu treats them as equivalent.

Waste and yield. The costed recipe assumes the whole case is usable. Trim, spoilage on a slow item, and anything made to par and discarded at close all belong to the dishes that caused them. An item that sells eleven a night and is prepped for twenty is subsidised by the rest of the menu.

Mix. A dish contributing $22 that sells forty a month matters less than a dish contributing $11 that sells nine hundred. Contribution per plate is only half the number. Contribution times volume is the number that pays rent, and it is the one that decides where the item should sit on the page.

Menu design is not random. Items get moved to the top of a section, boxed, photographed, or named as a signature because they sell, and because they sell they sell more. Servers learn to recommend what guests already order, because it is the easy recommendation and it is what the kitchen is fast at. Every one of those decisions is made on a popularity signal, and popularity is the only signal available.

The consequence is a menu optimised for order volume, which is not the goal. The goal is contribution dollars per cover. Those two align sometimes, by luck, and diverge more often than operators expect, because guests choose on appetite and price and nothing about a guest's preference is informed by your prep schedule.

What has to be connected

The pieces are all in the building, in four places that do not talk:

Item sales counts and prices, in the point of sale, per location.
Invoice prices for every ingredient, in the accounting file, changing weekly.
Recipes and portion specs, usually in a binder, a spreadsheet, or the head chef's memory.
Prep times and waste, which are observed by the kitchen and recorded almost nowhere.

Once those sit in one connected picture and stay current on their own, contribution per item stops being an annual project and becomes something you can look at on a Tuesday. That matters more than it sounds, because ingredient prices move constantly. A dish costed in January at 29 percent may be at 38 percent in June and still be printed in the same box on the menu, and the only way anyone finds out today is when the food cost line moves and someone goes looking.

A look at a restaurant group

Take a restaurant group doing about $8 million across four locations, casual full service, a shared head chef, and roughly 260,000 covers a year. Recipes are costed once a year. Food cost runs where it should. The menu was last redesigned around the four items that sell best, which is what everyone advises.

Suppose the group costs every item properly, including prep minutes and waste, and sets that beside the actual sales mix. You would expect the most-ordered entrée to look different than it does now. Say it sells at $26 with a 33 percent plate cost of $8.58, and it takes about twelve minutes of prep labor per portion at a loaded $22 an hour, another $4.40, plus a couple of dollars of trim and end-of-night discard. It contributes around $9 and it is the single most promoted thing on the menu.

Then look at an item further down the page. It sells at $24 with a 26 percent plate cost of $6.24, comes together on the line in four minutes, about $1.47, and generates almost no waste because it shares components with three other dishes. It contributes closer to $16, and it sits below the fold in a section nobody reads.

The difference is seven dollars a plate. If repositioning the menu, retraining the servers on what to recommend, and adjusting the specials moved even 8 percent of entrée orders, roughly 20,000 a year across four locations, from the first dish to the second, that is around $140k of additional contribution on identical covers, identical rent, and identical staff.

The head chef would likely not delete the popular dish, and should not. It brings people in and some items earn their place by drawing traffic rather than by contribution. What would probably change is that the price moves a dollar, the prep gets simplified to cut four minutes, and the menu real estate that was given to it on instinct gets shared with the item that was quietly earning more all along.

How to start

You can do this on one menu section before you do it on all of them.

Cost one section properly. Take the entrées. Add current invoice prices, prep minutes at a loaded rate, and an honest waste allowance to each.
Multiply by the mix. Pull actual unit sales per item for the last quarter and produce total contribution, not contribution per plate.
Rank in dollars and find the split. Put the popularity ranking and the contribution ranking side by side. The items that sit high on one and low on the other are the whole opportunity.
Change the design, not just the price. Move menu position, retrain the recommendation, and adjust the specials before reaching for a price increase.
Let the costing keep itself current. Set automation to recost recipes as invoice prices change and flag any item whose contribution drops below a threshold, so a dish cannot drift for six months unnoticed.

The takeaway

A restaurant group at this size does not usually have a food cost problem, it has a mix problem it cannot see. Percentage costing hides the dollars, prep labor and waste never reach the item that caused them, and the menu gets designed around a popularity signal because that is the only signal the point of sale hands over. The result is a menu that pushes hardest on dishes that may earn least. Cost one section including the prep minutes, multiply by what actually sells, and compare the two rankings. The first honest comparison is likely to change the menu layout, the specials, and what the servers say before anything about the food has to change at all.