Which dishes to cut from your menu to gain profitability: before vs after with Masterestaurant

Which dishes go first: the high food cost ones or the ones with thin margin in dollars?

Cut by contribution margin in dollars, never by food cost percentage: the ratio is a purchasing signal, the dollar is what reaches the register.

Run both readings on the same dish and the contradiction shows up on its own. A $28 dish at 34% cost leaves $18.48 of margin per portion; a $9.50 dish at 22% cost leaves $7.41. The first one looks worse on the food cost board and delivers two and a half times more money per sale. Add rotation of 380 covers a month against 60 for the second, and monthly profit lands at $7,022 versus $445, so the ratio rule would have told you to kill the exact dish carrying the month. Dollar margin multiplied by rotation wins, and not by a little: percentage only rules when two dishes sell the same volume, which almost never happens. The rotation that decides is the one in the last ninety days of sales reports, not the one your team remembers, and the gap between them is usually brutal.

Ninety days of real sales mix against the chef's perception

On the 43-item menu that opened this engagement, the kitchen defended four dishes as "the ones that fly"; the system showed two of them barely reached 40 portions a month, under 1.4 a day in a 70-seat room. Behind that sits a tail effect: across most menus I review, roughly 20% of the items carry more than 70% of the portions sold. Perception overvalues whatever gets cooked with pride and underrates whatever gets ordered without thinking. Once we crossed all 43 items against measured rotation, the menu closed at 29 and gross margin climbed 6.2 points in eleven weeks. Register data wins; the team's memory is useful for picking the replacement, not for deciding the cut. A dish with strong margin that barely sells deserves a four-week relocation test before it goes, because very often the problem is the menu and not the dish.

Kill a high-margin, low-rotation dish, or relocate it before killing it

Move it to the top third of its section, rewrite the description around the ingredient nobody else on the block uses, and drop the currency symbol from the price: according to Cornell University, School of Hotel Administration (2009), guests spent 8.15% more per person when menu prices appeared without the dollar sign. Four weeks give you 200 to 400 checks in a mid-volume room, enough of a sample to see whether the mix moves. If the portion count still sits below the threshold you set, it leaves without argument. Relocation wins when margin per portion sits in the top third of the menu; outright removal wins when the dish also carries an inventory line nobody else shares. A long menu charges an operating toll that never shows up in food cost, and that toll is why trimming lifts margin even when no price changes. Each additional item eats an inventory line, a slot of refrigeration, a stretch of mise en place and a waste risk on slow-turning product.

The hidden cost: every extra item pays for space, time and waste

In a nine-square-meter kitchen with two cooks per shift, holding 43 dishes means prepping possibilities nobody orders, every single day. Dropping to 29 items freed prep time for the same crew, and the ingredients exclusive to the 14 cut dishes stopped being purchased. Market context tightens the screw: full-service menu price inflation closed at +3.6% through December 2024, according to the National Restaurant Association with BLS data, and the US cattle herd sits near 86 million head, the lowest since the 1950s, according to the USDA. With protein this expensive, holding product for slow movers is financing dead inventory. A market-cuisine restaurant walked in with 43 items, two cooks per shift and an owner convinced that removing anything meant losing sales; it walked out with 29 items and 6.2 more points of gross margin in eleven weeks. The method was boring on purpose: export ninety days of sales mix, compute dollar contribution margin per dish, multiply by portions sold, sort the list by monthly profit from top to bottom.

The case: 43 dishes, two cooks, 6.2 margin points in eleven weeks

The bottom 14 lines together contributed less than the top two. Two of those 14 ran food cost under 25%, meaning they were the "good ones" on the old board. No net sales were lost: guests who ordered the cut dishes migrated to neighboring items in the same section, already on the menu and running at higher margin. For years I cut whatever was expensive to produce and kept the cheap stuff, happy with a 27% average food cost, and the register never improved; that mistake was mine and it lasted a long time. There is one firm exception to the margin rule, and it deserves naming before somebody kills a menu with a spreadsheet: the dish that explains why people choose that restaurant does not get removed, even sitting in the bottom quartile of profit. If the room is known for one specific stew and that stew brings full four-tops that order a starter, drinks and dessert, the margin to watch is the CHECK, not the portion.

When NOT to cut: the anchor dish that carries the promise?

Dietary restrictions work the same way:

according to the Food Allergy and Foodservice study published in PMC, 36% of guests with food allergies always return to the same place, against 17% of those without them, so the single gluten-free option may earn little on its own and anchor an entire party every week. The answer is not keeping it untouched: raise its price or rebuild the recipe card until margin per portion stops being a subsidy. After the cut, resist filling the space by reflex: run the shorter menu for a full cycle and, if you add, add on proven demand rather than a house idea. Market data already shows where traction lives: hot honey appears on roughly 11% of US menus after growing 197% in four years, according to Datassential 2024 via CNBC; 71% of Generation Z prefers cold or iced drinks, according to Datassential 2025; and 58% of Gen Z and millennials pay more for beverages with health benefits, according to Hardtank 2025.

Replace with measured demand, or leave the gap open

A new dish costs an inventory line, so it enters only if its projected margin per portion beats the current menu median and it shares at least two ingredients with existing items. Diego F. Parra runs it this way inside the Masterestaurant method: cut first, measure eleven weeks, then decide whether the gap wants something or whether the gap WAS the improvement. If your kitchen runs under twelve square meters, two cooks per shift and more than 35 items, cut the bottom quartile by monthly profit in one move: the operating savings arrive before any sales loss does, and the 43-to-29 case with 6.2 margin points in eleven weeks is the typical outcome. If you run a brigade of five or more, high average check and a menu already under 25 items, go with progressive pruning: two or three items per season, with a four-week relocation test first for anything high-margin and slow-turning.

What to choose for your profile: aggressive short menu or progressive pruning?

And if you operate delivery under your own brand, the criterion gets harsher, because every extra dish stretches dispatch time and punishes your rating.

Start today with one thing only: export the last ninety days of sales mix, compute dollar margin per portion, multiply by portions and sort from top to bottom. The last line of that table is already telling you what to drop. Percentages mislead because you bank currency. A dish at 34% food cost priced at 28,000 leaves 18,480 of margin per portion; one at 22% priced at 9,500 leaves 7,410. If the first turns 380 times a month and the second 60, the monthly gap is 6.7 million against 445 thousand. Cutting by ratio kills the first. Turnover is not background context, it is half the calculation. Real menu mix from the last 90 days — not the chef's impression — turns marginal profitability per dish into monthly profit, which is the only thing that reaches the bank.

The five differences that move cash

A high margin dish that barely sells does not get cut on sight: it gets relocated. Moving it into the top third of its section, rewriting the description around the ingredient nobody else uses and dropping the currency symbol lifts its sales between 12% and 27% before any removal decision. Only if it stays flat after two menu cycles does it go. Fewer references does not mean fewer sales; it means less decision noise. With 29 dishes the guest chooses faster, the server suggests with judgment and suggestive selling stops sounding like a memorized script. Average check rises through order composition, not through price increases. The printed menu and the QR menu do different jobs and you need both. The printed menu controls the experience — service pace, menu narrative, suggestive selling, the hospitality of placing something in a guest's hand — while the QR complements with delivery, accessibility, price updates and analytics on what gets viewed and never ordered. Masterestaurant ALWAYS recommends keeping both, each in its role; dropping the printed menu for QR only costs you control of the table.