Which Menu Items to Cut to Gain Profitability: Myth vs Reality

Step 1: cost every portion before you touch the menu

Without a standard recipe costed to the gram there is no menu cleanup, only hunches. What this first step must produce is a sheet with the three columns that matter: selling price, portion cost and contribution margin in DOLLARS, never in percentage. And that is where the trade's oldest trap shows up, because food cost is a percentage and you do not deposit percentages at the bank. A ceviche running at 34 % that leaves 9,200 pesos per plate and sells 300 times a month delivers 2.76 million in margin; a spotless salad at 21 % leaving 2,100 pesos and selling 40 times delivers 84,000. The salad is the one you do not need. Verification: every menu reference has a technical sheet with waste included, and margin per dish times monthly sales reconciles against the gross margin on your income statement within 3 %. Theoretical share comes from a division that fits on a napkin: 100 divided by the number of references on the menu, and the result is the sales percentage each dish would take if everything sold evenly.

Step 2: work out each dish's theoretical share and flag whatever misses 70 %

With 60 references the share is 1.67 %, so the cutoff sits at 0.7 times that figure, meaning 1.17 %. Any dish below that floor AND carrying a low contribution margin goes on the elimination list. Watch the nuance closely: both conditions together, never one alone. An operator with 60 dishes who discovers that 22 of them hold 91 % of the contribution margin does not have a kitchen problem, he has dead inventory dressed up as variety. The deliverable is a table sorted from lowest to highest share, margin sitting in the adjacent column, and a red mark on everything that meets both tests. That one does not get cut, it gets rescued: Kasavana and Smith's menu engineering files it as a «puzzle», and the costliest mistake in the trade is killing it for low turnover when its margin leads the whole menu. Rescuing it costs zero in ingredients and takes three concrete moves.

What do I do with the high-margin dish almost nobody orders?

First, change its physical position, since the top right corner of a printed menu and the first two lines of each category own the reader's eye.

Second, rewrite the description with texture, origin and method, because words sell before photographs do. Third, drill it into the shift brief with a suggested sales line each server can be measured on. Give it 30 days, then measure share again. If it still sits under 0.7 of its share after that window, out it goes, with no mourning. A kitchen that does not measure confuses affection with profit. The dish you truly need to cut is usually invisible in the sales report, because its real cost lives in the inventory it drags along rather than in its technical sheet. Build this list: EXCLUSIVE ingredients, meaning any item that enters one single reference and nothing else. Each of them carries waste, cold room space, a purchase order line and an expiry risk that nobody charges to anyone.

Step 3: chase the invisible cost your sales report never shows

A dish with two exclusive ingredients turning over once a week may be throwing away 20 % or 30 % of that purchase every cycle, and the money vanishes from margin without leaving an accounting trail per plate. The deliverable: a fresh column named «exclusive ingredients» plus the monthly cost of the waste attached to it. Add that to portion cost and you will watch profitable dishes turn negative once the arithmetic is complete. Cut in one single move rather than dish by dish, because a shorter menu defends margin and it also defends ticket time, which is exactly why large US chains have spent years trimming references (FSR Magazine). When fifteen items leave at once, mise en place compresses, the line cook stops juggling fifteen simultaneous preparations and plates leave faster, which hands you back table turns during peak hour. The rule we apply at Masterestaurant with Diego F. Parra: no category drops below four options, so the perception of choice survives, and no category climbs past twelve.

Step 4: cut in blocks and protect the speed of the pass

Before printing, run the counterfactual: if the eliminated dish took its regular customer with it, what is that customer worth per month? When that number lands under the waste the dish leaves behind, the decision has already been made for you. Number one, and the most frequent, is cutting by high food cost: that ceviche at 34 % may be your best dish in dollars, and pulling it opens a hole in margin no cheap salad will ever fill. Number two is deleting the emotional anchor, the dish that brings the whole table in even though its margin is average, because you do not sell plates, you sell occasions. Number three is trimming without telling purchasing: three cases of an exclusive ingredient sit in the cold room and waste eats the first month of savings. Number four is doing this twice a year, when it belongs on a quarterly review with the same dashboard.

Four mistakes that wreck a menu cleanup

One warning about trends: 95.3 % of US menus already offer spicy options against 91.6 % back in 2015 (Datassential), so if you are cutting, do not cut what the market keeps asking for. Cut what your own register proved nobody orders. A trimmed menu without repricing leaves money on the table, because the surviving dishes just gained visibility and visibility carries value. Raise by 4 % to 7 % anything combining high margin with high turnover, since those absorb the adjustment without losing volume, and hold steady on high-turnover items with weak margin while you redesign their sheet. Beverages give you an even bigger lever: spirits work on a 400 % to 500 % markup against wine sitting near 200 % (Provi / Parts Town, 2024), so pushing signature cocktails in the shift brief moves ticket margin faster than any kitchen adjustment. And there is a commercial window barely anyone takes: mocktails grew 280 % over four years and still hold only 1 % menu penetration (Datassential via Restaurant Dive).

Step 5: reprice the survivors before you reprint

The deliverable here is a new price list showing projected margin per reference. You will know it worked when five questions get answered with a number in hand instead of an impression. First: did weighted AVERAGE contribution margin per dish climb against last month? Second: did average ticket hold or grow, proof that you did not cut the wrong anchor? Third: did exclusive-ingredient inventory fall, and monthly waste along with it? Fourth: did peak-hour ticket time drop, measured on the POS clock rather than from memory? Fifth: does every surviving reference carry a costed technical sheet signed by the chef? Holding that dashboard six months straight produces something worth more than the savings: a run of operating data a credit officer reads far better than a balance sheet polished for December, in a region where labor informality hovers near 55 % of employment (World Bank). Start today with the margin column in dollars.

Where the usual reasoning breaks?

Food cost is a percentage, and you do not deposit percentages at the bank, you deposit pesos. A ceviche at 34 % leaving 9,200 pesos per plate and selling 300 times a month delivers 2.76 million in margin;

a salad at 21 % leaving 2,100 and selling 40 times delivers 84,000. The salad is the redundant one, however immaculate its cost sheet looks. Classic Kasavana and Smith menu engineering sorts dishes into four quadrants by crossing popularity with margin, and there sits the nuance almost nobody applies: the high-margin, low-turnover quadrant is NOT cut, it is rescued. Rename it, lift it in the printed menu and give it a description with texture, and sales return without touching a single ingredient. The dish that truly needs cutting is usually invisible in the sales report because it sells little and costs little: nobody looks at it. Yet it drags three ingredients no other dish uses, forces a separate morning prep and occupies walk-in space that turns into waste.

Where the usual reasoning breaks — in practice?

None of that cost appears on its recipe card. Price psychology comes after pruning, never before.

Dropping the currency symbol, breaking prices ending in zero and avoiding a right-aligned value column shifts the mix upward by 2 to 6 %, but applied to an unpruned menu it only sells more of the wrong dish. Demand elasticity in food service is local and cannot be imported from a study: the same dish absorbs a 12 % increase in an office district and loses 30 % of volume in a residential neighbourhood. That is why step six of this guide measures before deciding, with two weeks of observed sales rather than a survey.