Profitable menu: criteria to build it (and the five mistakes eating your margin)

First criterion: calculate each dish's contribution margin in CURRENCY, not as a percentage

Start by turning every menu item into a cash figure: selling price minus raw-material cost from the recipe costing sheet, and that figure in currency is the only one your bank account ever sees. The case opening this guide shows it bluntly: the signature dish the chef defends leaves 4,100 pesos per unit while the pasta nobody promotes leaves 6,800, even though its food cost percentage looks worse. The deliverable here is a four-column table —dish, price, cost, margin in currency— covering all 64 references, and you verify it the same day by comparing the weighted margin of one real service against that night's deposit. When a menu carries a theoretical food cost of 29% and the operating account has not balanced since February, the reason is almost always that nobody built this column. Sales mix rules over the recipe, so the second step pulls units sold per dish from the POS across a full quarter rather than one good week.

Second criterion: cross that margin with the last ninety days of sales mix

At the restaurant in the case, fourteen dishes carried 78% of tickets and eleven never sold once in a month; with that distribution, fixing the costing sheet of a dead dish changes nothing. Multiply units by margin in currency and sort descending: what appears is your REAL menu, the one paying payroll. My control reference in audits comes from the National Restaurant Association's Operations Data Abstract 2024, where star dishes capture between 35% and 45% of orders per category; if your total-profit leaders fall short of that floor, the problem sits in the mix, not in the kitchen. Cut every dish selling under one unit a day while leaving margin below the menu median, because each extra reference costs frozen inventory, waste and seconds of station time during peak service. Going from 64 references to 22 is no cosmetic trim: it frees walk-in space, shortens the guest's decision time and lifts weighted margin without touching a single price.

Third criterion: prune the menu down to between 18 and 24 live references

Wingstop runs a deliberately short menu and closed 2024 with system-wide sales near USD 4.8 billion and 20% domestic same-store sales growth, according to its published annual results. The deliverable is the delisting sheet signed by chef and owner, with an exit date on it; you verify it by measuring real food cost two inventory cycles later. Pricing works backwards: begin with the margin in currency each dish must produce to cover its station cost, and arrive at the selling price from there, instead of applying a flat three-times-cost multiplier across the whole menu. That uniform multiplier is exactly what fills a menu with cheap sides carrying enviable percentages and miserable cash. Your control rule is a WEIGHTED AVERAGE food cost below 32%, weighted by units actually sold, and to track input costs lean on the USDA food price index rather than your supplier's mood.

Fourth criterion: price from the target margin and hold the weighted average under 32%

Raise high-demand, low-margin dishes by 4% to 7%, freeze the stars, and leave untouched the items anchoring your guests' price perception. Once the arithmetic is settled, the work moves onto the page: concrete descriptions —cut origin, technique, garnish— sell better than bare names, and Brian Wansink's work at Cornell's Food and Brand Lab measured up to 27% higher sales for dishes with descriptive labels versus the same dishes without them. Put your highest currency-margin references in the first two positions of every block, kill the right-aligned price column that invites guests to scan downward for the cheapest option, and reserve one boxed panel for the dish you want to push. What gets built is a layout with an assigned position per reference; verification arrives three weeks later, comparing the new mix against the previous quarter dish by dish. The costliest error of judgement is rewarding percentage over weight: a dish at 25% food cost leaving 3,000 pesos is worse business than one at 34% leaving 7,500, and yet it passes for accounting prudence.

The mistakes that sink execution, and how to dodge them

The second error is raising prices across the whole menu at once, which pushes your regulars into comparison mode; the third, shaving portion weights until the guest notices on the plate, because the drop in return visits takes two months to reach the P&L and by then nobody links it to the decision. And a fourth, quieter one: measuring mix over an atypical week. If your menu carries allergens, do not improvise either —food allergy emergency visits exceed 200,000 a year in the United States according to Food Allergy Research & Education— and put the warning in writing before you print. Picture raising the price of your three highest-turnover dishes by 6% and losing 5% of the units sold on each: with margins of 4,100 and 6,800 pesos, that exercise usually ends with more cash, because the extra margin per unit comfortably covers the units that walk away.

The scenario almost nobody simulates before printing

Now flip it: hold the price and shave 15 grams of protein to save a few pesos of cost; cash improves in month one and rots by month three, when frequency falls and average ticket follows it down. That is the real tension of this trade, and it resolves on the price side, never on the portion side, as long as the dish has proven demand. Delivery forgives nothing here: more than 40% of adults order delivery or takeout three to five times a month according to UpMenu, and the portion arrives photographed. The menu is ready when you can quote five figures without opening a file: margin in currency for dish number one and dish number twenty, weighted average food cost for the latest inventory cycle, count of live references, and ticket share held by your four stars. The control thresholds are plain: weighted average under 32%, between 18 and 24 references, no dish with margin below half the median, and stars capturing 35% to 45% of orders per category according to the National Restaurant Association.

Closing checklist: how you know the menu was built right

Diego F. Parra and the Masterestaurant method hold that a menu gets audited every ninety days with the rigour of a monthly close, because input prices move and so does the mix. Open your POS today and pull the currency-margin column. Percentages do not get deposited. A dish at 25% food cost leaving 3,000 in margin is a worse business than one at 34% leaving 7,500, and a menu that rewards percentage over currency fills up with cheap sides nobody orders as a main. This is the costliest criteria error on the financial side, because it looks like accounting prudence and is really misapplied arithmetic. Sales mix outranks the recipe.

Where the margin actually breaks?

You can hold the finest costing sheet in the country and still lose money when 60% of your tickets flow to the dish with the weakest marginal profitability;

according to Sheryl Kimes, professor emerita at the Cornell University School of Hotel Administration and a leading voice in restaurant revenue management, what moves a service period's result is less the unit price than the combination of what sells and how fast the table turns. Every extra reference costs money before it sells. One added dish drags its own SKUs, walk-in space, an extra line procedure and a waste risk. Past twenty-four references, complexity cost climbs faster than incremental revenue, and that shows up in waste long before it reaches the income statement. Price psychology is not decoration. Dropping the currency symbol, setting prices right after the description instead of in an aligned column and placing an expensive anchor above shifts average check without touching a gram of product.

Where the margin actually breaks — in practice?

Cornell's hotel school measured double-digit spending swings from menu price formatting alone. Price holds on perceived value, not on cost. Build the menu from cost outward and your supplier sets your ceiling.

Build it from perceived value inward and the guest sets it, and the guest almost always pays more than the owner assumes, provided the plate justifies the gap with something tangible.