How to design a menu that increases profits: the expensive mistakes and the method that survives an audit

The best call for a room under fifteen tables: cut the lineup before touching the print

If you run fewer than fifteen tables with a short crew and tight cash, the best menu decision is to CUT items, not redesign the paper. The arithmetic decides it: the Economic Commission for Latin America and the Caribbean estimates that roughly 70% of the region's microenterprises do not make it past five years, and in food service the clock runs faster because you buy perishables every day at prices that move every week. A forty-item card in a two-station kitchen forces you to carry slow-moving references that spoil in the walk-in, and that waste comes out of the same pocket that pays payroll. Diego F. Parra has argued from Masterestaurant for a sequence almost nobody enjoys: measure dish by dish first, cut second, and only then send to the printer something that already reflects a decision made with numbers. Food cost expressed as a percentage is the wrong unit for deciding what stays on the card, which is why so many owners pull precisely the dish that carries the register.

Why does food cost as a percentage make you pull the wrong dish?

Take two real items from any menu in the region: one at 38% cost that leaves eight dollars of contribution margin per unit, and another at 24% that leaves two.

The hallway conversation kills the 38% dish, the one that was paying rent. Michael Kasavana and Donald Smith formalized the right frame at Michigan State University in the early eighties: cross popularity with contribution margin and sort every item into star, workhorse, puzzle or dog. Four decades later it remains the most solid model available, and it remains the least applied one here. Fixing that single criterion usually moves operating margin by three to five points without raising a single price. When the kitchen has two stations and three people during peak service, the scarce resource is not money, it is MINUTES. Every dish on the card consumes inventory, station time, the cook's attention and the server's mental space, and those four run out before the budget does.

Best for two-station kitchens: a short card frees minutes, not just inventory

The FAO calculates that Latin America and the Caribbean loses around 220 million tonnes of food per year, worth close to 150 billion dollars, and a measurable share of that starts at the service stage: oversized cards that force you to hold references almost nobody orders. The IDB Group's #SinDesperdicio initiative works on SDG target 12.3 for exactly this reason. A short, properly costed menu lifts the operator's margin, cuts waste and steadies payroll, three outcomes no direct subsidy achieves at the same cost-effectiveness. Cutting is not always the play, and it is better said upfront. First scenario: the destination menu, where the slow-moving dish is the reason a party of eight picks your house; pulling it sinks the whole table, not one cover. Second: an operation with a high share of guests who have dietary restrictions, where a short card excludes the person who decides where the group goes — remember that eight food groups cause 90% of food allergies according to the US Food and Drug Administration under FALCPA, so a card without a safe alternative loses entire reservations.

When NOT to take the popular route: three scenarios where cutting the card costs you money?

Third: the business still in opening ramp, under six months of sales history, where you simply do not know yet what sells; cutting there is guessing.

In all three the right lever is pricing and description engineering, not the scissors. Four signals in this trade predict a bad menu project, and all of them surface in the first meeting. First: someone shows you a graphic proposal before asking for ninety days of item-level sales; without that file nobody can classify a dish. Second: they promise to «raise the average check» without naming contribution margin per unit, which is where the money actually lives. Third: they suggest tagging half the card as chef's favorite — the effect is real, an item marked as most popular lifts orders by 13% to 20% according to NeatMenu, but it dissolves once everything carries a badge. Fourth, and the most expensive: nobody asks about kitchen capacity at peak hour, and you end up with a beautiful card your crew cannot execute in twenty minutes.

This fits you if you sell delivery: one dish, two margins, and one card will not do

If more than 25% of your sales run through platforms, you need two menu engineerings, not one. The commission eats the item's contribution margin before you ever see the deposit, and a dish that is a star in the dining room can be a dog in delivery with exactly the same recipe cost. The underlying pressure is real: input costs rose roughly 35% on food and 35% on labor since 2019 according to the National Restaurant Association, so there is no cushion left to absorb a commission with the wrong dish's margin. What works is easy to say and annoying to execute: a reduced digital card holding only items that survive the commission and travel well, prices calculated on margin net of commission, and the high-margin puzzles reserved for the dining room, where you control the suggestive sell. Almost everyone redesigns the print first because it is the visible, gratifying part, and inverting that sequence is the costliest mistake in this trade.

The right order of the work, and what happens when you invert it

Suppose you hire a designer, print two thousand cards and then measure: you find eight dishes that do not cover their station cost, you pull those eight, and the two thousand cards are obsolete — you paid twice for the same job and lost the quarter. The order that actually holds the register runs the other way: ninety days of item-level sales, contribution margin per unit, Kasavana-Smith classification, cut, repricing of the puzzles and, last, design. Here is a judgment with no middle ground: the print is the final decision, never the first. And an honest concession, since I defended redesign as the main lever for years: a beautiful card does lift the check, but only once it sits on dishes that leave money behind. Open the point of sale and export units sold per item for the last ninety days; that file, not the chef's opinion and not yours, decides what stays.

What to do on Monday: the file that decides the card?

Next to each dish write the recipe cost updated with this week's prices and subtract: what remains is contribution margin per unit. Multiply by units sold and rank the list from highest to lowest.

The bottom of that ranking, if it adds nothing to the menu's photograph either, comes off; high-margin, low-selling items get rewritten on the card and drilled at the pass; low-margin, high-selling items get repriced or a new side. All of it is two hours in a spreadsheet, with no designer and no print shop. Design comes afterward, once you know what deserves to be printed. The underlying split is the unit of measure. The default thinks in food cost percentages, which is convenient because it fits into a hallway conversation, and so the owner pulls the 38% dish that yields eight dollars of margin and keeps the 24% one that yields two.

Where the two roads genuinely diverge?

The correct method thinks in contribution margin per unit multiplied by units sold, meaning the money that actually reaches the register.

An operator who corrects that single criterion usually moves three to five points of operating margin without touching a single menu price, and without the guest noticing anything at all. The second break point is sequence. Most operators redesign the printed piece first because it is the visible, gratifying part of the work, and only afterwards discover that the beautiful menu still carries the same twenty-eight ungovernable dishes. We invert the order: cut, recost, reorder for price psychology, and let graphic design come last, once it knows what it has to highlight. Diego F. Parra has held that sequence across two decades of Masterestaurant method implementations in the region, and the argument is not aesthetic: a designer handed an uncosted menu optimizes for the eye, never for the till.

Where the two roads genuinely diverge — in practice?

Third, and this is where resistance peaks: a shorter menu does not shrink sales. Intuition says fewer options scare guests away, and consumer-behaviour evidence has pointed the other way since Sheena Iyengar and Mark Lepper's Columbia work on choice overload.

Faced with too many alternatives the guest defers, picks the familiar and spends less. I got this wrong for years, recommending broad menus for neighbourhood spots on the belief that variety built loyalty; the sales-mix numbers eventually showed that 80% of revenue came from eleven dishes while the rest generated nothing but dead inventory. There is a fourth, institutional layer that rarely enters the operational conversation. A well-costed menu produces structured data —unit cost, turnover, waste, margin per item— and that is precisely the data commercial banks lack when they assess a gastronomic MSME and end up denying credit or charging a punitive risk premium. Alternative scoring built on operational data, which BID Lab has piloted across several regional programs, feeds on exactly that information.

Where the two roads genuinely diverge — key points

Building the menu rigorously improves this month's result; it also builds the record that later unlocks formal financing.