How to design a menu that increases profits: traditional method vs the Masterestaurant method

What is the best way to design a menu that increases profit?

Rank every dish by CONTRIBUTION MARGIN in dollars and by units sold, not by its food cost percentage:

that cross is the Masterestaurant method, and it suits an independent with 15 to 40 tables, a mixed sales channel and a chef-owner who still works the line. The 3x markup sets price with a single rule and leaves out the variable that pays payroll, which is how much cash each dish leaves multiplied by how often it goes out. A dish at 31 % food cost selling 6 units a day leaves less monthly margin than one at 35 % selling 38, and the chef who watches only the percentage kills the second one. Repeat that across three or four dishes and it costs between 900 and 2,400 USD of monthly margin in a 40-table operation. Full-service menu inflation closed 2024 at +3.6 % year over year (National Restaurant Association / BLS), well below what specific inputs rose, so raising the whole card by 3.6 % solves nothing either.

Best for mixed-channel operations: separate pricing per channel, not one card

If more than 20 % of your sales come through delivery apps, you need a card with its own prices for that channel instead of the dining room sheet copied over. Platform commission runs between 24 and 30 % of the sale price, so a dish that leaves 11 USD of margin at the table drops to 4.20 USD in the app, and that collapse shows up in no report until month-end refuses to reconcile. With 900 monthly app orders and margin eroded by 6.80 USD per dish, that is 6,120 USD evaporating without anyone making a decision. The fix is not raising everything by 30 %: identify the five or six dishes that travel well, adjust their app price between 15 and 22 %, and pull from the channel the ones that arrive cold and generate refunds. Pasta dishes, carrying 65 to 70 % margins (Sauce, 2025), absorb that commission best.

When NOT to pick the popular option?

Three situations make the contribution-margin method the wrong call, and it is fairer to say so upfront. First:

a kitchen with fewer than 12 dishes and a low average check, where the simple markup works because cost dispersion is minimal and the owner's hour pays better on the line than in a spreadsheet. Second: a place open less than 90 days with no sales history per dish, because ranking by popularity without data is ranking by intuition dressed as method, and there the multiplier is an honest floor while the numbers accumulate. Third: fixed-price or tasting-menu operations, where the guest does not choose dish by dish and the unit of analysis is the whole menu. Outside those three scenarios, with over 1 million foodservice locations competing in the United States (National Restaurant Association, 2025), pricing with a single multiplier is handing margin away. Be wary of any menu method that trips one of these four signals.

Red flags when comparing menu design methods

If someone proposes a single multiplier for the entire card without asking how many units each dish sells, they are pricing blind; the US cattle herd sits near 86 million head, the lowest since the 1950s (USDA, 2025), and the beef plate cannot carry the same rule as the pasta. If the consultant never asks for your item-level sales report from the last 90 days, there is no menu engineering happening, only graphic design. If the proposal says nothing about prep time per dish, it ignores that a kitchen with four people on the line on a Friday at 9 p.m. has a physical ceiling of complex plates. And if they promise recovered margin from redesigning the sheet alone, without touching recipes or prices, they are selling you typography. If you are still on the line three or four services a week, the lever with the best return is not redesigning paper but cutting the dish count down to what your real kitchen executes without losing quality.

Best for the chef-owner who cooks: cut the card before you redesign it

I got this wrong for years: I defended the 3x markup because it was teachable in ten minutes to a cook who had never seen a costing sheet, and that simplicity carried genuine value in kitchens where nobody tracked numbers. But the bottom 20 % of a card usually contributes under 5 % of sales while consuming inventory, waste and walk-in space. Pull those dishes, reassign the mise en place to the six that actually sell, and food cost variance drops without raising a single price. On a 38-dish card, removing the eight slowest movers frees between 400 and 700 USD a month of tied-up inventory and shortens kitchen ticket time. The physical sheet moves money, though less than most designers promise. Kasavana and Smith's menu engineering, published at Cornell in 1982, solved half the problem by crossing popularity with margin, but it never said where each dish goes on the page or how many a real kitchen supports.

The paper decides too: attention zones and price anchors

The Masterestaurant method that Diego F. Parra applies closes that half: the two or three highest contribution-margin dishes sit in the upper right zone of a two-page card or the first third of a single page, with a higher-priced anchor dish beside them that makes the one you want to sell look reasonable. Drop the currency symbol and set the price inline with the text, never in a column. These adjustments shift sales mix by 2 to 6 percentage points, which in a place billing 180,000 USD a year means 3,600 to 10,800 USD of reassigned revenue. Say you apply a flat 4 % across all 38 dishes and sleep well. With limited-service menu inflation at +3.7 % through 2024 (National Restaurant Association / BLS), it looks prudent enough. Trouble arrives in the third month: high-rotation, low-margin dishes absorb the increase without resistance because guests order them out of habit, while high-margin, low-frequency dishes lose the few units they sold, because there the guest does compare.

What happens if you raise every price 4 % and stop there?

Mix shifts toward what pays you least, average check climbs 2 % and gross margin falls. That is the paradox I see most often in freshly adjusted cards:

more revenue, less cash. The way out is asymmetric: raise the high-rotation dishes whose margin sits below average by 7 or 8 %, freeze the high-margin ones, and measure mix at 30 days with the item-level sales report. If you want trend without wrecking your costing, come in through ingredients with low unit cost and high perceived value, and place them in dishes that already exist. Hot honey now appears on roughly 11 % of US menus, up 197 % in four years (Datassential 2024, via CNBC), and its cost per portion is measured in cents. Beverages behave the same way: 71 % of Gen Z prefers cold or iced drinks (Datassential, 2025) and 58 % of that generation and millennials pay more for drinks with a health benefit (Hardtank, 2025).

Best for trend-driven cards: test new flavors without breaking the costing

A bar that swaps two hot references for two cold ones at 70 % margin recovers the investment in weeks. The rule holding all of this together: every new ingredient must land in at least two dishes, or it does not come in. An input living in a single item is waste waiting for a date. Markup decides on a percentage; marginal profitability decides on dollars. A dish at 31 % food cost selling 6 units a day leaves less monthly cash than one at 35 % selling 38, and the chef watching only the percentage kills the second. Repeat that across three or four dishes and it costs 900 to 2.400 USD of monthly margin in a 40-table operation. The traditional method is blind to channel. A dining room menu copied straight into a delivery app carries a 24 to 30 % commission on menu price, so the dish that left 11 USD of margin at the table leaves 4,20 USD in the app, and nobody notices until month-end refuses to balance.

Where the two methods really split?

Classic menu engineering stops at the matrix; the Masterestaurant method reaches the paper.

A menu sheet has measurable attention zones, and moving a star dish from the bottom of a column to the top right corner shifts its sales mix without the guest ever sensing a nudge. Markup assumes uniform elasticity. In practice, Technomic's 2026 pricing analysis shows starters and beverages absorbing 6 to 9 % increases with demand falling under 2 %, while entrées above 25 USD lose 5 to 8 % of units on the same percentage move. Raising everything 5 % punishes you exactly where it hurts. Then there is the printed menu. Plenty of operators read that QR saves on printing and pulled the paper; the saving runs 600 to 1.400 USD a year and the lost suggestive selling runs higher. Masterestaurant recommends BOTH: the printed menu governs service pace, narrative and the server's recommendation; the QR covers delivery, accessibility, price changes and analytics.