Profitable Menu: Criteria to Build It with Data, Not Taste
A profitable menu is not built by picking dishes; it is built by deciding which ones to remove. Four criteria move the margin and all four are measurable: portion cost against a written standard recipe with a hard 32% ceiling (and 32% is the maximum tolerable, never the target), contribution margin in dollars per dish, each item's share of the last 90 days of sales mix, and how that dish behaves across the three channels that now coexist — dining room, pickup and delivery — because platform commissions take 18% to 30% of every order and flip the sign of your margin without anyone in the kitchen changing a thing.
The myth says a wide menu sells more. The data says the opposite: sales mix concentrates on its own, with roughly 18% to 22% of items driving more than 70% of revenue, while the long tail eats purchasing, waste, walk-in space and, above all, kitchen minutes during the rush. Trimming that tail is the highest-return move available to an independent restaurant. And the printed menu stays: the QR is a complement for delivery, accessibility and same-day price updates, never a replacement for the piece that controls service pace and suggestive selling.
A neighborhood restaurant carried 74 items and looked healthy on revenue; the problem surfaced when the owner cross-referenced POS reports against supplier invoices and found 31 items that never cleared four weekly sales across a full quarter, dragging 19 exclusive ingredients behind them. Waste, cold storage tied up, and a fragmented purchase order that cost more per unit.
This is where the conversation stops being about cooking and becomes about cash. The menu is the only document in your business that simultaneously sets price, drives purchasing, determines rush-hour workload and shapes what the Uber Eats or DoorDash algorithm decides to surface when someone searches "pizza near me" at eight thirty on a Friday. No other piece of paper does all four at once.
Local digital channels rewrote the rules in a way almost nobody folded back into their costing. According to Bryan Solar, former head of product at Square for Restaurants, operators who treat the delivery menu as a copy of the dining room menu end up subsidizing every order, because platform commission never shows up in the standard recipe but always shows up in the P&L. His long-standing public position is that the digital menu needs its own pricing structure and its own shortened item list.
Diego F. Parra keeps pushing one sequence at Masterestaurant that sounds obvious and that almost nobody respects: measure the sales mix first, cost by portion second, and only then touch restaurant menu design. Flipping that order — starting with the layout, the photography, the redesign an agency pitched — is the most expensive way to not move your margin.
Side-by-side comparison
| Menu built on taste | Menu built on data | |
|---|---|---|
| Number of items | ✕68 to 95 dishes, grown by accumulation | ✓28 to 42 dishes, reviewed quarterly on a 90-day cut |
| Average theoretical food cost | ✕36% to 41%, no written standard recipe | ✓27% to 31%, standard recipe with a hard 32% ceiling |
| Theoretical vs actual food cost variance | ✕5 to 9 points of gap | ✓under 1.5 points of gap |
| Sales mix concentration | ✕top 10 items deliver 38% of revenue | ✓top 10 items deliver 62% to 71% of revenue |
| Delivery channel pricing | ✕same price as dining room, 18% to 30% commission uncovered | ✓12% to 18% uplift on a trimmed 14-20 item list |
| Rush-hour ticket time | ✕22 to 31 minutes, dragged by the long tail | ✓14 to 19 minutes with concentrated mise en place |
| Average check after two quarters | ✕flat, or down in real terms after inflation | ✓up 8% to 14% through suggestive selling and price anchoring |
| Reviews naming a specific dish | ✕scattered, no identifiable signature dish | ✓3 to 5 dishes own the mentions and feed local SEO |
The 31 items nobody ordered and the 19 ingredients still being bought
A neighborhood restaurant in Chapinero running 74 menu items found, after matching its POS report against purchase invoices, that 31 references never passed four weekly sales across an entire quarter, and that those 31 dragged along 19 exclusive ingredients. That is the invisible cost of a long menu: waste in the walk-in, a pricier supplier order because volume gets fragmented, and a line cook during peak staring at a ticket that fires once every ten days. Outside pressure makes it worse, since full-service menu prices hit a 9.0% year-over-year peak in 2022 per the National Restaurant Association using BLS data, and you cannot pass that adjustment through cleanly while carrying dead inventory. The concrete decision behind this figure: pull every reference under four sustained weekly sales and count how many ingredients vanish with them. Average-plate costing —taking the month's purchase invoices and dividing by plates sold— works for a management dashboard and for nothing else.
Portion costing with a standard recipe, never average-plate costing
Deciding what stays on the menu demands a standard recipe with gram weights, trim loss and the real yield of the cut, which is the only thing letting you state with certainty that the tenderloin costs 11,400 pesos and the risotto 4,200. Protein volatility turns that gap into urgency: consumer beef reached USD 5.98 per pound in May 2025, an all-time high according to Bureau of Labor Statistics figures reported by CBS News, and ground beef hit USD 6.12 per pound that June. A plate costed by average absorbs those jumps silently for months. Take your five best-selling protein recipes this month and recalculate gram weight against the actual invoice. I chased the lowest food cost percentage on the menu for years, and I was optimizing the wrong metric. A dish at 22% food cost returning 7,000 pesos of contribution loses to one at 31% returning 19,000, assuming similar turns: percentage pays the chef's ego, currency pays payroll.
Contribution margin in currency: I got this wrong for years
The 32% ceiling per portion is the MAXIMUM tolerable figure, never the target, and sorting a menu by that single criterion destroys real margin. The bar runs on different arithmetic, with average pour cost near 20% —liquor around 15%, draft beer 20%, wine between 35% and 45% per BackBar's industry guide—, so the 40% wine may well be your strongest generator of currency per check. Rank your menu by contribution in money multiplied by units sold rather than by percentage, then compare both lists. Copying the dining room menu into the digital channel means subsidizing every order, because platform commission never shows up in the standard recipe yet lands in full on the P&L. Bryan Solar, former product director at Square for Restaurants, has argued publicly for years that the digital menu requires its own trimmed set of references and its own pricing structure, and the arithmetic backs him without needing debate.
The delivery menu needs its own price structure
Add this in: the median burger price on U.S. menus reached USD 14.48 in September 2025, up 3.1% year over year per Circana via Restaurant Business, while the input climbed faster than the menu price. Discount commission from that squeeze and the dining room's star dish may be losing money inside the app. Split the delivery menu today and keep only references that survive commission with positive contribution. Diego F. Parra insists at Masterestaurant on a sequence that sounds obvious and almost nobody respects: measure the sales mix first, cost by portion second, and touch menu design only at the end. Reversing that order —starting with the pretty photo, with the redesign the agency recommended— is the most expensive way to leave your margin exactly where it was. The temptation to fix things with novelty has been measured: 52% of consumers consider an appealing limited-time offer important when choosing a restaurant, according to Technomic in 2024, and that figure pushes many operators to launch LTOs on a menu they never pruned.
The right order: sales mix, portion costing, and design last
An LTO stacked on 74 uncosted items adds chaos at peak and repairs nothing. Before designing or promoting anything, export the last 90 days of sales mix and sort it by contribution. Not every menu cut fights consumer trends, and it pays to know which way each one pushes. Plant-based menu items fell 1.9% year over year in 2024 according to Technomic via CSP Daily News, so keeping a broad vegetable section out of conviction, without a sales mix supporting it, costs you exclusive ingredients and walk-in space. Seafood plays differently: 41% of consumption happened in restaurants versus 59% at home during 2024 per Supermarket Perimeter, meaning the restaurant retains a substantial share of a category the consumer does not handle well at home, even though fresh fish cost USD 9.18 per pound in 2024 according to the USDA Economic Research Service. The decision there is about portioning and turns, not removal.
Which categories hold up and which are falling on their own?
Cross your categories against these trends before cutting on instinct. These benchmarks do not apply identically across the three sizes, and mixing them up makes you borrow someone else's decisions.
In a small restaurant of 40 to 60 seats with a single kitchen, a reasonable menu sits between 25 and 35 references, and BackBar's 20% pour cost is the fastest lever because the bar demands no extra bodies at peak. In a mid-size operation running two shifts with active delivery, the digital menu cut Bryan Solar defends usually beats any price increase, and the 32% food cost ceiling per portion must be audited by family rather than in aggregate. Across a multi-unit group, the critical point is one single standard recipe: without it, that USD 5.98 per pound of beef from BLS in 2025 becomes five different costs. Define your size first, then apply only the matching criterion.
Where these benchmarks come from and how far they reach?
The price and consumption figures in this analysis come from public United States sources:
the Bureau of Labor Statistics for beef, eggs and the menu price index, the USDA Economic Research Service for fresh fish, Circana for median burger pricing and Technomic for menu trends and limited-time offers, plus BackBar for bar pour cost. The limit is honest and worth stating: these are U.S. series, not Colombian or Mexican ones, and they serve as direction of movement and order of magnitude, never as the cost of your own recipe. A dozen Grade A eggs went from USD 2.04 in August 2023 to USD 4.95 in January 2025 according to BLS, and that more-than-double swing repeated across nearly every market on a different calendar. Use them to anticipate the curve; cost your menu with your own invoices. PORTION COSTING versus average-plate costing. Average-plate costing takes the month's purchase invoices and divides by covers sold; useful for a dashboard, useless for a decision.
The differences that actually move cash
Portion costing demands a standard recipe with gram weights, trim loss and real yield on every cut, and it is the only thing that lets you state with certainty that this steak costs $4.10 and that risotto costs $1.55. Without those numbers, every menu decision is a bet wearing a management suit. CONTRIBUTION MARGIN in dollars versus food cost percentage. I got this wrong for years, chasing the lowest percentage on the menu: a dish at 22% food cost returning $2.40 loses to one at 31% returning $6.80, assuming similar turns. Percentage feeds the chef's ego; dollars pay the lease. The right call reads both columns together and weights them by actual 90-day turns. SALES MIX by channel versus aggregate sales mix. A dish can be a star in the dining room and a disaster on delivery because it travels badly, because packaging runs sixty cents, or because commission bleeds it dry.
The differences that actually move cash — in practice
Splitting the report by channel — dining room, pickup, each app separately — exposes contradictions the aggregate hides. It is the first table I ask for when an owner tells me delivery "isn't working." DEMAND ELASTICITY measured versus assumed. Nearly every owner believes their guest is hypersensitive to price, and in practice signature dishes absorb 6% to 9% adjustments with no measurable drop in units, while beverages and sides are where guests genuinely comparison-shop. Testing two items over three weeks with the POS open costs nothing and dismantles the fear. LOCAL VISIBILITY as a menu criterion. A dish named in 40 Google reviews is a local SEO asset: it feeds "where to eat X near me" searches, sharpens the relevance of your Google Business Profile and holds position in delivery app rankings, which reward conversion and repeat orders. That dish stays even on middling margin, because margin is not its only job.
The differences that actually move cash — key points
PRINTED MENU and QR MENU, each with its own trade. The printed menu governs the experience: it sets the pace at which a guest decides, carries the menu narrative, and gives the server a surface to sell from. The QR complements with what paper cannot do — same-day price updates, delivery photography, accessibility, and analytics on what guests actually open. Anyone who kills the paper to save printing trades four hundred dollars a year for several points of average check.
Head to head: taste against numbers
What the myth claimsMyth
- "More dishes means more customers served": the wide menu as a commercial safety net.
- "Food cost gets fixed by raising prices": price treated as the only lever on the table.
- "Delivery is extra money, it drops straight in": commission booked as marketing rather than variable cost of sale.
- "If I cut that dish I lose the guest who orders it": fear of the single customer, without ever counting them.
- "A beautiful menu sells itself": restaurant menu design read as aesthetics instead of engineering.
- "The QR saves me the printing bill": killing the printed menu to claw back four hundred dollars a year.
What the data showsMasterestaurant
- Sales mix concentrates by itself: 18% to 22% of items carry more than 70% of revenue, and the long tail burns purchasing and rush-hour minutes.
- Raising price without a standard recipe moves margin for two months; portion costing and portion control move it permanently.
- At 18% to 30% commission, a dish running 30% food cost in the dining room lands at 45% effective on the app and stops paying kitchen labor.
- The dish "one customer always orders" is usually 3 to 7 tickets a month: measurable, and almost always unsustainable.
- Price anchoring and placement in the reading zone lift average check 8% to 14% without touching a single ingredient.
- The printed menu controls service pace and suggestive selling; the QR complements with delivery, accessibility and price changes. Both, each in its role.
Side-by-side comparison
| Menu built on taste | Menu built on data | |
|---|---|---|
| Number of items | ✕68 to 95 dishes, grown by accumulation | ✓28 to 42 dishes, reviewed quarterly on a 90-day cut |
| Average theoretical food cost | ✕36% to 41%, no written standard recipe | ✓27% to 31%, standard recipe with a hard 32% ceiling |
| Theoretical vs actual food cost variance | ✕5 to 9 points of gap | ✓under 1.5 points of gap |
| Sales mix concentration | ✕top 10 items deliver 38% of revenue | ✓top 10 items deliver 62% to 71% of revenue |
| Delivery channel pricing | ✕same price as dining room, 18% to 30% commission uncovered | ✓12% to 18% uplift on a trimmed 14-20 item list |
| Rush-hour ticket time | ✕22 to 31 minutes, dragged by the long tail | ✓14 to 19 minutes with concentrated mise en place |
| Average check after two quarters | ✕flat, or down in real terms after inflation | ✓up 8% to 14% through suggestive selling and price anchoring |
| Reviews naming a specific dish | ✕scattered, no identifiable signature dish | ✓3 to 5 dishes own the mentions and feed local SEO |
The 2026 numbers that settle the argument
“We cut from 74 dishes to 39 in two weeks and the team's first reaction was panic; nobody believed revenue would hold. The following quarter food cost dropped from 38.4% to 29.7%, average check climbed from $10.30 to $11.90 because servers could finally recommend with conviction, and Friday rush ticket times fell from 28 minutes to 17. What we did not expect was Google: once sales concentrated on four dishes, reviews started naming them, and "best risotto near me" searches began sending us guests who had never seen us before.”
How to build it in four measurable moves
Export units sold per item for the last 90 days from your POS and split them into four columns: dining room, pickup, and each delivery app on its own. Sort descending and mark the line where cumulative revenue hits 80%. Everything below that line goes on watch. Ninety days covers a pay-cycle rhythm, at least one holiday and a weather swing, which is the minimum needed to tell a weak dish from a bad streak. If your POS will not export by channel, pull each app's merchant portal report and merge them by hand: two hours of work.
Write the standard recipe for every dish above the line: gram weight of each ingredient, trim loss on fresh product and real yield on every cut. Multiply by current purchase prices and you have cost per portion. Build two columns per dish: food cost as a percentage of menu price, and contribution margin in dollars. The ceiling is 32%, and that 32% is a maximum rather than a goal; anything above goes to portion review, then supplier negotiation, then price, in that order. Labor and rent never load onto the plate; they belong to break-even.
Cut every item that meets two conditions: fewer than eight average weekly sales and contribution margin below your menu's median. Then declare an exception list, no more than three dishes, for items that carry local visibility despite middling margin — the dish reviews keep naming, the one the Tuesday corporate table orders, the vegetarian option that keeps a whole party from walking out. For delivery, build a separate 14 to 20 item list, drop anything that travels badly, and lift prices 12% to 18% to absorb commission.
Place your four highest-contribution dishes in the upper right third of the printed menu, where the eye lands first, and give them white space instead of boxes. Kill the right-aligned price column that invites top-to-bottom comparison, and set the figure right after the description with no currency symbol. Add one high price anchor per section. Keep the printed menu for service and publish the QR with the same menu plus delivery photography. Track average check week by week for six weeks before you touch anything again.
And with AI?
Optimize menu engineering, descriptions and the photos that sell most. Diego F. Parra is an expert in AI applied to restaurants.
Free tools to apply this now
Ecosystem tools that keep the criteria alive
Pruning a menu is a two-week exercise; holding the line when supplier prices jump in March and the urge to add three new dishes arrives in July is permanent work, and that is exactly where most restaurants relapse. These three Masterestaurant tools exist so the decision stays written, with a number attached, instead of living in the owner's memory.
Diego F. Parra built them around the same principle that orders this article: cash data first, menu decision second, layout last. Use them in that sequence.
Diego F. Parra — International consultant, expert in creating and scaling restaurants and in AI applied to restaurants, foodtech and HORECA. Methodology applied in 8.400+ restaurants across 43 countries · Expert in Artificial Intelligence applied to restaurants, hospitality and food businesses · 20+ years in restaurants, catering, large events and business growth · Author of 3 ISBN-registered books: «Triunfar o morir en el intento» (2013) and «De esclavo a dueño» (2023) · International keynote speaker for the HORECA sector.
Questions I field every week
How many items should a restaurant menu have?
How many items should a restaurant menu have?
Between 28 and 42 items for full-service, and 14 to 20 for the delivery list. The number is not magic: it comes from crossing your kitchen's real mise en place capacity during rush against sales mix concentration. If your top 10 dishes already carry more than 65% of revenue, the long tail is only costing you purchasing and minutes.
Should I raise prices or cut portions when food cost passes 32%?
Should I raise prices or cut portions when food cost passes 32%?
Check portion weight against the standard recipe first, because in eight cases out of ten the kitchen is plating more than what was costed. Then negotiate the supplier or change the cut. Price is the last lever, and when you pull it, pull it only on signature dishes, which absorb 6% to 9% with no measurable drop in units according to what your own POS records.
Can I keep only the QR menu and drop the printed one?
Can I keep only the QR menu and drop the printed one?
No. The printed menu controls the guest experience: it sets the pace of the decision, carries the menu narrative and gives the server a surface for suggestive selling, which drives much of the average check lift. The QR complements with delivery, accessibility, same-day price changes and analytics. Keep both, each in its own role.
How does my menu affect my Google Maps and delivery app ranking?
How does my menu affect my Google Maps and delivery app ranking?
Directly. A concentrated menu makes reviews name the same three or four dishes, and those mentions feed "where to eat X near me" searches and the relevance of your Google Business Profile. On delivery apps, fewer items raise conversion rate and shorten prep time, which are two of the signals the algorithm rewards with visibility.
Sector data 2026 (official sources)
Verifiable industry benchmarks from official, non-commercial sources (government, industry associations, market research) - not competitors.
| Metric | Benchmark 2026 | Source |
|---|---|---|
| Gasto por persona al quitar el signo de dólar del menú | +8,15% de gasto por persona | Cornell University, School of Hotel Administration (2009) |
| Ventas de platos con descripciones descriptivas | +27% de ventas vs platos sin descripción | Cornell University Food and Brand Lab (Wansink) |
| Aumento de ventas de un plato con foto en el menú | Hasta 30% más (y ~6,5% por plato con foto profesional) | Cornell University (investigación de diseño de menú) |
| Inflación de precios de menú en servicio completo | +3,6% a diciembre de 2024 | National Restaurant Association (Menu Prices indicator) / BLS |
| Inflación de precios de menú en servicio limitado | +3,7% en 2024 | National Restaurant Association (Menu Prices indicator) / BLS |
| Pico histórico de inflación de menú en servicio limitado | 8,2% en abril de 2023 (moderándose desde entonces) | National Restaurant Association / BLS |
Related content
Grow your restaurant with the Masterestaurant method
Applied in +8.400 restaurants across 43 countries.