How to price a restaurant menu: one dish, from the recipe card to the number on the page

Menu price = plate cost ÷ target food cost percentage. That is the formula every guide leads with, and it is genuinely the right starting point. It is also about a third of the job. The number you print has to survive three separate tests: it must cover what the dish actually costs to produce, it must earn enough gross profit dollars given how often the dish sells, and it must look reasonable next to the restaurant down the street.

This guide runs one dish through all three, end to end, with every formula written out and every benchmark attributed. It also does something most pricing articles skip: it says clearly which pieces of menu psychology have research behind them and which two of the most repeated ones have been tested and failed.

The five formulas, with units

Everything below is built from five calculations. Keep them straight and the rest is arithmetic.

1. Plate cost — what one serving costs to produce, in dollars.

Plate cost = Σ (AP cost ÷ yield % × portion) + Q-factor

2. Food cost percentage — what your ingredients consumed as a share of food sales, over a period.

Food cost % = (Beginning inventory + Purchases − Ending inventory) ÷ Food sales × 100

3. Target menu price — the price that hits your intended food cost, in dollars.

Target menu price = Plate cost ÷ Target food cost % (as a decimal)

4. Prime cost — the control number that includes people, in dollars.

Prime cost = COGS + Total labour cost (wages + payroll taxes + benefits)

5. Contribution margin — gross profit dollars per plate sold.

Contribution margin = Final menu price − Plate cost

Two of these are routinely confused. Food cost percentage and prime cost percentage are not the same metric. Food cost measures purchasing, portioning and waste discipline. Prime cost measures whether the whole operation can pay for itself. Two dishes can both run 28% food cost while one takes six hours of prep and the other takes five minutes, and only prime cost sees the difference.

Step 1: calculate the plate cost

Plate cost is what one serving costs you to put in front of a guest. It is not the invoice price of the ingredients, because ingredients arrive in a state you cannot serve. The gap between the two is where most under-pricing starts.

Two adjustments close it. Yield percentage converts an as-purchased cost into an edible-portion cost, accounting for trimming, butchering and cooking loss. Buy a whole tenderloin at $18 a pound and yield 70% usable portions, and your real cost is $18 ÷ 0.70 = $25.71 per usable pound, not $18. Skipping this step guarantees your calculated food cost sits below the one your profit and loss statement reports, permanently.

The second adjustment is the Q-factor, sometimes called cover cost: the untracked things every guest consumes that appear in no recipe. Bread and butter, salt, pepper, tabletop condiments, fryer oil degradation, and a baseline allowance for misfires. It is commonly applied as 5% to 10% of raw ingredient cost. It is an estimate, and it should be labelled as one, but omitting it entirely is a bigger error than estimating it.

Worked example, part one

A filet mignon plate at an independent full-service restaurant. All figures are illustrative and should be replaced with your own invoices.

Component	Edible-portion cost
Beef tenderloin, 8 oz EP after yield testing	$12.50
Haricot vert, 2 oz EP	$0.87
Potato pavé, 3 oz EP	$0.34
Truffle demi-glace, 1.5 oz	$1.47
Coffee rub, 1 tsp	$0.25
Raw ingredient total	$15.43
Q-factor at 10%	$1.54
Plate cost	$16.97

Step 2: apply a target food cost percentage

Once you have a plate cost, the target menu price falls out of one division. At a 35% food cost target:

Target menu price = $16.97 ÷ 0.35 = $48.48

That is a mathematical starting point, not an answer. It ignores what guests will pay and what the three comparable restaurants nearby charge for a steak. In the worked example, the operator sets the final price at $49.00, which gives an actual food cost of $16.97 ÷ $49.00 = 34.6% and a contribution margin of $49.00 − $16.97 = $32.03 per plate.

Notice which number matters at service. You bank $32.03, not 34.6%. Percentages tell you whether your purchasing is disciplined; dollars tell you whether the restaurant survives.

What is a good food cost percentage?

The honest answer is that no universal target exists, and the widely repeated 30% rule has no traceable source. What does exist is a real, recent, named benchmark from a large operator sample.

The National Restaurant Association’s 2025 Operations Data Abstract, drawn from financial and operating data provided by more than 900 restaurants nationwide, reports that in 2024 food and non-alcohol beverage costs ran at a median of 32.4% of sales in limited-service restaurants and 32.0% in fullservice restaurants. Higher-volume fullservice operators, at $2 million or more in annual sales, came in lower, at a median of 31.0%.

Those medians are close enough together to be worth pausing on, because they contradict the tidy segment tables that circulate online. Limited service does not automatically run a lower food cost than full service. Its advantage is elsewhere, in labour.

Metric, 2024 medians	Limited service	Fullservice
Food and non-alcohol beverage cost	32.4% of sales	32.0% of sales
Salaries and wages including benefits	31.7% of sales	36.5% of sales
Prime cost (sum of the two)	64.1% of sales	68.5% of sales

Source: National Restaurant Association, 2025 Operations Data Abstract, reporting on 2024. The prime cost row is our own addition of the two published medians, not a figure the Association publishes.

For context on how elevated that is, the Association notes that between 2010 and 2016 labour averaged roughly 33% of sales for fullservice and 28% for limited-service operations. Roughly three and a half points of margin have moved from the operator to the payroll in under a decade, and no amount of ingredient discipline recovers it.

A caution about segment tables. You will find plenty of pages listing food cost by segment to the percentage point — QSR 28 to 32, fast casual 27 to 30, fine dining 30 to 35. Those ranges are planning heuristics circulated by software vendors and consultancies, not audited statistics, and they rarely carry a year or a sample size. Use them as a sanity check, not as a target. The full picture of published cost and margin data by service type is in our restaurant profit margin statistics.

Beyond food cost: why prime cost is the number to watch

Prime cost is cost of goods sold plus total labour, and it is the single most useful control figure in an independent restaurant. Food cost alone cannot see the two things that most often sink a menu item: prep labour and the opportunity cost of a bottlenecked station.

Run the comparison on two dishes with an identical 28% food cost. One is a composed pasta assembled in four minutes on the line. The other is a scratch terrine requiring three hours of prep, a dedicated cooler shelf, and a two-day lead time. Their food cost percentages are the same. Their actual cost to the business is not remotely the same, and only prime cost surfaces it.

This is why serious operators review prime cost weekly, sometimes daily, rather than waiting for a monthly P&L. A one-point improvement in prime cost flows straight to the bottom line without touching a single menu price, which makes it the cheapest margin available to most kitchens.

Three pricing methods, and when each one is right

Cost-based pricing is where you start, but no restaurant prices its whole menu that way. In practice, operators blend three methods, and the blend is what makes a menu profitable rather than merely accurate.

Method	How it works	Best used for	Its blind spot
Cost-plus	Plate cost ÷ target food cost %	Establishing a floor for every item	Ignores what guests will pay and what neighbours charge
Competition-based	Price anchored to the local comparable set	Commodity items guests can price-compare (burger, margherita, house wine)	Copies competitors’ mistakes; ignores your actual costs
Value-based	Price on perceived value to the guest	Low-cost, high-perceived-value items: soups, pasta, fountain drinks, cocktails	Overreaches quickly if the product does not support the claim

The workable menu uses all three at once. Low-cost items priced on value (running a 10% to 15% food cost) subsidize the centre-of-plate proteins that must run at 40% to 50% because guests know roughly what a steak costs. Cost-plus sets the floor nobody goes below. Competition-based pricing constrains the handful of items a guest can price-check from memory.

One method to treat with care: algorithmic dine-in surge pricing. When Wendy’s announced a $20 million digital menu board investment in early 2024 and used the phrase “dynamic pricing”, the market read it as peak-hour surge pricing, competitors ran campaigns promising stable prices, and the company spent days publicly clarifying that it meant off-peak discounting. Off-peak discounts, happy-hour pricing, delivery-channel pricing and overstock promotions are widely accepted. Raising the dine-in price of a burger because the dining room is full is not, and the reputational cost tends to exceed the margin.

Menu engineering: classify every item

Menu engineering plots each item on two axes — how often it sells, and how many gross profit dollars it contributes — and assigns a specific action to each quadrant. The framework comes from Michael Kasavana and Donald Smith at the Michigan State University School of Hospitality Business in 1982, adapting the Boston Consulting Group growth-share matrix to restaurant menus. Its foundational insight is the one from earlier in this guide: you bank dollars, not percentages.

The two thresholds are calculated per category, not across the whole menu:

Popularity threshold  = (1 ÷ number of items in the category) × 0.70
Profitability threshold = average contribution margin of the category

An item above both is high on that axis. Plotting each item gives four quadrants:

Quadrant	Popularity	Profitability	What to do
Star	High	High	Protect it. Hold quality and specification, keep it prominent, resist casual price changes
Plowhorse	High	Low	Raise the price modestly or adjust portion or garnish cost. It carries traffic, so do not remove it
Puzzle	Low	High	Reposition, rename with sensory description, train staff to recommend it. Volume is the missing piece
Dog	Low	Low	Remove. If it must stay for dietary or menu-completeness reasons, reformulate or reprice and de-emphasize

Two refinements worth knowing. First, the original model treats labour and overhead as fixed, which understates the true cost of labour-intensive dishes. Pairing menu engineering with activity-based costing, where prep time is traced to specific dishes, routinely reclassifies a scratch-made item from Star to Plowhorse once the back-of-house hours are attributed. Second, the classification is worthless without accurate plate costs, which is why Step 1 comes first. The published data behind popularity and profitability patterns sits in our menu engineering statistics.

Pricing psychology: what the evidence actually supports

This is where restaurant advice is least reliable, because the consulting industry has circulated several claims for decades that were eventually tested and did not hold. Here is the split.

Removing the dollar sign — supported

Dropping currency cues from menu prices is one of the better-evidenced interventions in the field. Sybil Yang, Sheryl Kimes and Mauro Sessarego of the Cornell Center for Hospitality Research tested three price formats in a working upscale casual restaurant, comparing numerals with a dollar sign, bare numerals, and prices written out as words. Guests reading the format without monetary cues spent about $5.55 more per check, roughly 8% more, than those reading prices with a dollar sign ($ or Dollars: Effects of Menu-price Formats on Restaurant Checks, 2009).

Charm pricing — supported, with limits

Prices ending in .95 or .99 exploit left-digit bias: the brain anchors on the first digit, so $19.99 reads as meaningfully cheaper than $20.00. The effect is well documented in retail and translates into quick-service and casual dining. It also carries a signal about positioning, which is why fine dining generally avoids it. Charm pricing communicates value; if value is not what you are selling, it works against you.

Descriptive labels — supported, with a caveat about the literature

Sensory and provenance descriptions (“slow-braised”, “stone-ground”, “from Hood River”) lift both purchase intent and perceived quality relative to a bare item name. The single most-cited study on this is one of several papers by Brian Wansink, whose broader body of work suffered retractions and serious data-integrity findings in the late 2010s. The specific effect has been supported by later independent work, so the practice stands, but do not attach a headline percentage from the original paper to it.

The golden triangle — tested and refuted

The claim is that a diner’s eye lands in the middle of the menu, then moves to the top right and top left, and that high-margin items belong in that triangle. It traces to an unverified 1987 diagram rather than to research. In 2012, Sybil Yang published an eye-tracking study in the International Journal of Hospitality Management that tested it directly and found diners read menus sequentially, top to bottom, like a book, with no sweet spot where gaze lingered (Eye movements on restaurant menus, 2012).

The useful part of that finding is the inverse. Yang’s data did identify low-attention areas, mostly in the margins. So the actionable rule is negative: nothing about layout will levitate a high-margin dish into extra sales, but a badly placed one can be effectively hidden. Order and grouping matter more than geometry.

Choice overload — weaker than its reputation

The idea that long menus paralyse diners is usually anchored to the jam-tasting study, which was a grocery display, not a restaurant. A meta-analysis by Scheibehenne, Greifeneder and Todd examined 63 conditions across 50 published and unpublished experiments with 5,036 participants and found a mean effect size of virtually zero, with high variance between studies (Can There Ever Be Too Many Options?, Journal of Consumer Research, 2010).

For menus specifically, that means the fix is structural rather than subtractive. Guests arrive with preferences and cravings; clear categories and sensible chunking help far more than cutting the menu to seven items per section because a consultant said so. If your menu is genuinely too long, cut it because the kitchen cannot execute it or because the Dog quadrant is full, not because of choice overload.

The summary table

Technique	Evidence	Primary source	What it actually shows
Removing currency symbols	Strong	Yang, Kimes & Sessarego, 2009	About 8% higher spend per check without monetary cues
Charm pricing (.95 / .99)	Strong in retail, applies to QSR and casual	Left-digit bias literature	Communicates value; unsuitable for fine dining positioning
Descriptive labels	Moderate	Multiple, some tainted by later retractions	Lifts perceived quality and purchase intent; treat headline percentages with caution
Decoy / anchor items	Weak, contested	Attraction-effect replication work	Reliable in the lab, unreliable in naturalistic menu choice
Golden triangle layout	Refuted	Yang, 2012	Diners read sequentially; no sweet spot exists
Choice overload	Near zero on average	Scheibehenne et al., 2010	No general effect; categorization beats arbitrary menu shortening

Waste is a pricing lever you do not have to explain to guests

Every point of waste is a price increase you did not have to announce. ReFED reports that 29% of the roughly 240 million tons in the US food supply goes unsold or uneaten, with total surplus valued at about $381 billion, and that in foodservice specifically 70% of waste is plate waste — food that guests were served and did not eat.

That last figure reframes portion sizing. If most foodservice waste leaves on the plate rather than in the bin behind the line, then over-portioning is simultaneously a cost problem and a guest-satisfaction problem, and correcting it improves both. The full set of published numbers is in our restaurant food waste statistics.

When and how to raise prices

The mechanics are simple; the timing and the friction are what matter.

1. Recost first, then decide. Reprice off updated invoices, not off a percentage applied across the board. A blanket 5% increase raises the price of your Stars, which needed no help, and leaves your Plowhorses under-earning.
2. Move in small, frequent steps. Several small adjustments read as normal; one large annual jump reads as a decision made about the guest.
3. Never move everything at once. Guests price-check a handful of anchor items from memory. Hold those, move the rest.
4. Do the highest-margin item last. If a dish is already a Star, a price rise risks the volume that makes it one.
5. Say nothing, but be consistent. A price change explained on the menu invites scrutiny of the price. A price change that is simply current does not.

The operational obstacle is usually printing. A restaurant that has to reprint and re-laminate every menu, insert and table tent to move four prices will delay the change until the arithmetic is badly overdue — and that delay is expensive precisely when costs are moving fastest. The National Restaurant Association projects $1.55 trillion in industry sales for 2026 against just 1.3% real growth, which is another way of saying most of the top-line movement in the market is price, not extra covers.

That is the practical argument for keeping the menu in a format you can edit. When prices live on a page rather than on card stock, the recosting cycle and the publishing cycle are the same length, and a Tuesday price correction goes live on Tuesday. Our guide to digitizing a restaurant menu covers the migration itself.

The worked example, end to end

Pulling one dish through every step:

Step	Calculation	Result
1. Raw ingredients, at edible-portion cost	$12.50 + $0.87 + $0.34 + $1.47 + $0.25	$15.43
2. Add Q-factor at 10%	$15.43 × 0.10	$1.54
3. Plate cost	$15.43 + $1.54	$16.97
4. Target price at 35% food cost	$16.97 ÷ 0.35	$48.48
5. Final price after competitive check and rounding	—	$49.00
6. Actual food cost	$16.97 ÷ $49.00	34.6%
7. Contribution margin	$49.00 − $16.97	$32.03
8. Menu engineering position	Above category average contribution margin, sales mix to be measured over a full period	Star or Puzzle, pending mix data

Step 8 is the one most operators never reach, and it is the one that decides what to do next. If that filet is above the average contribution margin and above the popularity threshold, it is a Star and you protect it. If it is above on margin and below on mix, it is a Puzzle, and the fix is menu position and staff recommendation rather than another price change.

Two more things worth pairing with this process. Menu format changes how predictable your costs are before you price anything: a fixed-price menu tells the kitchen what most covers will eat, which is why our guide to a la carte vs prix fixe sits upstream of this one. And if you are unsure which cost benchmarks apply to your operation at all, start by pinning down what kind of restaurant you are running with our guide to the types of restaurants.

Duckhub gives you a menu you can recost and republish the same day: edit a price, mark an item sold out, and every guest sees the current version instantly. Free plan, 30-day trial on the paid tiers, and 0% commission on orders.