Restaurant pricing: the 2026 numbers, before and after

How much do input costs really weigh when you set prices in 2026

The producer price index for all food in the United States sits 35% above its February 2020 level, according to USDA ERS using BLS data through May 2026, and that figure —not the food cost percentage you carry in your head— is the floor from which the menu has to be rebuilt. The same BLS reported the final demand producer price index rising 3.0% in 2025, after 3.5% in 2024, with services running at 3.2% against 2.5% for goods. Translated into a kitchen: labor and utilities get expensive faster than chicken does, and an operator who adjusts prices looking only at the supplier invoice is fixing half the problem while the other half, the one that leaves through payroll and energy, eats the margin quietly. A growing sector does not mean your restaurant earns more, and the numbers prove it with uncomfortable bluntness. The National Restaurant Association projects roughly US$1.55 trillion in United States sales for 2026, despite cost pressure, after projecting $1.5 trillion for 2025.

Market size will not save you: the sector grows and your profit does not

In Canadá, Statistics Canadá recorded food and beverage services sales of CAD 96.5 billion in 2024, up 4.0% over 2023. Mexico moves 300 billion pesos, per CANIRAC in 2024. Every one of those curves points up, and still the average operator closes the year with single-digit profit. The reason is arithmetic: growth is measured in sales, survival is measured in CONTRIBUTION, and a badly priced menu turns each point of growth into more volume at a margin that does not cover the bill. Monday's decision: stop celebrating monthly revenue and start reporting margin pesos per plate sold. ReFED valued foodservice food surplus at US$157 billion during 2024, equal to 14% of the channel's sales, and that 14% alone explains why so many prices that look right on the recipe card are wrong at the register.

Waste is an invisible tax you are already paying

If your recipe calculates a 30% theoretical food cost while your operation dumps, overproduces or misportions the equivalent of a tenth of what it sells, real food cost lands near 34% or 35% and you are subsidizing every plate with money you believed you had collected. The fix is not raising price: measure waste by station for two straight weeks, load the REAL waste onto the recipe card of the plate that creates it, and only then reprice. A plate that produces systematic waste is neither expensive nor cheap; it was mis-costed from the start. Delivery changed the geometry of price and very few operators have redone the math. UpMenu documented in 2024 that 37% of adults order delivery at least once a week and that more than 40% order delivery or takeout three to five times a month. That volume arrives through a channel that charges commission, demands packaging and punishes station time with orders that cannot be deferred.

Price gets defended through the channel, not only through the menu

When you publish the same price at the counter and on the platform, you are giving away the entire commission differential, which in most markets runs between 15 and 30 points of the order. The solution I apply with Masterestaurant clients is simple and unpopular: a dedicated delivery menu, fewer references, prices built on margin NET of commission, and dishes chosen for their resistance to transport. Consultant's read: if a plate cannot survive twenty minutes in a thermal bag, it does not belong on that menu at any price. Run the scenario. The National Restaurant Association, citing BLS data from 2024, counted 6.2 million 16-to-19-year-olds in the United States workforce, 900,000 more than in 2019, and that rebound eased wage pressure in the entry segment for a while. Now assume the flow reverses, which is exactly what happened between 2000 and 2019.

What would happen if your margin depends on a labor pool that is evaporating?

Your labor cost climbs three points, your menu price stays frozen because you review it once a year, and the contribution margin holding the operation together vanishes across two quarters without a single supplier invoice announcing it.

That is the blind spot of pricing on food cost: food cost warns you, payroll does not. Which is why price has to be recalculated against full prime cost —ingredient plus direct labor— and reviewed on a quarterly calendar, not whenever the accountant gets nervous. There is a tension most menus never resolve: the plate with the lowest food cost percentage is usually the worst business in the house. A dish carrying 24% cost that leaves 7,200 pesos of margin and blocks the grill for eight minutes at Friday peak yields 900 pesos per station minute. Another one at 28% cost that leaves 9,000 pesos and clears in four minutes yields 2,250 pesos per minute, two and a half times more, with a percentage your accountant would call worse.

The paradox of the cheap plate that bankrupts restaurants

For years I defended percentage as the control metric, and that is where I was wrong: it works for negotiating with suppliers, not for designing a menu. The bridge between both ideas is the unit of measure. Control percentage by ingredient family, but set PRICE on contribution pesos over your kitchen's bottleneck, which is almost always a station and almost never the register. Crestmont Capital, in its analysis of SBA loan default rates by industry published for 2026, found a regional variation of 8.7 percentage points in restaurant sector delinquency. Eight point seven points across regions, inside the same country, with the same credit instrument and in many cases with similar concepts. Demand does not explain that spread; what explains it is the price and cost discipline with which each operator defends margin in markets with different cost structures. A leveraged restaurant that adjusts its menu with a flat 8% increase punishes its anchor dishes, pushes guests toward lower-contribution references and arrives at the bank payment with less cash than it forecast.

The financial risk of bad pricing is not theoretical

Granular adjustment —three points here, twelve there, two dishes redesigned and one retired— produces the same incremental revenue with half the traffic damage. First: 14% of sales disappears into foodservice food surplus, per ReFED 2024. Action: measure waste by station for fourteen consecutive days, charge it to the recipe card of the plate that produces it and reprice on real cost, not theoretical cost. Second: 35% above February 2020 is where the United States food producer price index stands, per USDA ERS with BLS data through May 2026. Action: put a quarterly menu review on the calendar with a per-dish adjustment ceiling and stop waiting for the pain. Third: 37% of adults order delivery weekly, per UpMenu 2024. Action: split off the platform menu, price on margin net of commission and pull out everything that does not survive transport. Diego F. Parra brings these three into every Masterestaurant diagnostic because together they move contribution margin between 4 and 9 points in a single quarter.

The 3 numbers you should tattoo on yourself

Start this week with waste: it is the only one that requires no customer to accept anything. The unit of measure. Scenario A argues about food cost percentages; scenario B argues about pesos of contribution margin per dish sold. No restaurant deposits percentages at the bank. It deposits pesos, and that distinction — which sounds semantic — is what separates a 3% P&L from an 11% one. The decision horizon. Before, price gets reviewed when it hurts; after, it gets reviewed on the calendar. FAO data shows the food price index swinging up to 2,3 points month over month during 2025, and an operator who reprices annually absorbs eleven months of that drift out of pocket. The granularity of the move. An 8% bump across the whole card is a lazy decision that punishes the dish that can take it and the one that can't, identically. Quadrant-level adjustment raises where elasticity is low, holds where the guest anchors price perception, and removes what never belonged.

What genuinely separates the two scenarios?

How CapEx and OpEx get treated. In the miscalibrated operation, the remodel loan slips into dish cost and pushes the theoretical price out of the market.

In the calibrated one, CapEx amortises against structure and variable OpEx is the only thing touching the recipe. The conversation with the floor team. Once a server knows the risotto leaves 21.000 pesos and the pasta leaves 9.400, the recommendation stops being a script and becomes a margin lever pulled forty times a service. Reaction speed to capital leakage. With a live spec sheet, yield loss jumping from 4% to 9% on the beef cut shows on next week's board. Without it, it shows three months later, disguised as 'we sold less this month'.