Operating costs vs menu prices: the numbers before and after fixing the structure

Why do costs climb faster than the menu?

The gap is a calendar problem, not a market one: ingredients reprice weekly while the menu reprices every eighteen months, so the distance widens on its own even when nobody makes a bad call.

The producer price index for all food in the United States sat 35% above its February 2020 level as of May 2026, according to USDA ERS using BLS data, and the final-demand PPI closed 2025 at +3.0% after +3.5% in 2024. In Colombia, Acodrés reported a 9.8% increase in dishes and products during February 2025 alone. Run that arithmetic and a restaurant that leaves prices untouched for a year and a half hands over six to nine margin points without noticing. The concrete decision: put a menu review on the calendar every six months, crisis or no crisis. A venue billing 92,000 dollars a month at 68% prime cost faces arithmetic, not commerce: thirty-two points remain to cover rent, utilities, maintenance, insurance, technology and debt, and those lines eat between 22 and 28 points in most markets.

A 68% prime cost is not a sales problem

What survives is EBITDA, and it lives between 4% and 10% according to the National Restaurant Association Industry Report 2026. Do the subtraction: two careless points of food cost —2,000 dollars monthly at that volume— erase half the annual profit. Selling more against the same cost structure rarely fixes anything; it multiplies the problem and adds work. Before chasing new traffic, measure your prime cost over the last four weeks and set it against 60. That single figure decides whether your next move is marketing or a knife. I got this wrong for years: I recommended raising prices the moment food cost crossed 32%, as if that percentage measured health. It does not. Rent gets paid with money, never with percentages, and any well-read menu proves it: a dish at 38% food cost leaving 14 dollars of contribution and turning eighty times a week delivers 1,120 dollars weekly; another at 24% food cost, 5 dollars of contribution and twelve turns, delivers 60.

The thermometer is contribution in money, not food cost

The first one funds a shift's payroll, the second barely covers the walk-in's electricity. Technomic reported in 2024 that 46% of surveyed operators name alcohol among their highest-margin categories, and the same principle governs there. Sort your menu by total monthly contribution rather than percentage, and the priorities move to different places. Theoretical cost comes from the recipe card; actual cost comes from counted inventory, and the distance between them measures theft, waste, bad portioning and sloppy purchasing all at once. In uncontrolled operations that gap runs between 4 and 7 points of food sales: at a restaurant doing 92,000 dollars monthly with 30% coming from food, that means somewhere between 1,100 and 1,900 dollars vanishing every month with no invoice to explain it. Closing that gap is worth MORE than any price increase, because it costs zero customers and forces nobody to defend anything at the table.

The gap between theoretical and actual cost

Start with your five highest unit-cost items, weigh what goes in and out for two straight weeks, then compare against the card. If the difference clears two points, your process is broken, not your supplier. Adding 8% to the whole menu is the most expensive way to make money: it destroys traffic on elastic dishes —the ones guests compare and remember— and gives away margin on inelastic ones, where nobody would have noticed three extra dollars. Selective repricing touches between 15% and 25% of items, usually upward on high-turnover stars and downward on the ones carrying good margin with weak movement, to push them. With 60 dishes on the menu, that means changing nine to fifteen, not sixty. Delivery follows different math: UpMenu documented in 2024 that 37% of adults order delivery at least weekly and over 40% do so three to five times a month, so that channel carries its own pricing.

Across-the-board increases versus selective repricing

Never run the same menu in the dining room and in the app. Ranges shift with size, and applying them unadjusted produces bad decisions from good data. At a small venue up to 40,000 dollars monthly, healthy prime cost holds between 58 and 62 points because the owner works inside and absorbs payroll nobody invoices; there, repricing outweighs purchasing, and a semiannual menu review moves the needle. At a mid-size operation between 80,000 and 150,000, scheduling becomes the main lever: TimeForge documented labor cost reductions of 8% to 12% in 2025 with forecast accuracy above 90%, and on 90,000 dollars that means 2,200 to 3,300 monthly. In a group of three or more units, consolidated purchasing and a shared theoretical cost take over; there, one hard-won food cost point beats any promotion you can run. Be honest about the origin of these figures before moving them into your budget.

Where these benchmarks come from and what they miss?

Industry sales data and EBITDA ranges come from the National Restaurant Association's State of the Restaurant Industry, which projects roughly 1.55 trillion dollars in sales for 2026 based on surveys of U.S.

operators; the price indexes come from the Bureau of Labor Statistics and USDA ERS, which measure markets rather than your kitchen. Acodrés covers Colombia and UpMenu measures consumption habits, not profitability. None of those sources knows your rent, your mix or your installed capacity, which is why Diego F. Parra insists at Masterestaurant on treating benchmarks as an external thermometer and never as a target: the number governing your decision is the prime cost of your own last four weeks, measured with counted inventory. Carry the scenario to its end, which is where it stops being theory. Take the restaurant at 92,000 dollars monthly, 68% prime cost, 6% EBITDA: about 5,500 dollars of profit.

What happens if you touch nothing for another twelve months?

If food climbs another 3.0% annually —the final-demand PPI pace the BLS reported for 2025— and payroll rises 4%, prime cost reaches 70 or 71 points by month twelve without any price move, and profit drops below 3,000.

The paradox is that this owner will be selling the same volume, serving the same food and working the same hours for half the result. And should a line cook walk, StaffedUp puts replacement at 150% of salary. Open your menu today, rank items by monthly contribution in money and move the top ten. That is the entire plan. Theoretical cost comes from the recipe card; actual cost comes from inventory. The distance between them is the only figure that measures theft, waste, bad portioning and sloppy purchasing at once, and in uncontrolled operations that gap runs 4 to 7 points. Closing it beats any price increase, because it costs you no guests.

Five differences that move EBITDA

A flat 8% raise across the whole menu destroys traffic on elastic dishes and gives away margin on inelastic ones. Selective repricing touches 15% to 25% of the lineup, usually upward on the stars and downward on high-margin, low-turnover items. Food cost is a percentage; contribution margin is money. Rent gets paid with money, not with percentages, and that detail explains why so many menus optimized to 28% leave the operator short of cash by December. Loading labor and rent into the plate is the most widespread costing error in the industry. Those lines do not vary with the dish sold, they belong in the break-even, and folding them in produces inflated prices that scare guests without fixing the structure. Cadence matters as much as method. Quarterly reviews with a 1.5% threshold on key inputs produce 2 to 4% adjustments nobody notices; waiting eighteen months forces a 12% jump that guests do register and that does cost traffic.