How to make a restaurant profitable: the 2026 numbers that actually move margin: delivery and local SEO

Input costs are not coming back down: the number that reframes everything

The food you buy costs 35% more today than it did in February 2020, according to the all-food producer price index published by USDA ERS with BLS data as of May 2026, and that level is not falling.

Add to it that the producer price index for services closed 2025 at +3.2% against +2.5% for goods (U.S. BLS, Producer Price Index 2025 M12): what got expensive faster was the labor that transforms the input, not the input itself. In Colombia the picture looks harsher still, because ACODRÉS measured a 9.8% rise in menu prices since February 2025 and the trade group still called it a crisis.

Translated into cash: raising prices stopped being a margin lever and became a defense. Anyone still costing recipes with 2022 prices is handing away six to nine points of contribution per dish without noticing.

Prime cost under control: the real ceiling, not the pretty target

The case that opened this piece shows it without mercy: food cost of 29.4% and payroll of 27% add up to 56.4%, well inside range, and the place still lost 4.1 million pesos in March. That result is not a contradiction, it is proof that food cost at 32% per dish is a CEILING and never a target: squeezing portions below 29% does not buy profitability, it buys small-portion reviews.

With food 35% above 2020 levels (USDA ERS/BLS 2026) and services climbing 3.2% a year (BLS 2025), the break-even point moves on its own. The mini-conclusion these figures trigger: if your prime cost is in range and you are still losing money, leave the kitchen alone and go check where the sales come in from.

Where the sale is decided: delivery, commission and the expensive cover?

Some 37% of adults order delivery at least once a week and more than 40% do it three to five times a month, according to UpMenu (Food Delivery Statistics 2024).

That demand is not optional, it is the new split. The trouble shows up when that share arrives through a channel charging 26% commission, as in the Bogotá restaurant serving 180 covers a day: with 38% of sales flowing through two apps, the place hands over nearly ten points of total billing before touching a gram of product. Run the counterfactual all the way.

If that same owner shifts just a third of that volume to direct ordering —updated Google listing, own menu, WhatsApp answered within minutes— he recovers roughly 3.3 points of margin on total sales, which in his March numbers turn a 4.1-million loss into a positive result without changing a single recipe. I would rather fight that battle than the one over portion weights.

The local listing: fourteen months without photos cost more than a food cost point

A Google Business Profile with 3.6 stars across 211 reviews and fourteen months without a fresh photo is not a marketing oversight, it is a measurable cash leak. Diego F. Parra presses a point that makes plenty of owners uncomfortable: local ranking is a number you manage exactly like food cost, with data, with frequency, and with somebody accountable for it. At Masterestaurant we treat it that literally.

When 37% of customers already order food through a digital channel every week (UpMenu 2024), your position on a map hands out covers the way the street corner used to. And if you want the cheapest channel available to reactivate that base, look at email: average open rate was 25.1% in 2023 according to Omnisend (Email, SMS & push marketing report 2024), with no per-transaction commission. An owned channel at 0% against a rented one at 26%: there is your decision.

Payroll: 8-12% of savings sits in the shift, not in the layoff

Scheduling shifts with AI demand forecasting cuts labor costs by 8% to 12% with forecast accuracy above 90%, according to TimeForge (2025). On a payroll running at 27% of sales, that range means trimming 2.2 to 3.2 points of prime cost without touching anyone's wage, which is precisely the opposite of what a frightened owner does.

The other side of the same coin weighs even more: every departure you prevent saves 150% of that person's salary in replacement costs, according to StaffedUp (Restaurant Professional Development 2025). There is raw material to work with, because the young workforce came back: 6.2 million Americans aged 16 to 19 are working, 900,000 more than in 2019 (National Restaurant Association/BLS 2024). The decision these three numbers trigger together is a single one: before you cut people, cut badly placed hours.

The menu rules: alcohol, coffee and the star-dish paradox

Alcohol is named among the highest-margin menu categories by 46% of surveyed operators, according to Technomic as reported by Nation's Restaurant News (2024). On the opposite side, arabica coffee jumped 70% during 2024 (Bellwether Coffee), so the cup that used to hold the table after dessert stopped being comfortable margin.

Here sits the tension no owner wants to resolve: the best-selling dish is almost never the one contributing most, and yet it is the one defended whenever somebody suggests touching it. You settle that with arithmetic, not affection.

Rank dishes by contribution in pesos per unit sold rather than by food cost percentage, push the top two on the menu and in the service script, and raise the price of the bottom two until the math justifies them or they go. With food 35% above 2020 (USDA ERS/BLS 2026), keeping a dish out of nostalgia is a cash decision, and you make it every month.

The clock: why measuring every seven days is worth millions

The deep difference between a profitable restaurant and one bleeding out is not the tool, it is how often it closes its numbers. An operation that calculates prime cost every seven days catches a supplier leak when 1.4 million pesos are gone; the one reading it in the month-end balance catches it with six million on top. The same error, measured four times later, costs four times more, and that arithmetic allows no argument.

With the food price index 35% above 2020 levels (USDA ERS/BLS 2026) and suppliers revising lists every few weeks, a quarterly costing describes a restaurant that no longer exists. At Masterestaurant the weekly close is the first deliverable of any engagement, before we touch the menu or the org chart, because without that clock every other decision gets made blind. The mini-conclusion: frequency is the cheapest lever in the business.

The 3 numbers you should tattoo on yourself

Three numbers, three concrete actions for this week. Second: the share of sales leaving through commission apps. Action: measure today what that channel is worth —if it tops 25% of your billing, like the 38% in the Bogotá case, every point you move to an owned channel returns clean margin, and email opens at 25.1% charging no commission (Omnisend 2024). Third: 150% of salary, which is what replacing each person who walks out costs you (StaffedUp 2025). Action: name a retention owner today with that figure written into the target.

The differences that change the month's number

The underlying difference is not the tool, it is the CLOCK. A restaurant closing prime cost every seven days catches a supplier leak at 1.4 million pesos lost; the one waiting for month-end catches it six million deep. The arithmetic is blunt and does not negotiate: the same mistake, measured four times later, costs four times more. The traditional method treats revenue as something that arrives and cost as something you control.

That asymmetry made sense when demand depended on the corner you stood on. Today demand is allocated by a ranking, and a ranking is a number you manage exactly like food cost: with data, with cadence, with somebody accountable for it. Diego F. Parra keeps pressing a point that makes owners uncomfortable: raising menu prices is almost always cheaper than lowering food cost.

A sustained 4% lift in average ticket moves operating margin further than three months of fighting your protein supplier, and it does not degrade the plate.

The differences that change the month's number — in practice

It only works, though, when the listing, the photos and the reviews hold that price up; a 3.4-star restaurant cannot raise prices without bleeding traffic. The real tension of the trade in 2026 reads like this: delivery brings volume and destroys margin at the same time. Solving it is not about picking a side, it is about splitting the books.

Masterestaurant runs two parallel P&Ls, one for the dining room and one for digital channels, because blending them hides the fact that the dining room is subsidizing a delivery operation losing 4% per order. And one difference in judgment outweighs the rest: the traditional owner asks how much did I sell; the MR method asks what did each cover cost me and which door did it come through. Only the second question has an answer you can act on the same day.