Plate costing: the myths that eat your margin and the data that kills them

The dish he costed well and the channel he never costed

Costing a dish once is no longer enough: you have to cost it three times, once per channel where it sells. The home-style restaurant in Chapinero that closed last year had recipe cards for everything, gram by gram, and still ended the year with a 29.4% theoretical food cost and a 4.1% operating loss, because 46% of its sales came through a marketplace charging 28% commission, plus 1,400 pesos of packaging per order, plus a two-for-one promotion the owner accepted himself to rank higher in the listing. The industry reference number helps you place yourself: full-service food cost at operations with sales under 2 million dollars ran 33.7% in 2024, against 31.0% at those above 2 million, according to the National Restaurant Association's Restaurant Operations Data Abstract 2025. That starting point holds for the dining room. For the app it is worth nothing.

Why does a 30% food cost dish run at 58-64% on delivery?

Because commission, packaging and algorithmic discounting are variable cost of the dish, not administrative overhead.

Take a dish at 30% ingredient cost, add between 25% and 30% marketplace commission, then add packaging, which in mid-ticket Colombian operations runs around 1,400 pesos per order, and total variable cost settles between 58% and 64% of that sale. What remains, 36% in the best case, still has to cover kitchen payroll, rent, utilities and profit. It does not stretch. The National Restaurant Association puts the healthy food cost range at 28% to 35%, with a 32.4% median in limited service during 2024, and those ranges were built on structures where the channel did not take a third of the sale. Diego F. Parra keeps pushing at Masterestaurant an order that sounds obvious and almost nobody applies: cost the channel first, then set the dish price inside that channel. When contribution per order is negative, every new order buys you labor, not profit.

Volume does not fix a negative margin, it multiplies it

This is the arithmetic I ask of any owner who wants to enter an app: contribution in pesos per order, divided by the kitchen minutes that order consumes. An order leaving 1,600 pesos of contribution and occupying 6 minutes of line time yields 267 pesos per kitchen minute, while the same dish in the dining room, with no commission and no packaging, can yield four or five times that. With operator food spending at 34% of sales in 2024, per TouchBistro cited by Apicbase, and with food-away-from-home inflation at +4.1% during 2024 according to the USDA Economic Research Service, the cushion for absorbing an expensive channel got thinner. The question is not how many orders come in. It is how many pesos per minute each one leaves. Charging differently on the app and in the dining room is the correct practice, and the customer already assumes it.

Channel pricing is not deceiving the guest, it is surviving

I got this wrong for years, recommending price parity for the sake of transparency, until commissions moved from the 12-15% of 2018 to today's 25-30% and parity stopped being an ethical gesture and became a subsidy running from your dining room to the marketplace. The math is plain: if your dining-room dish sells at 32,000 pesos with 30% food cost, holding it at 32,000 inside an app that charges 28% leaves you a 12-18% contribution margin, when that same dish at the table leaves you above 45%. A channel surcharge of 15% to 20% across the delivery menu restores contribution without touching the perceived value of the guest eating seated, because these are two different experiences carrying two different cost structures. Raise it by category, not dish by dish. No benchmark applies the same way across three sizes of operation, and using one unadjusted does more damage than having none.

How to read these numbers in YOUR operation?

In a SMALL restaurant, below 2 million dollars in annual sales, the 33.7% food cost the National Restaurant Association reports for 2024 is your realistic ceiling, not your target:

purchasing power gives you nothing more, so the lever sits in channel mix, and under 35% of sales through delivery you can still breathe. In a MEDIUM operation, one billing above 2 million, the benchmark drops to 31.0% and commission negotiation genuinely comes into play, usually 3 to 5 points against committed volume. In a GROUP of several locations, the number that governs is not food cost but contribution per kitchen minute per channel, measured store by store, because one dish performs differently depending on delivery geography and how loaded the line runs. These numbers come from public United States sources and must be read knowing where they break.

Where these benchmarks come from and what they do not tell you

The National Restaurant Association's Restaurant Operations Data Abstract 2025 gathers income statements from United States operators, with a segmentation cut by sales volume, 2 million dollars, that does not translate directly to a Latin American market where the same threshold describes a far larger operation in units sold. Food-away-from-home inflation published by the USDA Economic Research Service, +4.1% in 2024 and +3.8% projected for 2025, measures a United States basket and works as direction, not as magnitude. And the packaging and delivery commission figures each operator builds from their own invoices, because no serious public benchmark for that exists in the region. Use the industry to place yourself. Use your own P&L to decide. You would lose 46% of sales and probably start making money, and that paradox has to be resolved before touching anything.

What would happen if you switched the marketplace off tomorrow?

Follow the Chapinero restaurant through: switch the app off, sales fall by half, but the 28% commission disappears along with the 1,400 pesos of packaging per order and the two-for-one;

with the kitchen line freed up, those same cooks turn more tables and the operation's effective food cost returns near the 29.4% theoretical figure. The trouble lies in fixed-cost absorption, since rent and base payroll did not move. The way out is not switching the channel off, it is splitting it in three, dining room, direct order through WhatsApp or your own site, and marketplace, then deliberately migrating guests from the third to the second, where commission falls to the payment gateway rate, between 2.5% and 3.5%. Every point that migrates is worth 24 to 26 points of contribution. Your recipe card needs three price columns and three contribution columns, one per channel, or you are not costing: you are taking inventory.

A recipe card that works carries three columns

The structure that works lists ingredient cost once, since that part does not change, and from there opens dining room, direct and marketplace, each loading its own commission, packaging and average discount as variable cost of the dish. With opening investment for an independent full-service restaurant in the United States running between 275,000 and 425,000 dollars according to Square, and with the SBA guaranteeing between 75% and 85% of those loans, the capital at stake demands that the number you look at weekly be the right one. Start this week with your ten best sellers, open them into three columns and sort them by contribution in the most expensive channel. The bottom three come off the delivery menu on Monday. MYTH: «my food cost is 30%, I am fine». REALITY: 30% is the input cost, not the cost of the dish sold. Add packaging and channel commission and that same dish runs at 58-64% variable cost on a marketplace, where nothing is left to pay payroll.

Five myths that destroy the most margin

MYTH: «I join the apps for volume, and volume fixes margin». REALITY: volume multiplies a negative margin. When contribution per order sits at $0.48 and preparing and packing it burns six minutes of kitchen time, every extra order buys you work, never profit. MYTH: «app prices must match the dining room out of transparency». REALITY: marketplaces allow differentiated pricing and guests already expect it; Christina Donnelly, VP of Restaurant Product at DoorDash, has argued publicly that operators adjusting menu prices in the digital channel protect profitability without measurable conversion loss. MYTH: «a five-star review has nothing to do with costing». REALITY: reputation is the variable that lowers acquisition cost. A Google Business Profile at 4.6 stars with 300 reviews draws up to 4.3 times more route requests than one at 3.9, and that organic traffic pays zero commission. MYTH: «costing is an annual exercise». REALITY: with 2.1% year-over-year food inflation in 2026 and commission contracts renegotiated quarterly, a twelve-month-old recipe card describes a restaurant that no longer exists.