Hybrid dine-in + delivery model: the numbers the traditional method never separates

Why does a profitable dining-room dish lose money on the app?

Because the platform commission eats your contribution margin before you touch a single fixed cost, and almost nobody measures it channel by channel. Take a dish priced at 48,000 pesos with a 31% recipe food cost:

in the dining room it leaves roughly 33,000 pesos of gross margin, enough to absorb its share of rent and payroll. That same dish, sent out through an app charging 27% commission, with 1,800 pesos of packaging and a portion the kitchen enlarges to offset heat loss, drops below 19,000 pesos of margin. The gap never shows up in any report because the average P&L adds total sales against total cost. When we opened the income statement by channel at Masterestaurant, the Bogotá pizzeria that billed 412 million in one semester found out its 48,000-peso delivery order left 1,400 pesos before rent. Treat every commission point as paid marketing and the decision turns into simple arithmetic.

A 27% commission is customer acquisition cost, not a sales channel

A restaurant selling 38% of its revenue through two platforms at 26% and 29% is handing over close to 10.5% of gross sales in acquisition, every month, forever, without keeping the customer's email, phone or purchase frequency. On 412 million per semester that adds up to 43 million pesos: budget enough for a direct-order program with its own payment gateway, outsourced couriers at a fixed fee per drop, and a database that actually belongs to you. The National Restaurant Association reported that in 2025, 74% of operators expected ordering technology to keep gaining weight in their mix. Gaining weight is not gaining margin. It means whoever fails to control the direct channel will pay that CAC indefinitely. A dish costed at 29% on the spec sheet lands between 34% and 38% once it goes out for delivery, and that gap hides in inventory instead of on an invoice.

Recipe food cost and travel food cost are two different numbers

Four items explain it: packaging, rarely charged to the recipe and worth 3 to 5 points; the sauce that ships separately and doubles portions; the gram weight the kitchen quietly raises so the plate survives twenty minutes on a motorbike; and the 2% to 4% of orders remade after a complaint, a cost paid twice —product lost plus product replaced— and once more in your app rating. If your food cost ceiling is 32%, that delivery order is already out of range before commission. Load packaging and shrinkage onto a separate delivery spec sheet and you will see which half of your menu should never travel. Dining room and delivery use the same kitchen in different time bands, and that is the only real advantage the hybrid model offers. The dining room peaks between 12:30 and 2:00 p.m., then again from 7:30 to 9:30 p.m.; delivery stretches to 11:00 p.m.

Break-even runs by the hour, not by the month

and holds up Sunday afternoon, when the room sits at 20% occupancy. An 80-seat restaurant running two shifts pays the same hot-line payroll, the same rent and the same refrigeration across fourteen hours, so every order landing in a valley only has to cover its variable cost —real travel food cost plus commission— to contribute margin. That same order at the 1:00 p.m. peak, tying up a griddle a dine-in guest would pay for at 68% margin, destroys value. The rule I apply: delivery open in the valleys, closed or differentially priced at the peaks. The channel threshold shifts with size, and using the wrong benchmark beats having none only in theory. In a small restaurant, one to three million pesos of monthly sales with no backup kitchen, delivery should stay under 20% of the mix: past that you are paying commission with the same grill the dining room needs.

How to read these numbers in YOUR operation: three scenarios?

In a mid-size operation running two full shifts and a dedicated packing station, the healthy range sits between 25% and 35% of the mix, and that is where negotiating volume rates pays off.

A group with three or more locations has the only structural way out: its own direct channel with outsourced logistics at a flat fee, because cost per drop stops being a percentage and becomes an absolute figure, between 4,000 and 9,000 pesos depending on the city. Work out your contribution margin per channel before picking a scenario. Raise platform prices 20% and you will recover the dish margin, though two consequences come with it and both deserve a straight answer. Volume falls first; in the operations where I have set this up, the drop settles between 12% and 18% of orders, which still leaves better cash because the surviving orders actually cover their full cost.

What happens if you raise app prices to offset the commission?

The second consequence cuts deeper:

if a guest sees 48,000 pesos on the app and 40,000 on your printed menu, you are training your own customer to order direct, exactly what you want, while handing the platform a reason to punish your ranking visibility. I got this wrong for years, recommending single pricing for brand consistency. Consistency does not pay rent. Differentiated pricing, framed as a service cost rather than a penalty, does. The commission, packaging and shrinkage ranges in this piece come from three separate sources, and knowing which is which matters. Platform fees and packaging costs are market values observable in operating contracts and invoices across Colombia and Mexico during 2025 and 2026; they move by city and by bargaining power, so read them as a range rather than a constant. The sector-structure figures come from public bodies: CANIRAC-INEGI counts more than 680,000 restaurants and 2.57 million economic units in Mexico in 2025, and the U.S.

Where these benchmarks come from and what they cannot tell you?

Bureau of Labor Statistics documents that only 51.4% of restaurants are still trading at five years and 34.6% at ten. What no benchmark can give you is your own contribution margin per channel.

That figure exists only once you split the P&L, and until you split it you are guessing. Split the income statement into two columns —dining room and third parties— and run it over the last ninety days of sales before touching anything else. Four lines per column will do: net sales, real food cost with packaging loaded in, commission, and shrinkage from complaints. In less than an afternoon you will know whether delivery adds or subtracts contribution margin, and in 70% of the cases I review the answer surprises the owner who swore the problem was payroll.

The decision that closes the month: what to move on Monday

If the third-party channel contributes under 15 points of contribution margin on net sales, you hold three levers and none of them is shutting down: raise platform prices, pull the dishes that do not travel off the digital menu, or start building direct ordering with the same money you hand over in commission today. Diego F. Parra puts it plainly in Masterestaurant audits: a channel you do not measure is not managed, it is subsidized. Customer acquisition cost hides inside the commission. An app charging 27% is not a sales channel; it is a marketing channel with a permanent 27% CAC on every transaction, and you never get the customer data. That same 27% spent on a direct channel buys a database, repeat orders and price control. Recipe food cost and travel food cost are two different numbers. Between packaging, sauce shipped separately, portions enlarged to compensate for heat loss and the 2-4% of orders remade after a complaint, a dish costed at 29% lands somewhere between 34% and 38% in its delivery version.

Four differences that move the cash

Nobody sees it because the shrink lives in inventory, not on an invoice. Break-even moves by hour, not by month. The dining room peaks 12:30-14:00 and 19:30-21:30; delivery stretches to 23:00 on a kitchen whose payroll is already paid. Read properly, that stretch is the best news in the hybrid model: incremental sales on absorbed fixed cost, provided the channel's contribution margin beats the variable cost of keeping the line open. Restaurant financial maturity is measured by how many channels you can switch off and survive. An operator with 55% of sales on two apps does not run a hybrid model; he runs a dependency with tables. The day a platform lifts commission three points or reshuffles its visibility algorithm, that decision never reaches your board.