Virtual brand profitability: the numbers before and after measuring each brand on its own

Why the platform payout hides the brand that loses money?

The consolidated payout hides your money-losing brand because the platform settles per store while you run four digital storefronts inside that same store.

An operator in Chapinero closed July with 41,200 USD billed across Rappi, iFood and Uber Eats, the deposit matched the bank to the cent, and the cash still would not add up: once we opened the report brand by brand, two of the four returned less than they cost to produce. Averaging works like anesthesia here. When restaurant commission plans offered by the apps run from 15% to 30% depending on the tier contracted (CloudKitchens, 2024), averaging three separate agreements at 28% gifts margin to the expensive brand and takes it from the cheap one. Start with the simple cut: pull last week's settlement and separate the actual commission charged to each digital storefront. Assign packaging to each brand before you argue about any other line, because it is the cost that varies most between storefronts sharing a kitchen.

Packaging is the cost nobody assigns and the one that swings most between brands

A burger in a kraft box consumes 0.42 USD of packaging; a bowl with sauce on the side and a sealed lid climbs to 1.10 USD. On an average 13 USD ticket, that 0.68 USD gap is worth 5.2 margin points, more than most operators earn in a year negotiating protein suppliers. The bowl brand usually shows the prettier food cost on paper too, so it looks like the winner until somebody books packaging where it belongs. Concrete decision: if your bowl brand cannot hold 5 extra points of gross margin against the burger brand, packaging is eating it, and you either redesign the container or raise the price. Inside the app you are not buying reach, you are buying POSITION within a three-to-five kilometer radius, competing against restaurants with their own kitchen, lower imputed rent and a food cost below yours. The same dollar performs differently depending on which brand spends it.

In-app geotargeted advertising behaves nothing like Meta advertising

This matters because 58% of customers prefer ordering through the restaurant's own app or website (NCR Voyix, 2024): you are paying for placement customers do not even prefer. My read is blunt — in-app advertising buys borrowed volume, never brand. Charge the week's ad spend to the brand that consumed it, divide by that brand's orders, and if acquisition cost per order exceeds that brand's contribution margin, switch the campaign off on Monday and do not turn it back on out of fear of the ranking. Read these benchmarks against your size, because the same figure means different things across three different operations. Small operation, one kitchen and two brands billing under 15,000 USD a month: skip analytical accounting, just enforce the packaging cut and the real commission per brand — two properly assigned lines already show you which one bleeds. Mid-size operation, three or four brands at 35,000 to 60,000 USD monthly, which is the Chapinero case: here you do need a per-brand P&L with commission, absorbed coupons, advertising and packaging, reviewed every fourteen days.

How to read these numbers in YOUR operation (three scenarios)?

Multi-kitchen group:

consolidate by brand ACROSS locations, because one brand can perform in one radius and sink in another due to local competition, and the group average will lie to you exactly the way the platform payout lied to that single-kitchen operator. Sector projections explain why the per-brand cut is worth the pain: the global dark kitchen market is projected at USD 171.30 billion by 2033 (Global Growth Insights), cloud kitchens at USD 203.72 billion by 2033 (Grand View Research), and online delivery in Mexico alone points to USD 18.27 billion by 2029 (Statista, 2024). Diego F. Parra keeps insisting at Masterestaurant that this growth pulls in new operators every month, and the practical consequence is that your three-kilometer radius will be more contested next year than it is today. A brand leaving 4 margin points today leaves zero once two more competitors advertise in that same radius.

The size of the market justifies the accounting work nobody wants to do

So the per-brand cut is not an accountant's exercise: it is the instrument you use to decide which storefront you defend and which one you close, before the market decides for you. Suppose you assign costs badly and kill the wrong brand. The bowl brand shows 2 margin points because you charged it the full packaging load plus the whole month's advertising, when that advertising brought orders the burger brand closed through cross-selling; you shut it down, free up 30% of kitchen production, and the following month discover burgers fell 18% because they lost the traffic the other storefront generated in the radius. Now you have idle capacity, one brand fewer and lower total sales. The real tension in this model is that each brand gets measured alone while operating inside a shared kitchen. Resolve it by assigning what is DIRECT per brand — commission, packaging, food cost, its own advertising — and leaving labor and rent in the location's break-even, never split by a rule of three.

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

Be honest about where the figures come from before you decide on them.

The 15% to 30% commission range comes from the CloudKitchens delivery app fee report (2024) and describes published United States plans, not the particular deal you negotiated in Bogota or Mexico City. The 58% preference for direct ordering comes from NCR Voyix (2024) and measures stated intent, which is not behavior measured in your radius. Market projections — dark kitchens at USD 171.30 billion in 2033 per Global Growth Insights, Spanish ghost kitchens at USD 1.379 billion in 2032 per Expert Market Research — are growth models, not promises. The packaging costs, tickets and decline percentages in this piece come from the operation described, not from a statistical sample. Use them as an order-of-magnitude reference and replace them with YOUR numbers as soon as you have two properly assigned closings. Commission is never one number.

Where the math actually breaks?

An operator with three brands usually holds three different agreements, because each one signed in a different month with a different launch promotion; averaging them at 28 % gifts margin to the expensive brand and steals it from the cheap one.

Packaging is the cost nobody charges properly and the one that swings hardest between brands: a burger in a kraft box costs 0.42 USD to pack, a bowl with sauce on the side and a sealed lid climbs to 1.10 USD. On a 13 USD ticket that gap eats close to five points of margin. Geolocated in-app spend does not behave like Meta advertising: you bid for position inside a three-to-five-kilometre radius against restaurants that may run their own kitchen and a lower food cost in that same radius, so the same advertising dollar performs differently per brand and per neighbourhood.

Where the math actually breaks — in practice?

The platform algorithm rewards acceptance and prep time and punishes cancellation; a virtual brand sharing a kitchen with three others inherits the kitchen's timings rather than its own, and that is where one good brand drags the rest down the ranking.

A physical restaurant amortises its dining room, a virtual brand has no dining room to amortise but pays a platform toll on every order; comparing dark kitchen vs physical restaurant on the same gross-margin line compares two businesses that do not even collect the same way.