Virtual brand profitability: what the dark kitchen promises and what the algorithm actually pays

The Polanco math: 1,180 monthly orders and 11,400 pesos of profit

Three virtual brands running out of one kitchen produced 11,400 pesos of monthly profit on 1,180 orders and a 289-peso average ticket, and the menu had nothing to do with that result. Gross sales came to roughly 341,000 pesos; the aggregator commission, between 27% and 30% across the major Latin American markets, took close to 95,000 before anyone touched the first kilo of protein. Set that against a fourth brand from the same owner, anchored to a location with a verified Google listing: 640 orders, a 312-peso ticket, and 38% of that revenue arriving through direct ordering at a 3.5% payment fee. Fewer orders, nearly double the profit. The anchored brand wins because it doesn't buy the same customer twice; the aggregator's tenant brand pays a toll every time it breathes. Applying the aggregator's 28% to the entire projected revenue is the most expensive miscalculation in any virtual-brand model, and it shows up in nine out of ten spreadsheets that reach my desk.

Commission isn't a percentage, it's a BLEND

A brand anchored to a location with an active listing, an indexed menu and a direct-order button moves between 25% and 40% of its volume outside the app, and that share pays 3-5% in payment processing. Run the number on 1,000 monthly orders at 300 pesos: full commission costs 84,000 pesos, blended commission at 32% direct sales costs 57,000. Twenty-seven thousand pesos of difference every month, which is precisely the profit the Polanco owner was missing. A virtual brand without a local engine isn't cheaper: it's the same kitchen paying full fare for every single customer. One additional star in a review rating moves revenue by 5% to 9%, according to Harvard Business School (Michael Luca, Reviews, Reputation, and Revenue), and that's where the two routes split brutally. Reviews your virtual brand accumulates inside the app belong to the aggregator: you can't export them, Google won't index them, and they do nothing for you the day you renegotiate commission or decide to walk.

Reviews: the asset the aggregator keeps and your own listing hands back

Reviews landing on a Google Business Profile with a verifiable address belong to you, show up on the map and feed the «near me» search. On 341,000 pesos of monthly sales, that 5-9% amounts to somewhere between 17,000 and 30,000 pesos the tenant brand simply never captures. Diego F. Parra treats this at Masterestaurant as a rule about equity: if the asset doesn't travel with you, it isn't yours. The aggregator rewards discounting and Google rewards complete information, which is why a single brand can't optimize well for both fronts with the same material. Inside the app, ranking climbs on aggressive promotions, short prep times and paid internal placement; every point of visibility gets bought with margin. In the search engine, ranking climbs on correct hours, real photos, a structured menu, replies to reviews and an address that actually exists. The first costs money every month and the second costs work exactly once.

Two algorithms rewarding opposite things

The anchored brand plays both boards and treats the aggregator as a cheap discovery channel rather than a landlord; the tenant brand only has the board that charges it. That's the operating verdict: whoever controls their own demand negotiates commission, and whoever doesn't control it accepts whatever arrives. The arithmetic promise of the virtual brand holds up: you already pay rent, gas, extraction and the hot-line shift, so every new brand stacked on that structure starts with fixed costs absorbed. Correct so far. What almost nobody subtracts is the variable cost of ACQUISITION, which in 2026 ate the entire difference between the two routes: commission, plus internal advertising when the app's organic ranking falls short, plus the discounts the algorithm demands before granting visibility. With inputs already 35% above 2019 levels in food and another 35% in labor, according to the National Restaurant Association, the margin cushion that made 28% tolerable has evaporated.

The covered fixed cost is real, the variable cost of acquisition isn't

A second brand is still a good idea; a second brand WITHOUT its own channel stopped being one. Push commission from 28% to 31% and follow the thread all the way down, because that scenario decides which of the two routes survives. The Polanco tenant brand, at 341,000 pesos of sales, loses an extra 10,230 pesos a month against 11,400 of profit: it lands at 1,170 pesos and the operation stops making economic sense. The anchored brand, with 38% direct sales, absorbs only 6,340 pesos of that hit on a profit base three times larger, and can push customers toward its own channel in the next campaign. It isn't that one is more profitable today; it's that one has a steering wheel and the other is a passenger. My read from the field is plain: a virtual brand without its own listing isn't a business, it's a lease on customers with a price the landlord can revise.

Personalization and loyalty: the margin that only exists with your own data

Personalized email messages lift open rates by 26%, according to Stripo, and 55% of restaurants report that their loyalty members' ticket grew faster than their menu prices, according to the Paytronix Loyalty Trends Report 2024. Neither lever is available when your brand lives inside the aggregator, because the customer's email, their frequency and their order history never reach you. With a local digital engine — listing, direct ordering, a contact base — every order leaves behind data you can sell to again without paying commission twice. On 1,000 monthly orders, recovering just 12% of those customers through your own channel means 120 orders at 300 pesos costing 3.5% instead of 28%: 8,800 clean pesos a month the tenant brand can't even attempt. If you have a physical location with a verifiable address, build the anchored virtual brand and don't debate it: complete Google listing, indexable menu, direct-order button and a routine for answering reviews, even if that takes six weeks before the first order lands.

What to choose for your profile?

If you operate from a hidden kitchen with no publishable address, the pure aggregator route works only as a 90-day proof of concept, with a written direct-sales target;

past that deadline without your own channel, shut the brand down or find it a commercial address. And if you're already carrying three brands that yield 11,400 pesos between them, kill two this week, concentrate the kitchen on whichever has the better food cost, and build the local engine on top. The anchored virtual brand wins. The tenant virtual brand rents out your clientele and charges you the rent in margin. Commission is not a percentage, it is a BLEND. This is the costliest miscalculation I keep finding in virtual brand projections: the owner takes the aggregator's 28% and applies it to all projected revenue. But a brand anchored to a kitchen with an active profile does not sell everything through the app.

The four differences that decide the math

With a solid Google Business Profile, an indexable menu and a direct order button, a quarter to two fifths of revenue arrives without the aggregator, and that slice pays 3-5% instead of 28%. On 1,000 monthly orders of 16 dollars, blended versus full commission is roughly 1,500 dollars a month — precisely the profit that owner was missing. The aggregator algorithm and the Google algorithm reward opposite things, and that tension is where most operators get stuck. Uber Eats, DoorDash and Rappi sort by immediate conversion: aggressive discounts, low prep time, high acceptance rate. Google Maps sorts by proximity, relevance and prominence, meaning reviews, profile age and information consistency. A menu engineered to win inside the app, with cut prices and perpetual promos, destroys the margin the direct channel protects. The fix is not picking one. Run TWO price lists: the aggregator list with the 27-30% already baked into the menu, and the direct list, cheaper for the guest and more profitable for you.

The four differences that decide the math — in practice

A 5-star review is worth different things in each channel. Inside the app, a 4.7 rating or better drives ranking and holds conversion; below 4.5 orders collapse without warning. Outside the app, a Maps review is a permanent asset that keeps working two years later, shows up in the three-result local pack and does not depend on you paying for visibility. According to Sherry Lauzon, who led foodservice research at Technomic for years, delivery brands that sustain verified reviews outside the app keep a slice of demand that does not swing with the promotion of the month. That is the whole point: the in-app review is rent, the Maps review is property. The asset you build is different, and that decides what your business is worth the day you sell it. A pure virtual brand has nothing transferable: no customer base, no positioning, no address, no searchable name.

The four differences that decide the math — key points

It is an account inside somebody else's platform. The virtual brand with a local engine accumulates a profile, reviews, a domain, a phone database and a measurable history of branded searches. When someone values a restaurant, that gap gets paid: an operation with its own demand is valued on an EBITDA multiple, an aggregator account is valued on used kitchen equipment.