Masterestaurant Dark Kitchen Index LATAM 2026: a real 6.8% net margin on pure delivery

Finding 1 — The verdict after auditing 214 delivery operations

A dark kitchen doesn't beat a traditional restaurant on margin: it delivers a different and more fragile margin. Between 2023 and 2026 we audited 214 operations at Masterestaurant across Mexico, Colombia, Chile, Peru, Argentina and Ecuador, and the ghost kitchen returned an average net of 6.8% (range 3.1%-11.4% by segment) versus 9.2% for full service with a dining room. That's 2.4 points less, and the culprit has a name: the aggregator's effective commission, which averaged 27.9% of the ticket before taxes in our sample. The owner who launches a dark kitchen believing it slashes fixed costs has merely swapped a fixed rent for a variable rent tied to every single sale. That swap scales worse than expected when volume grows without an owned channel. Diego F. Parra puts it plainly: don't pick the model by the margin snapshot, pick it by which of the three metrics almost nobody measures you can actually control.

Finding 2 — Post-discount average ticket: the 39% gap nobody recovers

The real ticket after coupons and promos is the number that decides viability, and in pure delivery it drops 39% versus the dining room. In our sample we measured $14.20 USD per ticket at the table against just $8.60 USD in delivery, after discounting the coupons and promotions the aggregator imposes to win visibility. No rent or waiter savings offset that gap. The rule we apply at Masterestaurant is blunt: any dark kitchen with a ticket below $9.80 USD enters the red zone, because the 27.9% commission leaves too little per order to cover food cost plus packaging. With food cost at 30% and packaging at $0.70 USD per order, an $8.60 ticket yields a contribution margin that evaporates in the first slow month. The dining room, with a higher ticket and high-margin drinks, absorbs the hit; the ghost kitchen cannot. The share of orders through your owned channel is the variable that moves a dark kitchen's net the most, full stop.

Finding 3 — The owned channel: the lever separating 4.1% from 10.9%

It is literally the difference between paying a 27.9% commission or paying zero. In the sample, ghost kitchens that pushed more than 40% of orders through their own channel —WhatsApp with payment, web checkout, proprietary app— lifted their net from the 6.8% average to 10.9%. Those that stayed 100% tied to the aggregator averaged 4.1% net, on the verge of destroying cash in any low-demand month. The figure is decisive: every point of sales you migrate from the aggregator to your own channel is worth more than any reasonable food cost cut. The mistake I see over and over is building the kitchen and outsourcing all demand to Rappi or Uber Eats, when that demand is rented and expensive. Without an owned channel, the dark kitchen isn't a business, it's a supplier to the aggregators. The commission that matters isn't the rate you sign (18%-30%), but the effective one after adding advertising, forced promotions and ranking penalties.

Finding 4 — Effective commission is not the nominal rate

Across our 214 audits, the nominal rate averaged 24% but the effective one reached 27.9% of the ticket, nearly four points the owner never sees in the contract. That gap comes from the campaigns the aggregator "suggests" to hold your position in the app, the discounts the restaurant finances, and the delivery charges in far zones. At Masterestaurant we always measure the effective commission by dividing what the aggregator retains by the channel's gross sales, month by month: it's the only honest number. One operator thought he paid 22% and actually paid 29.4%. On $40,000 USD in monthly volume, those 7.4 hidden points are $2,960 USD vanishing with no explanation. If you don't audit the effective rate, your P&L lies. A dark kitchen does cut rent and eliminate the dining room, but it shifts the cost to a pricier, less controllable place.

Finding 5 — Fixed costs: the illusion of variable rent

A ghost kitchen shrinks operating space by 60%-70% versus a venue with a dining room, and it removes waiters, hostess and much of the service; in our sample that dropped fixed costs from 34% of sales to around 19%. The problem is that the variable commission eats that saving. A traditional restaurant pays a fixed rent that, at high volume, dilutes and falls as a percentage of sales. The dark kitchen pays a variable rent —the 27.9% commission— that does NOT dilute: every new order pays the same. That's why full service scales better once volume arrives: its fixed costs spread out. The ghost kitchen scales worse without an owned channel, because its biggest cost grows one for one with sales. The opening snapshot deceives; the twelve-month film is what rules. A dark kitchen wins only if you can guarantee an owned channel and a sufficient ticket; otherwise the dining room takes it.

Finding 6 — When each model wins: the real decision tree

With data from the 214 operations, the tree is clear: if your delivery ticket exceeds $9.80 USD and you can push the owned channel above 40%, the ghost kitchen returns up to 10.9% net with less capital tied up and a launch in weeks, not months. If your ticket lives below $9.00 USD or you depend 100% on the aggregator, full service with a dining room —at 9.2% net— is more stable and survives the slow months. A hybrid model, a dining room launching a second virtual brand from the same kitchen, captured the best of both in 31 cases: it added delivery revenue with no new rent and shared food cost. The Masterestaurant method recommends no universal model; it recommends measuring your three metrics —post-discount ticket, owned channel and effective commission— before signing any aggregator contract. Every dark kitchen must build an owned channel from day one or it stays trapped at 4.1% net.

Finding 7 — The cash plan to avoid going broke in month six

The sequence we run with our clients is concrete: in quarter one, capture each aggregator customer's WhatsApp with a repurchase incentive; in quarter two, launch a web store with its own payment and offer a 10% discount that's still cheaper than the 27.9% commission; in quarter three, migrate at least 40% of volume to those channels. On $40,000 USD monthly, moving that 40% off the aggregator frees roughly $4,464 USD a month previously lost to commission. That money is the difference between a 6.8% net and a 10.9% one. Diego F. Parra's hard cash rule: never let the aggregator exceed 60% of your sales, because beyond that you don't have a business, you have a customer who can raise your commission whenever it wants. Post-discount average ticket. In the dining room we measured $14.20 USD; in pure delivery, $8.60 USD after coupons and aggregator promos.

Finding 8 — The three numbers that decide whether a dark kitchen is worth it

That 39% gap is not recovered by any rent saving. A dark kitchen with a ticket below $9.80 USD enters the red zone: the 27.9% commission leaves too little per order to cover food cost plus packaging. Share of orders via owned channel. It is the difference between paying 27.9% commission or paying zero. Dark kitchens that cleared 40% owned channel (WhatsApp, web with payment, own app) lifted their net from 6.8% to 10.9% in the sample. Those stuck at 100% aggregator averaged 4.1% net, on the edge of destroying cash in any soft month. Effective aggregator commission. It is not the nominal rate (18%-30%): it is the effective one after in-app advertising, visibility fees and co-funded discounts. We measured 27.9% real against a 22% declared nominal. That 5.9-point gap is money leaving your margin without showing up on any obvious invoice.