Own delivery vs app commission: myth vs reality

1. App commission or your own fleet? The answer hinges on your order density

App commission wins below a certain volume threshold; in-house delivery only pulls ahead once order density dilutes your fixed cost. The mistake I see over and over is comparing a «30% commission» against «gas and one driver» as if they were the same P&L line: they are not. Platform commission is a pure variable cost that bites EVERY order, while a salaried driver is a fixed cost that exists whether you run 20 or 200 deliveries a day. With a sector net margin of just 3–9% per Statista, that distinction decides whether the delivery channel adds positive or negative EBITDA. Food-away-from-home inflation rose +3.8% in 2025 per USDA Economic Research Service (2025), compressing that cushion further. Deciding on instinct, without computing your volume cross-over point, is the fast lane to subsidizing orders you thought were profitable. Platform commission is a variable tax that scales one-to-one with your channel sales, and in-house delivery flips that equation: capex and fixed payroll that only amortize with volume.

2. The app's variable cost vs. the fixed cost of your fleet

If you pay 30% to the app, every 40 USD ticket loses 12 USD before touching your food cost —which should never exceed 32% of the plate— or your prime cost. A driver, by contrast, costs the same whether you run 30 or 90 orders: below that floor, your fixed fleet bleeds you worse than any commission. With sector pre-tax operating margins averaging 10.66% per NYU Stern (Damodaran, 2024), the channel forgives no accounting errors. Diego F. Parra and the Masterestaurant framework demand a managerial P&L per channel: dine-in, take-away, in-house delivery and aggregators, each with its own food cost, prime cost and contribution margin isolated. That's the only way to see which channel funds you and which bleeds you. Platforms and in-house delivery don't compete for the same thing: one acquires customers you didn't have, the other retains and monetizes the ones you do.

3. Apps sell new demand; your own channel protects the margin of the demand you already have

An aggregator puts you in front of a diner who'd never have found your brand; you pay 25–30% commission as the acquisition cost of that incremental demand. Your own fleet, instead, defends the margin of demand that was already yours: the customer who was going to order from you anyway shouldn't cost a 30% toll every time. With a typical restaurant EBITDA margin between 12% and 30% per WhippleWood CPAs (Restaurant Financial Benchmarks 2026), handing commission on your captive base is torching EBITDA points needlessly. The consultant's move is to segment: let the app hunt cold customers and migrate the repeat buyer to your direct channel. Confusing acquisition with retention makes you pay commission for people you'd already won. With in-house delivery you own the customer data and can reactivate them without paying a toll; on the apps the customer belongs to the platform and every repurchase costs commission again.

4. Who owns the customer data and why that's worth money

That asymmetry is strategic, not cosmetic. When the order runs through your channel, you capture phone, address and frequency, and can reactivate by WhatsApp at near-zero cost; when it runs through the app, every second order from the same diner pays 25–30% again. In a ghost-kitchen market worth 72.06 billion USD in 2024 per Credence Research (2024), whoever controls the data controls the repurchase —and repurchase is where margin lives—. With an average profit margin of 9.8% in 2024 per TouchBistro (via Apicbase), reactivating your own customer commission-free can double that order's profitability versus the same ticket via aggregator. The app rents you the storefront; your own data is the asset that actually capitalizes your business. In-house delivery carries capex and per-location operating risk; apps scale with no investment but create dependence on a third party that controls your digital storefront.

5. Capex, territory risk and dependence on a third party

Building a fleet means vehicles, insurance, shift management and territory risk: a low-density zone won't generate enough orders to amortize that fixed structure. Consider that opening an independent full-service restaurant already costs 275,000–425,000 USD per Square (2024); adding your own delivery before you have volume is leveraging a channel that doesn't yet pay off. Apps avoid that capex but turn you into a hostage: if the algorithm changes your ranking or raises the commission, your demand evaporates without warning. In 2025 at least 8 restaurant brands filed Chapter 11 in the U.S. per Restaurant Business, and On The Border closed 40 of ~120 stores: structural fragility kills. The decision isn't «own vs. app», it's how much strategic dependence you tolerate. In-house delivery starts to win once your daily volume passes the point where fixed cost per order drops below the app's variable commission.

6. The cross-over point: when in-house delivery starts to win

It's boardroom arithmetic, not ideology. If your fixed fleet costs, say, 200 USD a day and you run 20 orders, each delivery costs 10 USD; at 40 orders it falls to 5 USD; at 80, to 2.50 USD. Compare that against the 25–30% the app charges per ticket and you'll see your exact threshold. Diego F. Parra insists with the Masterestaurant framework that this cross-over is computed per location and per time slot, not globally. With fast-casual valuation multiples of 4x–7x EBITDA per Sofer Advisors, every EBITDA point you rescue from the delivery channel multiplies when you sell the business. The operating rule: use apps until you cross the density threshold, and only then build a fleet. Do it backwards and you break the channel before scaling it. App commission is a pure variable cost that hits EVERY order; in-house delivery is a fixed cost that only dilutes with volume.

7. The differences owners confuse

Comparing the percentage against fuel is a channel-accounting error. Apps sell DEMAND you didn't have; your own channel protects the MARGIN on demand you already had. They don't compete for the same thing: one acquires new customers, the other retains and monetizes yours. With in-house delivery you own the customer data and can reactivate them toll-free; on apps the customer belongs to the platform and every repurchase costs commission again. In-house delivery carries capex and operational risk per unit (territory risk); apps scale without capex but create strategic dependency on a third party that controls your digital storefront.