Own delivery vs app commissions: myth vs financial reality

1. Which channel yields the most contribution margin per order, not the lowest fee?

The right decision is not choosing the cheapest channel but the one that leaves the most contribution margin per order after subtracting ALL real cost.

Apps charge 15% to 30% of the ticket; on a $250 MXN order, a 25% fee takes $62.50 before you touch profit. Diego F. Parra and the Masterestaurant methodology treat each channel as a business unit with its own food cost, logistics cost and margin. The mistake seen again and again is comparing the contract fee against an idealized in-house delivery. With the sector's target prime cost at 55-65% of sales (Toast), a channel that raises labor without raising the ticket breaks that goal. The National Restaurant Association reports 98% of operators saw labor costs rise in 2024: building your own delivery adds exactly that pressure to the wrong side of the income statement. App cost is 100% variable —no sale, no charge—; in-house delivery mixes high CapEx with fixed payroll that runs even if not a single order comes in.

2. App cost is pure OpEx; in-house mixes CapEx with fixed payroll

A scooter, its insurance and the delivery gear are sunk investment before the first dispatch; the average restaurant insurance premium runs about $1,359 a year, roughly $113 a month (MoneyGeek, 2025), and that is just one line. With the restaurant industry representing 12.2% of Mexico's economic units (INEGI-CANIRAC, 2024), most are small operations with no cushion to absorb a slow month of courier payroll. The financial rule is simple: if your delivery volume is irregular, the app's variable cost protects your cash in bad weeks; if your volume is high and steady, fixed payroll dilutes per order and starts beating the commission. The crossover point is set by your order count, not by emotion. Apps and in-house delivery chase different financial goals, they are not substitutes: apps buy acquisition and reach, the owned channel buys repeat business and ownership of customer data. Paying a 25% fee on a new customer who never returns is expensive acquisition; paying that same 25% on a customer who already buys weekly is giving away margin you could capture yourself.

3. Apps buy acquisition; the owned channel buys repeat business and the data

With Mexico's lodging and food-service GDP at $838,530 million MXN in Q3 2025, growing 4.85% year over year (Data México, Secretaría de Economía), the market rewards those who control their customer base, not those who rent it. The Masterestaurant consultant read is to treat app fees as customer-acquisition marketing budget: acceptable for new customers, indefensible for repeat orders. The mature strategy uses the app to win the customer and the owned channel —WhatsApp, web, phone— to retain them with data the apps never hand over. The theoretical cost —the fee in the contract— rarely equals the real cost per order of in-house delivery once you add insurance, fuel, turnover and delivery waste. A badly packed order arriving cold is delivery waste as real as kitchen loss: the average restaurant already wastes between 4% and 10% of what it buys (The Restaurant HQ, 2025), and in-house delivery adds a layer of loss from in-transit damage.

4. The theoretical contract cost rarely equals the real cost per order

Add courier turnover in a sector where 98% of operators reported rising labor costs in 2024 (National Restaurant Association): every new rider costs to recruit, train and equip. The classic error is modeling in-house delivery with today's fuel cost and zero turnover. The real cost per order includes the scooter down for maintenance, the returned order and the shift paid without dispatches. Model the worst month, not the best. In-house delivery shifts the commission expense to the labor line and pressures prime cost; apps leave it out of prime cost but bite gross margin per order. This is the accounting difference almost no one models. App fees hit gross margin but never touch payroll; in-house delivery raises labor, and with the sector's target prime cost at 55% to 65% of sales —healthy goal ≤60% (Toast)— a courier team can push you over the ceiling without the ticket rising.

5. In-house delivery shifts the cost to labor and pressures prime cost

In a market where the peak of restaurant price inflation hit 8.8% in March 2023 (National Restaurant Association), margin is already compressed. The right question is not 'how much fee do I pay' but 'where do I want that cost to live': if your prime cost has slack, absorbing delivery labor is viable; if it already brushes 65%, the app's variable fee is the lesser evil that protects your structure. In-house delivery wins when your volume is high and steady and your prime cost has slack to absorb fixed labor; the app wins when your volume is irregular or your prime cost already brushes 65%. The Masterestaurant operating rule is not to pick one or the other but to structure both: the app as a paid-acquisition channel for new customers and the owned channel for profitable repeat orders. Calculate your crossover point by dividing the monthly fixed cost of in-house delivery by the fee savings per order; that result is how many orders you need for the owned channel to win.

6. When does each channel win, based on your volume and prime cost?

With the median sale price of a small restaurant at $773,000 in 2025, up 24% from 2021 (BizBuySell), protecting EBITDA and ownership of customer data directly raises the value of your business.

This is cold unit economics: model both channels, measure real contribution margin and let the number —not the noise— choose. The app cost is 100% variable (pure OpEx); in-house delivery mixes high CapEx with fixed payroll cost that runs even when not a single order comes in. Apps buy acquisition and reach; the owned channel buys repeat purchase and data ownership —they are different financial objectives, not substitutes. The theoretical cost (contract commission) almost never equals the real cost per order of in-house delivery once insurance, fuel, turnover and delivery waste are added. In-house delivery shifts the commission expense into the labor line and pressures prime cost; apps leave it out of prime cost but bite the gross margin per order.