Selling on apps without costing the channel vs delivery unit economics and own channel

In consulting I find owners who say 'delivery is going really well' and when I check the numbers, that channel is losing money. Sales grow, cash doesn't. The problem is the platform commission never enters the channel cost: it's seen as a general 'marketing expense' and no one crosses it with the margin per order.

I'm not saying to abandon platforms — in many models they're the most efficient acquisition channel. What I'm saying is you need to know exactly what each order costs you per channel, and make that decision with the data in hand.

The commission nobody includes in channel cost accounting

Delivery platforms charge between 15% and 30% per order, and that cost rarely makes it into real channel cost accounting. On a $25 order at 25% commission, $6.25 leaves before you've spent a cent on food, packaging, or payroll. If your food cost is 30%, you've already committed $13.75 on that same order ($6.25 in commission plus $7.50 in food cost) and you still haven't counted energy, proportional rent, or cook time. The mistake Diego F. Parra sees over and over in consulting: owners record the commission as a general 'marketing expense,' spread it across all channels, and never cross-reference it with the margin per order. Delivery grows in sales, cash doesn't show up. That gap isn't bad luck; it's a costing problem. Compare two identical $25 orders: one through a platform at 25% commission, one through your own channel (WhatsApp, proprietary app, or web with a payment gateway).

Platforms vs your own channel: what stays in your register

On the first, the platform keeps $6.25 and you receive $18.75 gross. On the second, the payment gateway charges between 2.5% and 3.5% ($0.63–$0.88), and the rest stays in your register. The net difference per order is roughly $5.37–$5.62. If your operation handles 80 delivery orders per day, that gap is between $430 and $450 daily you're currently handing to the platform. Per month: $13,000–$13,500. With those numbers on the table, the question is no longer 'should I be on platforms' but 'how much volume can I migrate to my own channel without losing customers.' That is the decision the Masterestaurant method requires you to make with data. Most restaurants track total sales by channel, not margin per order by channel. Those are different metrics, and the second one is what matters. A channel can represent 40% of your sales and 10% of your real margin; another can move 20% of the volume and contribute 45% of the profit.

Margin per order as a survival metric

That asymmetry appears when you cross food cost plus commission plus packaging plus customer acquisition cost. In a dark kitchen with a $18 average ticket and a 28% commission, operating margin per order can fall to $1.80–$2.40 before covering rent and utilities. At that level, the channel's break-even requires a volume most operations can't reach. Masterestaurant builds the margin-by-channel table in the first week of consulting; it's the starting point, not an analytical luxury. Platforms make the commission visible, but there are three costs delivery generates that don't appear in any automatic report: specialized packaging ($0.45–$1.20 per order depending on format), assembly and dispatch time (2–4 additional minutes of labor per order), and rejection or refund rates for quality issues in transit (between 1.5% and 4.5% of GMV in temperature-sensitive categories). Add those to a 25% commission and the real cost of the channel can exceed 34–38% of the sale price before food cost.

Packaging, time, and the invisible costs of delivery

In that scenario, selling through a platform with a standard food cost of 30% means operating at a structural loss, not a tight margin. The own channel eliminates the commission and lets you optimize packaging because you control order flow and dispatch batches. Platforms are not the enemy; they are an acquisition channel whose cost you must know precisely. They make sense when the lifetime value of a customer acquired through the platform far exceeds the commission paid, or when you're in an area where you have no brand presence and need discovery. They don't make sense as a permanent, uncosted channel when you operate with low tickets (under $15), tight margins (food cost above 28%), and no strategy to migrate the customer to a direct channel. Diego F. Parra has reviewed operations where 70% of digital volume came from a single platform, creating a dependency that became a hostage situation when that platform raised its commission from 20% to 27% in one quarter.

When platforms make sense and when they don't

Diversifying the digital channel is not optional in 2026; it's operational risk management. Building a direct channel that moves 40–60% of digital volume without commission doesn't happen in the first month, but it's not a two-year project either if executed with method. Diego F. Parra works with restaurant groups that have achieved that transition in 9–14 months using three levers: a proprietary customer database with order history, a loyalty program with a differential incentive for the direct channel (10–15% discount vs platform price), and automated order capture through WhatsApp Business or a lightweight app. The Masterestaurant Growth Map defines the exact sequence: first activate the direct channel for repeat customers who already know you, then reduce platform exposure by zone and time slot, and finally renegotiate rates from a position of strength because you no longer depend on the channel. Artificial intelligence applied to order data analysis can identify in hours what used to take weeks of manual Excel work.

AI applied to channel profitability analysis

With platform APIs and POS data, it's possible to calculate the real margin per order by channel, by time slot, and by delivery zone. In operations with more than 150 daily delivery orders, that analysis reveals non-obvious patterns: Friday orders between 8 pm and 10 pm carry a 23% higher ticket and a 60% lower refund rate than Monday midday orders. That data changes the pricing strategy by channel and availability by platform. Masterestaurant integrates AI not as a passive indicator dashboard but as an active decision tool: which channel to open, when, and with what menu, based on real profitability data. The concrete action Diego F. Parra recommends is simple and requires no technology: take the last 30 days of sales by platform, pull the total commissions paid (it's in each app's report), add estimated packaging costs, and divide by the number of orders. That number is the channel cost per order, and if you don't know it, you're operating on a fictional margin.

First step: cost every channel before this week is out

Compare that cost to the average ticket minus food cost and you get the real gross profit per delivery order. If that number is negative or below $2.50 per order, the channel is subsidizing sales with working capital. The Masterestaurant method starts with that diagnosis in the first session: without the real cost of every channel on the table, any growth decision in delivery is a bet, not a strategy. The most expensive delivery mistake isn't being on platforms: it's not knowing what each incoming order costs you. A 25% commission on a $25 order is $6.25 out the door before you've spent a cent on food or the cook. If your food cost is 30%, on that order you've already given up $13.75 ($6.25 + $7.50 food cost) and you still haven't counted packaging, energy or a single dollar of payroll.

Why costing by channel decides if delivery is an asset or a trap

Diego F. Parra works with restaurant groups that have built own channels moving 40–60% of digital volume commission-free. It doesn't happen in the first month, but with the Growth Map and the dark kitchen canvas, the transition is a planned process, not a leap into the unknown. AI applied to demand analysis and per-channel price optimization is the accelerator of that process.

Point-by-point analysis: selling on apps without costing (A) vs delivery unit economics with own channel (B)

Side-by-side comparison

The numbers that matter

How to move to costing by channel and building your own

Free tools to apply this now

Method tools for designing profitable delivery channels

The Masterestaurant method has specific tools for digital operations:

Diego F. Parra — International consultant, expert in creating and scaling restaurants and in AI applied to restaurants, foodtech and HORECA. Methodology applied in 8.400+ restaurants across 43 countries · Expert in Artificial Intelligence applied to restaurants, hospitality and food businesses · 20+ years in restaurants, catering, large events and business growth · Author of the book «From Slave to Owner» (Amazon) · International keynote speaker for the HORECA sector.

Full profile →

Frequently asked questions about delivery commissions and own channels

Sector data 2026 (official sources)

Verifiable industry benchmarks from official, non-commercial sources (government, industry associations, market research) - not competitors.

Related content