Your Own Delivery App: Stop Handing 30% to the Marketplaces

Every order that comes in through a delivery marketplace shaves off between a quarter and a third of the sale before you touch a cent. Your own delivery app becomes profitable above roughly 600-800 monthly orders through the marketplaces: at that volume, the commissions you pay in a single year exceed the cost of building your own channel (USD 12,000-25,000), and every order you migrate turns directly into recovered margin. In this article we run the numbers with you: how much you lose today, what the app needs at a minimum, and how to coexist with the marketplaces while you migrate customers.

What the delivery marketplaces really charge you

The advertised commission is never the final cost. This is what we see in the numbers of restaurants and dark kitchens across LATAM:

Item	Typical range
Per-order commission	25-30% of the ticket
Tax on the commission	+21% of the commission (Argentina)
In-app advertising (to show up)	3-8% extra on revenue
"Co-funded" promotions	The discount usually comes out of your margin
Total effective cost	28-35% of every sale

Let's put numbers to a typical case: a dark kitchen doing 800 orders a month at an average ticket of USD 11. It bills USD 8,800 through the marketplaces and hands over about USD 2,640 a month in commissions. That's more than USD 31,000 a year: a cook-and-a-half's salary, or the full build of its own channel, given away every single year.

And there's a cost less visible than the percentage: the customer isn't yours. You don't have their phone, their history, or any way to reach them. If tomorrow the marketplace raises its commission or buries your shop in the listing, you've got nothing.

Minimum features of your own delivery app

You don't need to clone the marketplaces. You need four things that work flawlessly:

1. Catalog and cart: menu with photos, variants (size, extras), opening hours and a delivery zone validated by address.
2. Online payments: a local payment gateway, with cash-on-delivery as an option. If the payment fails or confuses people, the order dies right there.
3. Order status: confirmed → being prepared → on the way → delivered, with push notifications. Real-time courier tracking on a map is nice to have, but most customers are happy with clear statuses and honest times.
4. Kitchen and admin panel: an order screen, per-product stock control ("out of stock" in one tap), sales reports.

If you run your own couriers, add a simple driver view (assigned orders, navigation, mark as delivered). If you outsource logistics, you integrate a last-mile delivery API and forget about it.

All of this falls into what in mobile app development we call a well-trimmed MVP: we cover it in detail in how an app is built step by step.

What it costs, and the mini break-even calculator

Realistic ranges for LATAM in 2026:

• MVP (orders + payments + statuses + panel): USD 12,000-25,000
• Full version (map tracking + driver app): USD 25,000-40,000
• Monthly maintenance: USD 300-800 (servers, iOS/Android updates, fixes — the breakdown is in how much it costs to maintain an app)

The break-even math is simple. Suppose you manage to migrate 40% of your orders to your own channel (a conservative target a year after launch) and the effective marketplace cost is 30%:

Monthly savings = migrated orders × average ticket × 30% − maintenance cost

Orders/month via marketplaces	Migrated orders (40%)	Gross savings/month (USD 11 ticket)	Break-even on a USD 18,000 app
400	160	USD 528	~5 years — not worth it yet
800	320	USD 1,056	~24 months
1,500	600	USD 1,980	~12 months
3,000 (small chain)	1,200	USD 3,960	~6 months

From there on, everything you save is margin, every month, forever. And if you migrate more than 40% — repeat customers are usually the first to switch over — the timelines shorten.

Want this math done with your real commission numbers? Book a 30-minute call and we'll build it together, no strings attached.

The hybrid strategy: coexisting with the marketplaces while you migrate customers

The classic mistake is framing it as "own app versus marketplaces." The right framing is by function:

• Marketplaces = acquisition. They put you in front of people who didn't know you. For that service, 30% on the first order can make sense.
• Own app = retention. The customer who has already ordered three times doesn't need to discover you: paying the middleman 30% on that order is giving away margin.

Concrete tactics that work:

• A QR code on every bag and every box: "Order from us directly and get 15% off." That 15% costs you half of the commission you save.
• First order in the app with free delivery, funded by the commission difference.
• Your own database: every direct order leaves you a phone number and a history. A push saying "Thursday's promo is back" to dormant customers costs nothing and reactivates between 3% and 8% depending on the category.
• Slightly better prices in your channel (marketplaces usually forbid publishing this in their app, but your packaging and your social media are yours).

The goal of year one isn't to kill the marketplace: it's to move 30-40% of your repeat orders to your own channel and stop paying a toll on customers who are already yours.

When your own delivery app is NOT worth it

Let's be honest, because this is where a lot of agencies oversell:

• Fewer than 300-400 monthly orders through marketplaces. The savings don't justify the investment yet; focus on growing volume first.
• Low repeat rate. If you sell something people order twice a year, there's no repeat customer to migrate.
• You haven't solved logistics. The app takes orders; someone has to deliver them. Without your own couriers or a reliable last-mile third party, the direct channel will generate complaints, not margin.
• Your brand doesn't exist yet. If 100% of your sales depend on showing up in the marketplace listing, first build a customer base that looks you up by name.

In those cases, a middle-ground alternative is a ordering PWA: it costs a fraction, doesn't require going through the app stores, and validates whether your customers are willing to order from you directly. We compare the two options in app vs PWA: which is better.

The next step

If you're already above 600-800 monthly orders, every month you spend without your own channel has a concrete, calculable cost: it's that month's commissions. At Deepyze we build delivery and ordering apps for restaurants, dark kitchens and chains across LATAM — from the MVP to the version with your own logistics — with a fixed price closed up front, a proposal within 24 hours and a team in your time zone. Tell us your case with your current order volume and we'll send back the exact number for when it pays for itself.