Selling on delivery apps: the traditional method versus the Masterestaurant method

When the delivery app stops working for you?

The number that exposes the problem isn't in sales, it's in the deposit:

if your gross billing on the app climbs month after month while the money landing in your bank account moves less than half as fast, the app stopped being a channel and became a majority partner you never chose. DoorDash's published restaurant plans run at 15%, 25% and 30% of the ticket according to CloudKitchens, and that percentage applies to the selling price, not to your margin. Add the co-funded promotions, where you pay half of a two-for-one the app designed, plus the welcome discount for every new customer. An operator billing 11,400 USD and depositing 7,980 has a 30% leak before buying the first pound of chicken, and no recipe fixes that. The first alternative doesn't ask you to leave the app, it asks you to stop charging dining-room prices inside it.

Rebuilding the channel price: for whom, and at what cost

A 12 USD dish with 3.60 USD of raw material shows a 30% food cost that looks healthy, until the 25% commission takes 3.00 USD and thermal packaging another 0.55: the effective cost jumps to 46%, and that dish, which paid payroll at the table, eats payroll on delivery. You build it backwards, starting from the margin you need and adding commission, packaging and waste before showing any figure at all. It works for any owner with their own menu and no franchise-regulated pricing. Startup cost: zero, plus roughly six hours of spreadsheet work. Learning curve: low. The only real risk is psychological, because nobody enjoys seeing two prices for the same plate. Setting up a direct channel is the alternative with the best arithmetic over twelve months, and the market already validated it: 58% of customers prefer ordering through the restaurant's own app or website rather than an aggregator, according to an NCR Voyix survey reported by Restaurant Dive in 2024.

Your own ordering engine with local SEO

You pay a payment gateway —typically between 2.9% and 3.5%— instead of a double-digit commission, and you keep the phone number, the email and the order history. The profile this serves is the owner with recognized neighborhood branding and at least eight hundred monthly orders, because below that volume the traffic simply isn't there. Startup cost: 400 to 1,200 USD in setup plus a platform subscription. Switching effort: high for the first ninety days, since traffic has to be earned with local content and a properly worked business listing. Launching a second brand cooked on your current stove turns idle capacity into billing without adding rent, and institutional capital has been betting there for years: Research and Markets projects the ghost kitchen market at 142.5 billion USD by 2029, and Grand View Research estimates cloud kitchens at 203.72 billion by 2033. The operating logic is easy to state and hard to execute.

A virtual brand on the kitchen you already have

You create a separate concept —wings, bowls, desserts— with a short six-item menu sharing 70% of your existing inputs, so you never open a second storeroom. It serves the owner whose kitchen sits underused between three and six in the afternoon, with a head chef able to hold two ticket lines at once. Startup cost: 300 to 900 USD across photography, packaging and registration. Effort: moderate, and the whole thing is decided by inventory discipline. Here is the break that Diego F. Parra repeats in every Masterestaurant audit: the app isn't selling you food, it's selling you a customer you didn't have yet, and it charges a steep referral fee that only makes sense ONCE. According to Anand Vaidyanathan, Chief Technology Officer at Portillo's, a U.S. chain with more than seventy locations, the value of owned digital channels lies in the customer data that lets you personalize the offer and sustain repeat purchase.

Whoever owns the customer sets the margin, not the commission

Translated into cash: if out of every hundred app orders you pull thirty customers over to your direct channel, the acquisition cost amortizes and the commission becomes tolerable. Pull over zero and every order costs what the first one cost, meaning you are renting your own clientele in perpetuity. The slip of paper inside the bag, offering 15% off for ordering direct, is the cheapest tool that exists for this. The most underestimated alternative is still the phone or WhatsApp order with in-store pickup, and the reason is pure arithmetic: no commission, no delivery packaging in half the cases, and immediate payment. The global market is far from saturated —Statista puts user penetration in meal delivery at 29.2% for 2026— so not every neighbor of yours is inside an app. An in-house rider with a motorcycle costs between 450 and 700 USD a month across most Latin American cities, and past roughly twenty daily orders that salary comes out cheaper than 25% commission on those same orders.

Counter, pickup and your own courier

The profile: a street-front location with neighborhood regulars and a short delivery radius, under three kilometers. Switching effort: low on technology, high on people management, which is exactly where most owners quit. Let's take the scenario all the way through, since almost nobody does. You switch off the app on a Monday: 40% of your orders vanish at once, and of those, the share that already knew you —say a third— searches Google, finds your listing and orders direct. The other two thirds order something else entirely, because they were never your customers, they were the app's customers that an algorithm parked in front of your menu. By the third month your gross billing has dropped 27% and your operating profit has risen, because that 40% was being sold at a negative margin. There sits the paradox almost nobody resolves: selling less and earning more is possible when the channel you cut was the one draining you.

What happens if you switch the app off tomorrow?

The condition without which none of this holds is having built the direct channel first. Switching it off with no net underneath closes the till, it doesn't optimize it.

Stay exactly where you are if your location opened less than six months ago, if the app brings you more than 60% of your volume and if you still have no brand recognition in the neighborhood, because at that point the commission isn't a tax, it's advertising that also hands you a prepaid order. With user penetration at 19.8% in Colombia's meal delivery segment for 2024 according to Statista Market Insights, there is demand there you cannot reach alone. Don't change either if your kitchen already runs at its limit during peak hours: adding a virtual brand onto a saturated stove produces nothing but long tickets and two-star ratings that take a year to repair.

When NOT to change anything?

And if your dishes can't survive fifteen minutes in transit, the problem isn't the channel, it's the menu. Fix the menu first and come back to this decision in a quarter.

The gap is not the commission, it is WHO sets the price. Under the traditional method the dining-room price comes first and commission eats whatever is left; under the Masterestaurant method the channel price is built backwards, starting from the margin you need and adding commission, packaging and loss before a figure ever reaches the guest. A 12 USD dish with 3.60 USD of raw material looks healthy until the app takes 3.00 USD and packaging another 0.55 USD: effective food cost jumps to 46%, and that dish, which funded payroll in the dining room, now consumes it. The second break is guest ownership.

Where the delivery business really breaks

According to Anand Vaidyanathan, Chief Technology Officer at Portillo's, the US chain with more than seventy locations, the value of owned digital channels lies in the first-party data they return to the restaurant so the offer can be personalised, a position he has argued publicly at industry forums, and it matches what operations show: whoever fails to capture the guest's phone number pays commission again for that same guest on every return. The third is fragility. An app can rewrite its ranking, raise its fee or open a competing dark kitchen in your zone without asking. Your Google Maps listing answers to signals you control: 5★ reviews, fresh photos, a correct primary category and consistent name, address and phone. I took years to accept this, and I was wrong for a long stretch recommending in-app advertising ahead of local listing work; the correct order runs the other way, because the listing compounds and in-app spend evaporates the day you stop paying.

Where the delivery business really breaks — in practice?

There is a genuine tension worth settling head-on: apps DO bring volume you cannot generate alone, especially for a new venue or a virtual brand with no history.

The answer is not picking a side. It is accepting 25% commission as the acquisition cost of the first order, then refusing to pay it on the second, the third and the tenth.