What the delivery apps actually cost a Toronto restaurant

Restaurant net margins are thin. After food cost, labour, rent, and the rest, a healthy independent in the GTA keeps somewhere in the low single digits to high single digits as profit. Now hand 20% to 30% of a delivery order to a third-party app and the problem is obvious: on a lot of delivery tickets, the commission is the entire margin and then some.

This is a post about the actual numbers, and about the decision that follows from them. It is not “delivery apps are evil, quit them today.” That advice is usually wrong. It is a post about when your own online ordering is worth building, and when the apps are still the right channel.

What the apps actually charge

The big three publish (or negotiate) commissions in roughly the same band. Current Canadian marketplace rates:

• Uber Eats runs three marketplace tiers: Lite at 20%, Plus at 25%, and Premium at 30% on delivery orders. Pickup is 10% if you submit proof your in-app prices match in-store, 15% if you don’t.
• DoorDash is tiered too: Basic at 20% delivery / 10% pickup, Plus at 25% / 8% (27% for DashPass customers), and Premier at 29% / 8%.
• SkipTheDishes negotiates per restaurant rather than publishing a card, but lands in the same 20% to 30% delivery range.

Two things worth saying out loud. First, the higher tiers are not optional in practice: the cheaper tier buys you a smaller delivery radius and worse placement, so most restaurants end up on the 25% to 30% plans to stay visible. Second, the headline commission is the floor, not the ceiling. Sponsored placement, in-app promotions, and “buy one get one” campaigns are all extra, and they are how the apps quietly push your effective rate past the sticker number.

Why the apps are still worth it

Here is the part most “ditch the apps” articles skip.

• They are a discovery channel, not just a checkout. People browse Uber Eats the way they browse a mall. A new customer who has never heard of you finds you there. Your own site does not do that on day one.
• They handle the hard logistics. Drivers, dispatch, live tracking, refunds, and the support call when an order goes wrong. That is real operational weight you would otherwise carry.
• Direct ordering has lost its old price edge. A few years ago, ordering direct was clearly cheaper for the customer. That gap has mostly closed, and surveys now show customers are about equally happy ordering either way. So you will not win first-party orders just by being a couple dollars cheaper.

The takeaway: the apps are a customer-acquisition cost. The mistake is paying that acquisition cost forever on customers who already know you and would happily order direct.

What first-party ordering actually costs

Your own online ordering is close to commission-free. The cost is card processing plus, if you deliver, a way to get the food there.

• Processing. Even Uber’s own first-party product, Webshop, is commission-free at a 2.9% order-processing fee. A custom storefront on Stripe or Square sits in the same neighbourhood, roughly 2.9% plus a small fixed fee per card transaction. Either way you are comparing about 3% against 25% to 30%.
• Delivery. This is the real catch, and the reason “just build a website” is incomplete advice. Pickup orders cost you nothing beyond processing. For delivery you either run your own driver or plug in an on-demand courier. Uber Direct, for example, charges no commission and prices by distance, and you can pass that fee to the customer the same way the apps already do.

So first-party ordering trades the 30% cut for a 3% cut, in exchange for you owning the marketing and arranging delivery. For your regulars, that trade is lopsided in your favour.

The break-even, on your own numbers

Run this with your real figures. As an illustration, take a restaurant doing $20,000 a month in delivery at an effective 28% commission. That is $5,600 a month going to the apps.

Now assume a third of those orders are repeat customers who already know you, roughly $6,667 a month. Move just that slice to first-party:

• App commission avoided: about $1,867 a month.
• First-party processing at ~3%: about $200 a month.
• Net kept: roughly $1,667 a month, or about $20,000 a year.

A first-party ordering setup is a one-time build plus a low monthly cost. In our pricing terms it is a marketing site plus one real feature, which we cover in what a custom small business website actually costs in Toronto. At savings like the example above, it pays for itself inside six to nine months, and everything after that is margin you keep.

The flip side, stated honestly: if your delivery volume is small, or most of it is genuinely new customers discovering you on the apps, the commission is buying you growth and the build can wait.

What we build for restaurants

When the math says first-party is worth it, the thing we build is not “a website with a menu PDF.” It is:

• A fast, mobile-first menu and ordering flow. Most orders are placed on a phone, so the cart, modifiers, scheduled pickup or delivery windows, and checkout all have to be quick on mobile.
• Real payments, Stripe or Square, with the kitchen getting a clean order ticket the moment someone pays.
• Delivery without the 30% cut, by wiring in an on-demand courier (Uber Direct or a local service) for the orders you do not want to drive yourself.
• A reason for regulars to switch. A QR code on the table, on the receipt, and on the takeout bag, plus simple loyalty or repeat-order nudges, is what actually moves customers off the apps and onto the channel you own.

We are not in the business of talking you off the apps entirely. We build the channel you control and point your repeat customers to it, so you stop renting access to people who are already yours.

Where we fit

We are a small Toronto studio that builds custom sites and ordering systems, the same way we built the booking and quote flow for Painter’s Crew. If you want a straight read on whether your delivery volume justifies building your own ordering, or whether you should stay on the apps for now, send us your numbers. Monthly delivery sales and your current effective commission are enough for us to tell you which side the math lands on.