Own Online Ordering vs Delivery Apps: Which You Need
Delivery apps bring strangers. Your own ordering keeps regulars. Here's how to tell which problem you actually have before you spend money.
A guy orders from you on Uber Eats on a Tuesday. He likes the food. Three weeks later he wants it again. Where does he go? Back to the app, because that's where you live in his head, and the app charges you again for a customer you already earned.
That's the whole case for direct ordering in one sentence. The case against rushing into it is just as short: if nobody orders from you twice, a direct ordering page is a button nobody presses.
So the question isn't whether commissions hurt. They do. The question is whether you currently have the thing that direct ordering is designed to capture. Let's work that out before you spend a dollar.
The two jobs are different
Delivery apps are a discovery channel. Someone opens the app hungry with no restaurant in mind, scrolls, and lands on you. You pay a finder's fee for that. It's expensive, but it's a real service, and no website you build will ever do it. Your site doesn't appear in front of a hungry stranger who wasn't looking for you.
Your own ordering page is a repeat channel. It serves the person who already knows your name. They're not browsing. They're trying to get the same chicken they had last month with the least possible friction.
When owners frame this as "apps versus my site," they usually end up angry and stuck. Framed as discovery versus repeat, the decision gets easy. Most restaurants need both, weighted differently depending on how old you are and how good your food is at making people come back.
We wrote separately about which delivery apps to keep and which to drop. That's the pruning decision. This is the building decision.
Before direct ordering pays, you need three things
Direct ordering is not a switch you flip to stop paying commission. It's a channel, and channels need inputs. Three of them.
A base of regulars. Not a huge one. A few dozen households that order from you more than once a month is enough to start. If you genuinely cannot name a category of regular — the Friday pizza families, the office lunch crowd, the late-shift nurses — then you don't have a repeat problem to solve yet. You have a repeat problem to create.
A way to reach them. A list. Phone numbers, emails, a receipt in a bag, a sign at the counter, a follow-up text after a direct order. Without this, your ordering page is a locked store with no address. Most restaurants that "tried their own ordering and it didn't work" skipped this step entirely. They launched the page, told nobody, and concluded that customers prefer apps.
A checkout as fast as the app's. This is where most of them lose. The app remembers your address, your card, and your last order. If your page asks for a full address retype, a password, and an account confirmation email, you will lose to the app even among people who want to support you. Guest checkout, saved details on return, Apple Pay and Google Pay, and a reorder button that repeats the last order in two taps. Nothing else matters if this part is slow.
If you have all three, direct ordering starts paying almost immediately, because the margin difference per order is enormous and you're applying it to orders you were already getting.
If you have no repeat business, fix that first
This is the part nobody wants to hear. If your orders are almost entirely first-timers who never return, building your own ordering system will not change your revenue. It'll change which invoice the money leaves on.
A restaurant with no repeat business has a product or an experience problem, not a software problem. Cold food on arrival. Portions that look nothing like the photo. A menu nobody can remember the name of a dish from. Packaging that leaks. Those things kill repeat business quietly and no checkout page fixes them.
The honest test: pull last month's delivery app data and look at how many customer IDs ordered more than once. The apps show you this. If that number is tiny relative to total orders, spend the next quarter on the food, the packaging, and the pickup experience. The ordering page will still be there, and it'll be worth far more when you build it.
Our restaurant services page walks through what we actually set up and in what order, which is useful if you want to see where ordering sits relative to everything else.
How to move app customers to direct, without getting cute
You can't message app customers directly — the apps own that relationship and don't hand it over. But the food leaves your kitchen in a bag you control, and the pickup counter is yours.
Things that work:
- A card in every bag with your direct ordering URL and one clear reason to use it. Not "save us money." Something that benefits them: a free item, a dollar amount off, skipping the app's service fee.
- A short, typable domain. If your direct URL is a 40-character subdomain of some platform, nobody retypes it. Buy a short one.
- A QR code on the receipt and on the counter that opens the menu, not your homepage.
- Naming the offer at the counter when regulars pick up. Staff saying "next time order from our site and I'll throw in a drink" converts better than anything printed.
We covered the bag side of this in detail in your takeout bag is your best marketing channel. The short version: it's the only ad every single customer definitely sees.
One thing to avoid — don't price your direct channel higher than the app to "cover the platform fee," and don't price the app items higher than your counter in a way that surprises people. Price consistency is a trust issue, and we've written about matching your online and in-store prices.
What the money actually looks like
Skip the fantasy math where every app order becomes a direct order. It won't. A realistic target in the first year is moving a meaningful slice of your repeat customers over, while apps keep doing discovery.
The arithmetic you should run on your own numbers:
- Take your monthly app commission total.
- Estimate what share of those orders came from repeat customers. The app dashboards will give you a rough read.
- Assume you convert maybe a third of those repeat orders to direct in year one, with consistent promotion.
- Compare the commission saved on that slice against your monthly ordering platform cost.
If the saved commission doesn't clear your platform cost in that calculation, you're too early. Keep the apps, work on repeat, revisit in six months. If it clears easily — and for most restaurants with an established regular base it does — the only reason to wait is bandwidth.
The version of this that actually works
Keep the apps running for discovery. Treat every app order as a paid customer acquisition, which it is. Then build a direct channel that's genuinely fast, put the URL in front of every person who touches your food, and give them a concrete reason to use it.
Over time the mix shifts. Your best customers move direct, the apps keep feeding you new ones, and your blended commission rate drops without you ever making a dramatic exit.
That's the whole strategy. It's unglamorous and it works, which describes most things that work in this business. Once ordering is live, the sequence of what to do in the first few months matters a lot — we laid that out in your first 90 days of online ordering.