Your website gets traffic. People search your name, tap the link in your Instagram bio, scan the QR code on the receipt. Then about 97 of every 100 leave without ordering. The traffic is fine. What's broken is the website, and that's the most fixable leak in restaurant marketing, because everyone who left already wanted your food.
A restaurant website has exactly one job
Most restaurant sites were built to look nice. A hero photo, an "our story" page, opening hours, maybe an embedded map. Ordering, the thing the visitor came to do, gets a button labeled "Order online" that opens a third-party widget or, worse, a marketplace listing where you pay commission on a customer you brought in yourself.
Now look at how the visitor actually behaves. They're hungry, on a phone, deciding in under a minute. Every extra tap, every redirect, every "create an account to continue" peels a few more of them away. Across the 3M+ orders processed on our platform the hierarchy holds steady: branded apps convert up to 35%, well-built ordering websites up to 15%, marketplace listings around 3%. A typical brochure-style restaurant site lands right down with the marketplace at about 3%, because it makes ordering just as hard.
The five leaks that cost you the other 12%
1. The PDF menu is unreadable on a phone, invisible to Google, and you can't order from it. Every dish should be a card with a photo, a price, and an "add" button, not a line in a scanned document.
2. A redirect breaks the flow. An "Order online" button that throws the visitor onto a different domain resets their trust and their patience. Operators who move ordering onto their own domain see a lift right away, because the visitor never has to wonder again whether they're in the right place.
3. Slow mobile load. Most of your ordering traffic is on phones, and if the menu needs more than a few seconds to render on a mid-range handset, a chunk of visitors is gone before a single dish appears.
4. Forced registration. Asking for an account before you'll even show the cart is asking a hungry person to fill out paperwork. Take the order first: name, phone, address, payment. Offer the account afterward, once they have a reason to want order history and cashback.
5. No answer to "when will I get it?" Delivery fee, minimum order, and estimated time belong on screen before checkout, not sprung as a surprise on the final step. Surprise fees are the classic reason carts get abandoned at the last moment, in food as much as in the rest of e-commerce.
Bring your current site to a 30-minute call. We'll walk through the leaks against real benchmarks.
What a 15% ordering site does differently
The sites that reach double-digit conversion share a short list of traits, and not one of them is decorative.
Start with the menu, because on these sites the menu is the homepage. It shows the real, orderable menu one tap from arrival, not a photo of the chef. Categories are named the way customers think ("burgers", "for two"), photos sit on the items that sell, and the modifiers work without reading instructions.
Checkout has to fit on one screen: address with autocomplete, saved details for returning customers, card or wallet payment. A repeat visitor should be able to reorder their usual in three taps, and repeat customers are where the 15% number actually comes from.
The site also has to remember people. A first-time buyer who had a good experience is worth a second order only if you can reach them, and on a marketplace you can't. On your own site every order builds your customer base, with name, phone, and history, which feeds cashback, push, and SMS campaigns that bring people back without ad spend.
And somebody has to watch the numbers. Conversion isn't a property you set on launch day; it drifts. A menu change, new photos, a tweak to the delivery fee, each one moves it. A weekly look at orders, conversion, and repeat rate tells you what helped and what hurt.
The commission math that makes this urgent
A $30 order through a marketplace like DoorDash or Uber Eats, at an effective 30–40% cost, leaves you $18–21. The same order through your own site costs you payment processing, roughly a dollar. If your site converts at 3% when it could convert at 12–15%, the loss isn't only the missing orders. You're also handing your own Google traffic, the people typing your name into search, to the channel that charges you the most for them. We ran the full line-by-line marketplace math in a separate piece; the short version is that every order you shift to your own site keeps about a third of the ticket in your pocket.
Do this next
Open your website on your phone right now and time yourself ordering your own bestseller. Count the taps and the surprises. If it takes more than two minutes, or you run into a PDF, a redirect, or a forced sign-up, you've found the cheapest growth project you'll take on all year. Fixing the ordering path costs less than a single month of marketplace commissions, and unlike those commissions, you pay for it once.
Menu, checkout, payments, delivery zones, and marketing tools — on your domain, with zero commission.