The year restaurants started taking their customers back

In March 2020 we wrote that restaurants should put an ordering page on their own site that week, and keep it when the dining rooms reopened. Most did not; the apps were easier, and easy won. Five years on, the arithmetic has caught up. Commission on the platforms sits at 13–15% for a basic listing and 25–35% when the platform delivers, and the increases have only gone one way. For a place doing €5.000 a month through the apps, that is somewhere between €8.000 and €20.000 a year, paid mostly for orders from regulars.

This is the year we have watched owners do the sum and change their minds.

What changed

• The regulars are on the app, and the app owns them. Their name, their address, their order history, their loyalty — all held by a platform that will happily show them your competitor's promotion at the top of the screen.
• The margins got thin everywhere else. Energy, rent, wages, produce. The one line that can be cut without hurting the food is the commission on people who would have ordered anyway.
• The tools got good. Own-site ordering used to mean a clunky plugin and an email to the kitchen. Now it means a menu that opens in place on any phone, time slots that know the opening hours, iDEAL, a screen in the kitchen, and a customer list that belongs to the restaurant. We build these, and so do others; the point is that they exist.

The split that works

Not "leave the apps". Use them for what they are good at — discovery, the person who has never heard of you — and move everyone else to your own channel. Regulars, offices, the neighbourhood: a QR on the receipt, a line on the bag, a message to the list. "Order from us directly next time; here is the link, here is 10% off, no commission." People do it, because it is cheaper for them too.

What own-channel makes possible that the apps never will

• Shared orders — a table, an office, a group, joining one order from their own phones, the kitchen seeing it as one ticket.
• Business customers on account — invoiced monthly, the way offices actually buy lunch.
• Loyalty that is yours — points, not a platform's badge.
• Your data — who ordered what, when, and how to reach them.

The honest caveats

An ordering page does not fix a kitchen that cannot handle time slots. It needs a menu somebody keeps current. And it needs the restaurant to actually tell customers it exists — the best system in the world earns nothing on page four of the site.

We will publish the numbers from the next one we launch. Five years late is still early; the commission is not going down on its own.