QR ordering arrived everywhere at once for reasons that had nothing to do with restaurants wanting it, and a lot of rooms have since quietly gone back to paper. The ones where it stuck did two things: they treated the code as a shortcut for the guest rather than a way to cut staff, and they printed the card properly. Here is what happens between the scan and the pass, and what it costs to set up.
What the guest actually does, in order
They point a camera at a card, a menu opens in a browser, they add items, they type a table number or the code carries it, and they pay. Every one of those steps is a place to lose them. The most common losses are a code that needs an app installed, a menu that loads slowly on a bad signal, and being asked to make an account before ordering. None of those is necessary and all three are common.
What the card has to survive
A table card lives in a room with spilled drinks, hot plates and children. Laminate it, print the code large enough to scan from a seated distance in low light, and put the table number on the same card so a guest reading it aloud to a server is saying something useful. Expect to replace a few every year for damage rather than for menu changes, because the code does not change when the menu does.
Where the time actually goes
The saving is not the order taking, it is the second trip. A server no longer walks back to a table to take a drinks order, and no longer walks a card machine over at the end. In a busy room that is a real number of minutes; in a quiet one it is nothing, and the guest would rather have talked to somebody. This is why the same product delights one operator and is turned off by another within a month.
What it costs to put a code on every table
Two numbers: the cards you print and the reprinting you stop doing. On the worked example in the sheet on this site, 24 tables with one card each cost £20.40 to code, against 240 paper menus a year at four menu changes, which is £336.00 of printing, leaving £315.60 avoided in the first year. Put your own table count and print prices in and the answer moves; the shape of it rarely does.
When to turn it off
The code is a tool for the busy half of the week, not a policy. On a quiet Tuesday, with two servers and eleven covers, a guest would rather be asked, and the room reads warmer for it. On a Friday with the garden open the code is the difference between a drinks order taken in a minute and one taken in nine. The rooms where it stuck are the ones that switch: cards on the tables always, the ordering page live when the room is full, and a server who says either is fine. The rooms that went back to paper are the ones that made it the only way, on the nights it was the wrong way.
Questions people ask about qr code ordering
Do guests need to download an app?
They should not. A code that opens a page in the browser works on every phone made in the last decade. An order and pay app is a bigger ask for a smaller return, and it is the quickest way to have a table wave at a server instead.
What about guests who will not use it?
Keep paper menus and take orders at the table as normal. Rooms where this works treat the code as an option, not a policy. The saving comes from the reprint run you stop doing, and that survives keeping a handful of paper menus.
Does the code have to be reprinted when prices change?
No. The code points at a page you edit, so a price change is live in seconds and the card on the table is still correct. That is the whole economic argument for it.
Should QR ordering replace the server taking the order?
No. It replaces the second trip and the walk with the card machine; it does not replace the welcome, the recommendation or the question about allergies. A room that removes the server from the first minute to save a wage loses the guests the code was meant to serve faster.