Guide
Upsells people actually take.
A euro fifty on the average order is worth more than most marketing campaigns, and the customer is already buying.
In short
The upsells that convert are relevant, cheap relative to the basket, and offered at the moment of adding rather than at checkout. A side or a drink suggested alongside a main typically converts several times better than a generic add-on prompt at the end. Adding €1.50 to an average order across 60 orders a day is roughly €2,700 a month in extra revenue at almost no extra cost.
Updated 28 August 2026
The arithmetic that makes this worth doing
Take a restaurant doing 60 orders a day at €24. Add €1.50 to the average and you have added about €2,700 a month. The food cost on a side is low and the labour is already paid for, so most of that lands as gross profit. Compare the effort against acquiring the equivalent revenue in new customers and it is not close. This is the cheapest growth on the list, which is why it is first.
Relevance beats persuasion
A suggestion that fits what is already in the basket converts. A generic prompt does not. Garlic bread next to a pizza works. The same prompt next to a salad does not, and it slightly damages trust in every other suggestion you make. If your system can only do one rule, make it pair-based: what do people who ordered this also order.
- Pair the suggestion to the item, not to the order total
- Keep it to one or two, a wall of add-ons gets ignored
- Price it well under the item it accompanies
Online outperforms the counter, and by a lot
A prompt on screen is patient, consistent and never embarrassed. A member of staff on a Friday is none of those things, and will stop asking by eight. This is the main reason online average order values usually run above phone or counter for the same restaurant. If you want the upsell to happen every time, it has to be in the ordering flow rather than in a briefing.
Timing inside the flow
Offer at the point of adding the item, not at checkout. By checkout the customer has mentally finished and is looking for the total; anything that appears there reads as an obstacle. The one exception worth making is a single low-friction prompt for something genuinely forgotten, like a drink or a dip, which works because it completes rather than interrupts.
What kills it
Three things. Suggesting something more expensive than the item it accompanies. Suggesting the same thing every time regardless of basket. And upselling into a slow kitchen, where every added side pushes the ticket time out and the food arrives cold. The last one is the least obvious and the most damaging, because you win €2 and lose the customer.
Questions, answered straight.
Not covered here? Just ask us.
What is a realistic lift?
A well-built pairing prompt commonly adds a few percent to average order value. Anybody quoting a large double-digit lift is describing a best case or counting a promotion as an upsell.
Should staff still upsell at the counter?
Yes, on one thing, scripted and specific. Asking whether somebody wants a drink converts. Asking whether they want anything else does not, because it is a question about their whole life rather than about a drink.
Do upsells annoy customers?
Relevant ones rarely do. Irrelevant ones reliably do, and the damage carries over to the ones that would have worked. This is a case where doing it badly is worse than not doing it.
Does this work for delivery?
It works best for delivery, because the ordering happens on a screen where the prompt can be consistent and the basket is visible. Just make sure the kitchen can carry the extra items at peak.
See it run your restaurant.
Twenty minutes on your menu and your numbers. We'll show you what actually changes in the first month and exactly what it costs. If it's not right for your restaurant, we'll tell you that instead.
- Live in days, not months
- We build your menu and train your staff
- No contract, and thirty days to change your mind
Rather just ring us? +353 87 438 8032
Tell us about your restaurant
We reply the same day, usually inside a couple of hours.