Home / Blog / Restaurant loyalty program: designing one people finish

Restaurant loyalty program: designing one people finish

Most loyalty schemes fail for the same four reasons, and none of them is the size of the reward.

· Co-founder, Get Menu · 2 September 2026

A restaurant loyalty program is a promise that coming back is worth something. Most of them fail, and almost never because the reward was too small.

This guide covers which mechanic to choose, the four mistakes that quietly kill schemes, and how to make one that runs without anybody at the counter having to remember it.

Choose the mechanic before the reward

There are three shapes, and they suit different businesses.

Stamps. Buy a number of things, get one. Simple, visible, and easy for a customer to hold in their head. Best where the purchase is frequent and roughly the same size each time: coffee, breakfast, lunch.

Tiers. Keep coming and you move up, and the higher tier carries a standing discount. Best where visits are less frequent but larger, and where feeling recognised matters as much as the money.

Points. Spend, accumulate, redeem against a range of things. The most flexible and the least motivating, because nobody can picture where they are.

Pick one. Restaurants that run two at once usually find customers understand neither.

The four mistakes

The card that is not there. A paper card is at home, in a coat, in a bin. If the customer has to produce something to earn, most of the earning never happens. This is the single biggest cause of a scheme quietly dying.

A goal nobody can see. "You have some points" motivates nobody. "Two more and the next one is free" motivates. If the customer cannot say where they are without asking, the scheme is decoration.

A reward that arrives too late. A goal ten visits away is invisible to somebody who comes twice a month. Set the first reward close enough to feel reachable, then a second one further out.

Depending on staff. If a stamp requires a member of staff to remember, it will not happen at eight in the evening, which is exactly when your best customers are in.

"Every restaurant we visit has had a loyalty scheme at some point. Almost none of them stopped because the reward was wrong. They stopped because earning required somebody to remember something on a busy night."

— Anzul Aqeel, founder of Get Menu

What removes all four at once

Put the card in the phone the customer is already holding.

A digital card lives in the phone's own wallet with no app to install, using Google's own loyalty card format. It cannot be left at home, it shows the count without anybody asking, and it updates itself when a stamp is earned.

Then make earning automatic. Stamps that accrue from the order itself, rather than from a member of staff pressing something, remove the last dependency. Restaurant loyalty cards that work this way are still working in month six, which is more than can be said for most.

How to set it up in practice

  • Choose what earns. A stamp per order, per dirham spent, or only on the dish you want to move more of. That third option is the most underused: it is a way to shift a specific item rather than to reward everything equally.
  • Set the first goal close. Something reachable in a few visits, so the scheme proves itself before the customer forgets it exists.
  • Add a tier above it. Regulars grow into it and it carries its own automatic discount, applied without a code and without a conversation at the counter.
  • Never stack. If a coupon and a tier discount both apply, the better one wins. Two automatic discounts landing together is how a promotion becomes expensive by accident.
  • Give it an expiry. A reward that never expires is a liability that grows silently on your books.

Where it sits in the rest of the setup

A loyalty rewards program for restaurants works best when it is fed by the ordering itself, not bolted on beside it:

  • Your own ordering page, with 12 page layouts so the menu reads the way you want it to.
  • A QR menu for the tables and the window, printed from a designer rather than a generic sticker.
  • Table routing. The code carries the table, so nobody writes a number down.
  • A menu you edit yourself, with options and add-ons carrying their own prices, and required choices the kitchen can trust.
  • Orders that price themselves. Every line is re-priced from your live menu at checkout.
  • Messages nobody types. A confirmation to the customer the moment they order, a ticket to the kitchen, and for delivery a link carrying the GPS pin dropped at checkout. That is 3 messages in about 2 seconds.
  • Live stock on items and on options, greying out what has gone.
  • One bill for a group, with everybody ordering from their own phone.
  • A wallet card in the customer's phone, with a designer for the card's colours and logo so it looks like your restaurant rather than a template.
  • VAT invoices carrying your TRN, downloaded as a PDF without the customer logging in.
  • Coupons, upsells and bundles with caps, exclusions and expiry dates.

Where a restaurant loyalty system usually goes wrong on day one

Two setup decisions cause most of the disappointment, and both are made in the first hour.

Rewarding everything equally. If every dirham earns the same, the scheme is a small general discount rather than a lever. Choose what you want more of, and weight the earning towards it. A dish that is profitable and under-ordered is the obvious candidate.

Launching silently. A scheme nobody is told about earns nothing for months. The launch is not a poster. It is the first order after it goes live, where the customer sees a card appear and understands immediately what it is counting.

Get those two right and the mechanics above will carry themselves.

Measuring whether it is working

Two numbers, checked monthly.

Repeat rate. What share of customers ordered more than once in the period. If a loyalty scheme is doing anything, this moves.

Time between visits. If the gap is shrinking, the scheme is pulling people back sooner, which is the actual mechanism.

Ignore the number of cards issued. It measures how many people were handed something, not whether anybody came back, and it is the metric most likely to make a failing scheme look healthy.

Our trial runs 15 days with no card, which is enough to set one up and watch a first cycle. Pricing is on the pricing page rather than in this article. Payment is cash or your own card machine at the door or the table, with no online payments in the system, so no card details are held in it and old order details clear automatically after 120 days.

Frequently asked questions

Stamps, tiers or points?

Stamps for frequent similar purchases like coffee. Tiers for less frequent larger visits. Points last, because customers cannot picture where they are.

Do customers need an app?

No. The card is added to the phone's own wallet, which is already on the device. Nothing to download and nothing for staff to hand over.

How many stamps should a reward take?

Close enough to feel reachable for somebody visiting at your normal frequency. If your regulars come twice a month, a goal ten visits away is invisible to them.

Can a tier discount stack with a coupon?

It should not, and in our system it does not. The better single discount applies, which stops a promotion becoming expensive by accident.

What is the one number to watch?

Repeat rate, not cards issued. Cards issued measures handing something over; repeat rate measures whether anybody came back.

Further reading

Keep reading

More on this

Your own ordering page, your own customers

Fifteen days with everything switched on. No card, no commission, ever.

Questions from other owners are answered in the community.

Chat with us