Restaurant loyalty program: what actually brings people back

Aug 21
3 min read

Updated: 2 days ago

Introduction

Restaurants adopt loyalty schemes more often than almost any other business, and abandon them just as often. The pattern is familiar: cards printed, staff briefed, enthusiasm for a month, then a drawer full of half-stamped cards nobody mentions.

The schemes that survive have two things in common. They are simple enough to explain in a sentence, and they were costed against food margin before launch.

1. A restaurant loyalty program has to fit how people actually eat

Match the mechanic to your visit pattern, not to what sounds generous.

Frequent, low value — a café, a lunch spot. Visit-based stamps work: buy nine, get one. People can count it, staff can run it, no technology required.

Occasional, higher value — a dinner restaurant visited monthly. Stamps are too slow to feel achievable. Spend thresholds or a members' benefit suit better.

Delivery-heavy — the platform usually owns the customer, so a scheme that only works in-house is really an incentive to order direct, which is a different and often better goal.

Choosing the wrong shape is why most schemes fail before the reward is ever earned.

2. Cost the reward against food margin, not the menu price

A free main course is not "worth" its menu price to you — it costs its ingredients.

A dish selling at $18 with $6 of food cost costs you $6 to give away. If the scheme requires nine visits at an average $24 spend and a 70% food margin, you have earned around $151 and spent $6. That works comfortably.

Run the same arithmetic with a $40 average spend, a premium reward and a thinner margin and it may not. Do it before printing anything, because it tells you how many visits the reward should require.

3. Reward the behaviour you actually want

Most restaurant schemes reward visit count by default. That is not always the problem worth solving.

If your issue is quiet weeknights, reward visits on those days specifically. If it is low spend per head, reward spend thresholds rather than visits. If it is that first-time diners never return, weight the reward toward the second visit — which is where the drop is steepest.

Decide which number you are trying to move before choosing what earns a stamp.

4. Keep it explainable in one sentence

If a diner cannot state the benefit while waiting for the bill, it will not change their behaviour, however generous it is.

"Buy nine coffees, get the tenth free" works. "Earn points redeemable against selected items subject to availability" does not — not because it is worse value, but because nobody carries it in their head.

Complexity, not insufficient generosity, is the most common cause of failure.

5. Make progress visible

People act on progress they can see. A card showing seven of ten stamps is a genuine motivator; a points balance nobody checks is not.

Physical cards have an advantage digital systems often lose: they sit in the wallet, need no login, and remind the customer every time they pay for something else. For a single-site restaurant that is usually enough.

6. Make sure staff mention it

A scheme nobody brings up does not exist. Enrolment depends almost entirely on whoever is taking payment saying one sentence.

Put it in the till routine, give staff the exact wording, and check occasionally that it is happening. This is a documented step with a named owner, like any other part of the operation — not a hope that people remember.

7. Run it as a test first

Before committing across every shift, run the mechanic with one defined group — one location, one service, one half of the week — and compare repeat visits against everyone else over a defined period.

If the number moves, expand it. If it does not, you have learned that for the price of some printed cards rather than a year of discounted covers.

Conclusion

Match the mechanic to how often people actually visit, cost the reward against food margin rather than menu price, and reward the specific behaviour you want more of.

Keep it explainable in one sentence, make progress visible, make sure staff mention it, and test it against a comparison group before rolling it out. A simple scheme that is measured beats a sophisticated one that is merely running.