Somewhere near your register is a stack of half stamped punch cards nobody will ever finish. A few are in wallets that got thrown out months ago. One regular is convinced you owe her a free coffee and there's no way to check. That's the loyalty program at most restaurants, and it runs entirely on the honor system, in both directions.
It's not that punch cards are a scam. They're just a piece of paper trying to do a job that needs a memory, and paper doesn't have one. It can't tell you who your top twenty customers are. It can't nudge someone who hasn't ordered in three weeks. It can't apply a reward automatically at checkout, so it depends on a server remembering to stamp it, which is the first place these programs quietly die.
A card in a drawer can't text anyone
The actual job of a loyalty program is simple: recognize someone who's already spent money with you, and give them a reason to come back sooner than they otherwise would. Paper can do the recognizing, sort of, as long as the customer remembers to bring the card and the staff remembers to check it. It can't do anything with that information afterward. No list of regulars to message when you launch a new item. No way to spot the customer who used to order weekly and quietly stopped. No record at all once the card is lost, which happens to most of them within a few months.
A digital version doesn't have to be complicated to fix this. It just has to attach a name, phone number, or account to every visit, so the recognizing part survives past the moment the card gets shoved in a wallet.
Stamps, points, or tiers: pick based on how often people already return
There isn't one right mechanic. There's a right one for how your customers already behave.
| Mechanic | Right when | The catch |
|---|---|---|
| Digital stamp card | Visits are frequent and roughly the same price (coffee, lunch counter, a regular order) | Needs a low, fast reward or people stop caring after visit two |
| Points per dollar | Ticket sizes vary a lot and you want bigger orders rewarded more | Needs a point value people can actually do math on in their head |
| Tiered or VIP | You have a real group of high frequency regulars worth treating differently | Only works with real perks at the top tier, not just a badge |
| Paper punch card | You know every regular by name already and just want a small thank you | No data, no messaging, and it dies the day the card gets lost |
Most restaurants overthink this step. If people already come in two or three times a month, a simple stamp style reward beats a points system nobody wants to calculate.
Start tonight, before you buy anything
This costs nothing and takes fifteen minutes to set up.
- Keep a notebook or a phone note by the register.
- For the next twenty customers who pay, ask if they'd like a text when their usual order's ready next time, and write down the number if they say yes.
- At the end of the week, look at who's on that list and who ordered more than once. That's your actual regulars list, built from real behavior instead of a guess.
- Text three of them yourself. Not a blast, just "hey, saw you're due for your usual, come by this week." See who shows up.
That list is worth more than a drawer of punch cards, because you can actually reach every name on it. Plenty of restaurants run on a version of this for months before ever paying for software, and there's nothing wrong with that.
The reward has to trigger itself, or it won't
Here's the detail that decides whether a loyalty program survives its first month: does the reward depend on a person remembering, or does it just happen. A punch card needs a server to notice, reach under the counter, and stamp it, on a night when the kitchen's slammed and the till has a line. It gets skipped constantly, not out of carelessness, just because it's one more manual step competing with everything else.
Loyalty built into an ordering flow doesn't have that problem. The points or stamp count updates the moment the order's placed, and the reward applies itself at checkout without anyone having to remember it exists. That's the whole advantage of running loyalty inside the same app people already use to order, the way Eatz handles it: earn and redeem happen automatically on every order, so the program keeps working on the nights nobody has time to think about it.
When a loyalty program is solving a problem you don't have
An honest boundary: if you run a small counter spot and already know your regulars by their coffee order, a formal program isn't buying you much. You're already doing the recognizing part in your head, and a notebook list like the one above covers the rest.
It starts earning its keep once there's enough volume or enough staff turnover that nobody could realistically hold all the regulars in their head, multiple servers, a second location, or an order volume where repeat customers are genuinely hard to spot by eye. At that point, tying rewards to the ordering system itself beats a side app or a paper card, because the reward and the order happen in the same place.
One thing loyalty won't do: fix a restaurant people don't want to return to for other reasons. It rewards people who were already coming back. It doesn't create a reason to.
Either way, the mechanics matter less than whether the reward actually triggers when it's supposed to. Plans for running loyalty inside your ordering app are on the pricing page, and if the bigger question is why you'd want your own ordering channel at all, start with what owning your customers actually looks like.
Frequently asked questions
Do restaurant loyalty programs actually work?
Yes, for a specific job: giving people who already like you a reason to come back a bit sooner. They don't work as a substitute for good food or service, and they don't create new customers on their own.
What's the simplest loyalty program for a small restaurant?
A notebook. Capture a phone number at checkout for a few weeks, see who orders more than once, and text a handful of them yourself. No software needed until that list gets too big to manage by hand.
Should loyalty be a stamp card, points, or tiers?
Match it to how customers already behave. Frequent, similar priced visits fit a stamp card. Wide ranging ticket sizes fit points per dollar. A genuine group of high frequency regulars fits a tiered program with real top tier perks.
Do I need an app for a loyalty program?
Not to start. You need one once the reward has to trigger reliably at checkout without a server remembering, which is where paper programs quietly fail after a few months.
How much should a reward cost to give away?
Enough that it feels worth chasing, small enough that you're not discounting your best customers into a loss. A reward around the price of your cheapest popular item is a reasonable starting point.