Most Restaurant Loyalty Programs Are Just a Discount for People Who Already Liked You

Here's the uncomfortable math on most loyalty programs.

Your regulars were coming anyway. You now give them 10% off, or a free item every tenth visit. You didn't buy a new customer — you bought a discount on the customer you already had. Meanwhile the person who came once and never returned, the one the program was supposed to win back, never signed up in the first place.

That's not a reason to skip loyalty. It's a reason to build it so it targets the behavior you actually want to change.

What a loyalty program is really for

There are only four things a program can do, and most operators are running one that does the first while hoping for the others:

• Reward existing loyalty — What it looks like: Points, punch cards, free tenth item · Who it targets: Your regulars

• Increase visit frequency — What it looks like: Time-boxed offers, streaks · Who it targets: People who come occasionally

• Increase ticket size — What it looks like: Thresholds, bundles, add-on rewards · Who it targets: Everyone at the register

• Win back lapsed customers — What it looks like: Targeted outreach after an absence · Who it targets: People who stopped coming

The last one is where the money is, and it's the one almost nobody runs. A customer who came three times and then vanished is your single best prospect — they've already decided they like the food. Getting them back is cheaper than any ad you'll ever buy.

Which means the most valuable thing your program collects isn't points. It's knowing who stopped coming.

Design the thing you can actually run

The best loyalty program is not the most generous one. It's the one your staff can explain in one sentence during a rush, and that you'll still be running in a year.

Rules I'd hold to:

• One sentence, no asterisks. If a customer at the window can't understand it in five seconds, you've lost the transaction time it takes to explain and they still won't sign up.

• Signup has to cost nothing but a phone number. Every extra field cuts enrollment. You don't need their birthday to start.

• Make the reward worth returning for. A 5% discount is not a reason to change your Tuesday plans. A free item people actually want is.

• Cap your exposure. Know what it costs you if every single member redeems. If that number scares you, the reward is too big or the threshold too low.

• Set an expiry. Points that never expire are a liability that grows forever and lose their urgency.

The reward that must never be on your list

I want to be direct about this one because I've seen it recommended constantly and we nearly got it wrong ourselves.

Never reward a customer for leaving a review. Not points, not a free drink, not a discount, not entry into a drawing. Google's policies prohibit incentivized reviews, and the downside isn't a slap on the wrist — it's your review profile, which is the single most valuable marketing asset an independent restaurant has. Risking it to juice a star rating is a genuinely bad trade.

What's fine: asking. A QR code at the register, a line on the receipt, staff who ask happy customers directly. That works, it's allowed, and it's how real review counts get built.

Also fine: rewarding things that aren't reviews. A tagged photo, a check-in, bringing a friend, signing up for texts. Reward the behavior, not the rating.

A program worth actually running

If I were starting one tomorrow, this is the shape:

1. Capture the phone number at the register. That's it. No app to download, no card to carry and lose.

2. Give a reward that lands on visit two, not visit ten. The gap you're trying to close is between first visit and second, because that's where most customers are lost forever. A tenth-visit reward is a gift to someone who was already a regular.

3. Then use the absence. Thirty days without an order from someone who used to come weekly is a signal. That's your one text, with a real reason to return.

4. Measure one number: repeat rate. What share of first-time customers come back within 60 days? If your program isn't moving that, it's a discount, not a loyalty program.

Step 4 is the part that gets skipped, and it's the only way to know whether any of this worked.

The honest caveat

Loyalty programs are not a fix for a slow week, and they don't work on a menu people don't want to return to. They amplify an existing reason to come back. If the food or the speed is the problem, a rewards program is an expensive way to avoid fixing it.

They're also slower than they look. You're building a list, and a list is worth very little at 40 people and quite a lot at 4,000. The operators who get value from this are the ones who kept collecting numbers for a year while it felt pointless.

And there's a real failure mode: the discount treadmill. Run offers often enough and you train your regulars to wait for one. If your customers now only come when there's a deal, that's not loyalty — you've repriced your menu and told yourself a story about it.

Where doing this manually falls apart

You can run all of this on paper. Punch cards work. A notebook of phone numbers works.

Where it breaks is step 3, because "who used to come weekly and hasn't been in for thirty days" isn't a thing you can see in a notebook, or in a spreadsheet you haven't updated since March. That question requires knowing every customer's order history and being able to sort it — which means it has to live next to your orders.

That's the part we built. On Avocado, loyalty points, customer profiles, and full order history sit in the same system as your sales, so you can ask which customers haven't ordered in thirty days, who your top customers are, and what your average order value looks like per customer — then text that group directly. Rewards and retention texts are part of the product, not a bolt-on you sync.

Start with the phone numbers, though. Even in a notebook. The list is the asset; the software just makes it useful.

Avocado is a POS built for independent restaurants and food trucks. We're operators first, and we'd rather tell you the honest version than the version that closes faster.