Advice about loyalty programs is mostly written about businesses with a marketing department. Some of it transfers to a shop with one till. A lot of it does not.
What you have that chains do not
Staff recognise people. This is the single biggest advantage and it is worth designing around. A chain's loyalty program has to work without any human recognising anyone. Yours does not.
The practical effect is signup rate. A member of staff who says "you're in here most Fridays, are you collecting stamps?" converts at a rate no poster achieves. Chains cannot do that at scale. You can, for free, today.
You can change the program on a Tuesday. No committee, no rollout, no legal review. If a threshold is wrong you can move it this week and see the effect within a fortnight.
Your rewards feel personal. A free coffee from a cafe where the owner knows you lands differently from the same coffee from a chain. That is not sentiment, it shows up in whether people mention you to friends.
What you do not have
Volume to absorb mistakes. A chain testing a reward across two thousand sites can afford for some of them to be wrong. If your threshold is badly set, it is wrong for every customer you have.
Breakage as a business model. Large programs count on a proportion of rewards never being claimed. At your scale, an unclaimed reward is not a saving, it is a customer who lost interest. Design for redemption, not against it.
Anyone to run it. There is no loyalty manager. Whatever you build has to survive being ignored for three weeks during a busy period.
Design choices that follow from that
Keep one mechanic
Chains run points plus tiers plus a subscription plus partner offers, because they have people to manage each. You should run one thing a customer can describe in a sentence.
If you cannot explain your program at the till in under ten seconds, it is too complicated for the environment it has to live in.
Set the threshold from your real visit frequency
This is where most small business programs go wrong, and it is arithmetic rather than judgement.
Look at how often a regular actually comes. Then set the threshold so they finish in six to eight weeks.
A cafe regular visiting twice a week: ten stamps takes five weeks. Fine. A barber's customer visiting monthly: ten stamps takes ten months. Not fine. Use six, or four. A salon client visiting every six weeks: use points on spend, or a reward at three appointments.
Copying a cafe's ten stamp card into a business with monthly visit frequency is the most common mistake in this category.
Make joining faster than making the coffee
Time it. If signup takes longer than the transaction it is attached to, staff will stop offering it during exactly the periods when the most customers are in front of them.
A QR code the customer scans themselves, joining in the browser with no download, is roughly twenty seconds. An app install is two minutes and a lot of people will decline. This is a large part of why browser based loyalty cards took over from apps for independents.
Put the code where the customer already looks
Next to the card machine, not on the wall behind you. Customers look at the payment terminal. They do not look at your wall.
Give staff the sentence, not the brief
Write down the exact words. Not "let customers know about the loyalty scheme" but:
"Are you collecting stamps? Ten gets you a free coffee, takes ten seconds."
Then check the signup rate weekly for a month. A rate that stalls is almost always about whether the sentence is being said, not whether customers want the offer. Training staff to bring it up naturally goes further on this.
By trade
Cafes and coffee shops. Stamps, nine or ten, on drinks only. Add double stamps on your quietest morning. This is the most solved problem in loyalty and there is no reason to be creative.
Hair and beauty. Spend based, because a fringe trim and a full colour should not earn the same. Add a birthday reward. Your booking system already knows the date.
Bakeries. Stamps on a category rather than the whole shop, otherwise a £1 roll earns what a £30 cake earns.
Barbers. Six stamps, not ten. Monthly visits make a ten stamp card a year long project.
Takeaways and food trucks. Spend based, with the reward pitched slightly above your average order so redeeming pulls the basket up.
Independent retail. Points, with a reward that is a product rather than a discount. Discounts train customers to wait for discounts.
Gyms and studios. This is one of the few independent categories where tiers genuinely work, because the relationship is long and members already think in terms of commitment.
The first ninety days
Weeks one and two. Sign up existing regulars only. No advertising. You want a real signup rate and you want the first redemptions happening while you are standing there to see what goes wrong.
Weeks three to six. Watch the signup rate. Fix the ask if it is under about a quarter of transactions. Do not touch the reward yet.
Weeks seven and eight. The first cards complete. Watch how redemption feels during a rush. This is where operational problems surface.
Weeks nine to twelve. Now the numbers mean something. Compare visits per member before and after joining. Check what percentage of member spend you gave back. Adjust the threshold, not the reward.
What to do when it goes quiet
Every program has a dip around month four. Members have joined, some have redeemed, and the novelty is gone.
The fix is not a bigger reward. It is contact. Send one message to members who have not visited in six weeks, offering a reason to come in this week. That single action recovers more customers than any change to the card, and it is the reason the customer list is worth more than the stamp mechanic.
If you have not built the program yet, how to create a loyalty program is the step by step version. If you are choosing between structures, start with loyalty program examples.