There are six loyalty program structures in common use. Each one behaves differently at the till, costs you differently, and suits a different kind of business.
Stamp and punch cards
The customer collects one mark per qualifying purchase and gets a free item when the card fills.
To a customer this is instantly obvious. There is no conversion rate, no balance to check and no maths. A ten stamp card with seven filled is a progress bar you can read at a glance. At the till there is almost nothing to do: scan or stamp, hand the drink over.
The cost is one free item per completed card, which at ten stamps on similarly priced items works out at roughly 10 percent of value, less if the free item is cheaper than the average order.
It breaks on uneven basket sizes. If a £2 purchase and a £40 purchase both earn one stamp, you overpay the small orders badly and underpay the large ones.
Best for cafes, bakeries, sandwich shops, car washes, barbers, nail bars and takeaways.
Points programs
Points accumulate per pound spent and convert to a reward at a threshold.
This feels fair to customers but stays abstract. Nobody knows what 340 points is worth without being told, so the reward is always one mental step away. At the till it is fine when software handles it and painful when a person is doing arithmetic during a rush.
What it costs is entirely up to your exchange rate, which is both the advantage and the trap. It is easy to set a rate so conservative that customers never reach anything, at which point you have built a database rather than a loyalty program. The test is simple: if a customer cannot tell you roughly what their balance is worth, it is not motivating them.
Best for salons, garages, veterinary practices, independent retailers and pet groomers.
Tiered programs
Customers move through levels, bronze to silver to gold, and each level unlocks better perks permanently.
The pull here is status. Partly the perks, and partly not wanting to drop back down. Airlines built the modern version and it works remarkably well on people who insist they are not motivated by status.
Whether it does anything depends on staff being able to see the tier. If they cannot, the tier is invisible, and recognition was the whole reason the customer climbed. The cost sits in design and admin more than in rewards, because each level needs enough perks to feel different from the one below.
It breaks on small customer bases. When most members sit in the bottom tier because nobody visits often enough to climb, the structure demotivates.
Best for gyms, wine clubs, members' spaces and subscription retail.
Paid membership
The customer pays a recurring fee for ongoing benefits. Pret's coffee subscription is the version most people have met.
Having paid, a customer walks past your competitor to use what they bought, which is exactly the point. The economics invert too: money arrives up front and the cost is the margin your heaviest users consume. Model your top decile rather than your average, because the average tells you nothing useful here.
It breaks on thin margins and small regular bases. You need enough people visiting often enough to justify the fee, and enough headroom to absorb the ones who come daily.
Best where frequency is high and margin allows it, which in practice means coffee above all.
Cashback and store credit
A percentage of each purchase returns as credit spendable with you.
Customers find this reasonable and forget it immediately. It reads as generous because a percentage sounds larger than it is, and it produces no moment worth telling anyone about. A £1.60 credit balance does not create a story. Free coffee does.
The upside is forecasting. It costs exactly the percentage you set, which makes it the easiest structure to budget for.
Best for retailers with a broad product range where a fixed free item makes no sense.
Coalition programs
Several businesses share one scheme, so points earned at one can be spent at another. Nectar is the obvious UK example.
Customers like it because one card covers many places and points accumulate faster. You pay a fee to the operator, and you pay something that matters more than the fee: you stop owning the relationship. The customer becomes loyal to the scheme rather than to you, and the scheme decides what your points are worth.
For an independent this almost always breaks on control. You cannot message your own members, you cannot change your own reward, and if the operator changes the economics you absorb it.
Best for larger chains, petrol forecourts and supermarket groups.
Mechanics you bolt on
These are not programs on their own. They sit on top of one.
A referral reward pays an existing member when someone they bring joins or spends. It is the only mechanic here that grows your customer base rather than deepening it.
A birthday reward costs almost nothing and produces a disproportionate amount of goodwill, because the customer knows you remembered.
A bonus multiplier on a chosen day moves traffic rather than increasing it. Double stamps on Tuesday shifts some of Saturday's queue into midweek, which is worth more than it looks if Saturday is already at capacity.
A welcome reward given at signup converts browsers into members at the counter, because there is something in it immediately rather than in six weeks.
Choosing between them
Two questions settle it for most businesses.
Is your average transaction roughly consistent? If yes, stamps. If no, points, or stamps restricted to one product category.
Do you have hundreds of customers or thousands? Under about a thousand active customers, tiers and coalitions will not have enough people in them to work. Stay with a single simple mechanic and add bolt ons.
Then pick a threshold your typical regular can reach within six to eight weeks, which matters more than the structure you picked. How to create a loyalty program works through that calculation, and loyalty program examples shows what each structure looks like in a real shop.