In brief
Reward with something you make, not something you buy. A made item costs you its ingredients; a gift card costs you its face value, and the customer values both about the same.
The short answer
Reward loyal customers with something you produce yourself, at a threshold that generates far more margin than the reward costs. A cafe should give coffee. A salon should give a treatment. Neither should give a gift card, because a gift card costs its face value while a flat white costs eighty pence, and the customer is roughly as pleased either way.
Everything else in this post is how to check that the specific reward you have in mind passes.
The margin test
One calculation, before anything else.
Work out the gross margin on a typical visit, multiply it by your threshold, and compare it to what the reward costs you at cost of goods. That ratio is the only number that decides whether a reward is affordable.
| Business | Margin per visit | Threshold | Margin earned | Reward at cost | Reward as % |
|---|---|---|---|---|---|
| Cafe, free coffee | £2.40 | 10 | £24.00 | £0.80 | 3.3% |
| Barber, tenth cut free | £24.00 | 10 | £240.00 | £4.00 | 1.7% |
| Retailer, free £50 item | £20.00 | 10 | £200.00 | £30.00 | 15% |
| Anyone, £20 gift card | £2.40 | 10 | £24.00 | £20.00 | 83% |
Under about 10% and the reward is comfortable. Over about 25% and you are running a discount scheme that happens to have stamps on it.
Notice which businesses have room. Hospitality and services have enormous headroom because the thing they give away is mostly labour they were paying for anyway. Retail has much less, which is why “tenth item free” is rare in shops and why retail programs tend to be points-based or built around access rather than product.
If your reward fails the test, the fix is usually the threshold, not the reward. Twelve stamps instead of ten moves the retailer above from 15% to 12.5%. Fifteen moves it to 10%.
Four classes of reward
1. Something you make
The default, and the right answer for most local businesses. High perceived value, low real cost, and it brings the customer back into the shop to collect it — where they usually buy something else.
The one rule: pick something you can hand over on your busiest shift without checking. A reward that requires the good machine, the trained person or the Tuesday stock is a reward that will be refused on a Saturday.
2. A discount
Weak, and worth understanding why. A percentage off is abstract, gets applied to a purchase the customer was making anyway, and trains people to wait for it. It also fails the margin test faster than it looks: 20% off a £25 basket at 40% margin costs £5 against £10 of margin.
Discounts do have one good use — as an unlock rather than a reward. “Members get the Tuesday price” is access, not a discount, and reads completely differently.
3. Access
Something not available to everyone: early booking, the new thing before it launches, a seat held back, the ability to skip a queue. Costs close to nothing, and it is the only reward class that gets more valuable as you get busier.
Underused by local businesses, and the one to reach for when the margin test kills the free-product idea.
4. Status
A named tier, a recognised regular, the order remembered. This is the reward that costs nothing and is hardest to fake, because it requires the business to actually know the person.
It is also the only one on this list that does not scale. That is fine — it is worth more than the others precisely because it does not.
What to never give away
Cash equivalents. Gift cards, credit, cashback. They cost you face value, and they anchor the relationship on price. The moment a competitor’s card is worth more, the loyalty was never yours.
Your scarcest item. Rewarding with the thing you sell out of every day means the reward fails at exactly the moment the customer came in for it.
Anything with a manager approval step. A reward that needs someone senior is a reward that fails on the shift where it matters.
Something you would not offer twice. If honouring the reward stings a little, it will show, and the customer will read the wince.
Rewarding loyalty versus buying a visit
These get treated as the same thing and are not.
A reward that arrives at a threshold buys visits — it is a mechanism, and it works, and its cost is on the budget. A reward that arrives unannounced recognises loyalty — a free pastry for the customer who has been in every Tuesday for a year, given because you noticed.
The second one costs almost nothing, is not gameable, cannot be compared to a competitor’s offer, and is what people actually tell their friends about. It also cannot be the whole program, because it does not scale past the customers you personally recognise.
Run both. The threshold reward is the machine; the unannounced one is the thing the machine frees you up to do.
How to increase loyalty without a bigger reward
When a program is not working, the instinct is to make the reward bigger. That is usually the wrong lever, and it is the expensive one.
Cheaper levers, roughly in order of effect:
- Lower the threshold. Getting the first reward into someone’s hands sooner does more than making it larger. Nobody is loyal to a card they have never completed.
- Make the balance visible. A customer who can see they are two stamps away behaves differently from one who has lost count. This is most of the real advantage digital cards have over paper.
- Ask at the counter, every time. Enrolment is a staff habit, not a marketing campaign. Programs stall at the ask far more often than at the offer.
- Message the people who stopped. A customer who was on visit seven and disappeared is the highest-value message you can send, and the only one worth sending. There is a worked version in winning back customers who stopped coming.
- Fix the redemption. If claiming the reward is awkward, the ninth visit is where the program ends.
Only after all five of those should the reward get bigger. And when it does, add a second reward earlier in the card rather than inflating the one at the end — the arithmetic in what a loyalty program actually costs is the same either way, but a reward at stamp three does more work than a bigger one at stamp ten.