Ten. That is the Kanz default, and it is a sound starting point for most businesses. But it is not a convention: the number of stamps is a discount in disguise, and it can be worked out.
Here is the sum, in one line. If your reward costs you C and you hand it over every N
visits, you are granting a discount of C / N per visit. A reward at 6 that goes out every
10 visits is 0.60 of discount per visit. Compare that figure with your margin on a visit,
not with the price of the reward: that is where the decision gets made.
The amounts in this article are in your own currency, whatever it happens to be. Only the ratios matter.
What the number of stamps really tells your customer
Too low, and you pay for visits that would have happened anyway. Three stamps for a reward, in a business where people already come twice a week, is a permanent discount for regulars with no change in behaviour bought in exchange.
Too high, and the card turns invisible. A reward at thirty visits, on a monthly rhythm, is two and a half years. Nobody builds a habit on a horizon they cannot see. Kanz caps the setting at 30, as it happens, and floors it at 1.
The right number is the one that puts the reward in sight without being within reach. Two constraints, and both have to hold.
In visits: between six and twelve. Below six, the reward lands before the habit does. Above twelve, the counter reads as a distant target rather than as progress.
In time: the whole cycle should stay under roughly six months at the normal rhythm of your customers. It is this second constraint that bites when they come in rarely. A weekly customer at ten stamps has their reward in ten weeks, which is perfectly fine. A monthly customer at ten stamps waits ten months and gives up long before that: for them, six is already a ceiling, not a floor.
Which means that if your customers keep two very different rhythms, one number cannot be right for both. Choose for the ones you want to keep, not for the average.
The 40% rule, and why it lives in the code
There is a second kind of reward, and it is not on the card: the gift you offer to bring back somebody who has stopped coming. There the question is not "how many stamps" but "how much am I willing to spend to buy a visit back".
Kanz answers with a hard limit: never more than 40% of the value of a visit. If a customer typically spends 20 with you, the gift meant to bring them back cannot exceed 8, however generous you feel at the time.
The reason is arithmetic. Above that threshold the visit you bought back costs more than it brings in, and the follow-up becomes a machine for losing money, all the faster the better it works. A ceiling expressed as a percentage of the visit, rather than as a fixed amount, adjusts itself between a customer at 5 and a customer at 80.
A second rule in the same spirit: among the eligible rewards, Kanz picks the cheapest one that stays credible, never the most generous one available. Software that always reaches for the biggest gift within its grasp is not optimising your return, it is optimising its own success rate with your money.
A reward that has been earned is never given away
That is the third rule, and it is the one protecting the card itself.
The rewards on your card, the ones that ask for a number of stamps, are excluded from follow-up gifts. Technically, a reward that declares a required number of stamps can never be selected by a follow-up, and there is no setting to lift that.
Without the rule, a customer who has patiently collected seven stamps receives a message offering them the reward they were halfway through earning. The gift pleases them for five minutes and destroys seven visits of effort, along with the credibility of the counter for every cycle that follows. A stamp card is only worth something if nobody can jump the queue, software included.
The stamp is not the reward
One last thing, often misread, and it counts for more than the exact number.
What brings a customer back is not the reward on the tenth visit. It is the counter on the third. A card sitting at three stamps out of ten creates a small debt to yourself that you do not like walking away from, and it is that debt that produces the extra visit, not the reward at the end.
That has two practical consequences. The first is that the counter has to be visible without effort: on a card in the phone wallet it is seen in passing, which a piece of card at the bottom of a bag is not. The second is that a first stamp given at sign-up is worth more than one extra stamp at the end: starting at 1 out of 10 is a start, starting at 0 out of 10 is a queue.
The sensible starting point
If you have no data, take ten, a reward costing less than half an average basket, and let it run for two months. You will then know two things nobody could have told you in advance: how many customers actually reach the reward, and at which stamp they drop out.
How the card works is described in another article, and what all of it costs in a third.