What a stamp card is actually for
A loyalty card is not a discount. It is a reason to come back on a Tuesday, and the difference between those two ideas is most of your margin.
Most stamp cards are run as a discount with extra steps. Buy nine coffees, get the tenth free, and the owner quietly books a ten percent cut on every regular they already had.
That is a real cost, and it buys nothing on its own. The regular was coming anyway. What the card is for is the customer who would have come four times this month and comes six instead.
Key takeaways
- A stamp card only pays when it changes behavior. Rewarding visits that were already going to happen is a discount, and it comes straight out of margin.
- Work the card from gross profit, not from the sticker price of the reward. At £4 of margin a visit and a £1.20 reward, one extra visit per nine covers the cost with £2.80 left over.
- The card is worth most on the customer who visits occasionally, and worth least on the daily regular who needs no reason to return.
- The card is not the loyalty program. The card is the delivery address. What it really buys you is a way to reach the customer after they leave.
The only number that matters
Work it from the margin, not the discount.
If a visit is worth £4 in gross profit and the tenth coffee costs you £1.20 in goods, the card pays for itself the moment it adds a single extra visit per nine. One extra visit buys £4; the reward costs £1.20.
| Visits without a card | Visits with a card | Reward cost | Net |
|---|---|---|---|
| 9 | 9 | £1.20 | −£1.20 |
| 9 | 10 | £1.20 | +£2.80 |
| 9 | 12 | £1.20 | +£10.80 |
The first row is the one to avoid, and it is what happens when a card is only ever handed to people already holding one.
Notice what the arithmetic does not depend on. It does not depend on the reward being cheap. It depends on the reward being smaller than the margin on the visits it creates.
The same sum in other trades
The ratio moves a lot by trade, and it is worth doing on your own numbers before picking a card length.
| Trade | Margin per visit | Reward cost | Extra visits to break even |
|---|---|---|---|
| Coffee shop | ~£4 | £1.20 | 0.3 |
| Bakery | ~£3 | £2.00 | 0.7 |
| Takeaway | ~£8 | £6.00 | 0.8 |
| Barber | ~£18 | £18.00 | 1.0 |
Read the last column as the extra visits, per completed card, that the reward has to buy before you are square. It is the reward divided by the margin, and nothing else.
The barber row is the interesting one. A free haircut costs a full visit’s margin, so the card must generate a whole extra appointment just to break even. The coffee row needs less than a third of one. That is an argument for a shorter card at the barber, not a cheaper reward.
Who the card is actually for
This is where most programs lose money, and it is a targeting problem rather than a math problem.
A customer who comes in every weekday morning has already decided. A card changes nothing about their behavior, and every stamp is a small subsidy on a visit you had for free. The card is close to pure cost on that person.
The customer worth the card is the one who comes in most weeks, likes the place, and has no particular reason to pick you over the shop nearer their office. That person is movable. Generally the more occasional the customer, the more a card is worth on them.
That does not mean refusing your regulars a card. It means not counting them as the return, and not designing the whole program around the people who need it least.
Give it time to show up
One caution before you judge the program on a month of data.
Studying 322 firms that introduced a loyalty program between 2000 and 2015, Chaudhuri, Voorhees and Beck found sales and gross profits rose within the first year and the effect held for at least three. The part worth knowing is the lag: gross-profit effects did not become statistically significant until the second quarter after launch. Their sample is publicly traded companies rather than independent shops, so treat the timing as a caution rather than a forecast.
The practical version is that a card killed after six weeks is a card abandoned before the thing you were looking for would have appeared. Give it a couple of quarters, and judge it on repeat visits rather than on how it feels at the counter.
Why paper loses
A paper card has no idea who is holding it, so it cannot do the one thing that changes the second row into the third: say something on a quiet Tuesday.
A wallet card can. It sits on a phone the customer already unlocks dozens of times a day. It can put a line on the lock screen when the shop is empty, rather than when it is full.
The card is not the loyalty program. The card is the delivery address.
That is the part worth sitting with. The stamps are a mechanism for getting someone to identify themselves and agree to hear from you. What you do with that afterwards is where most of the value is, and a paper card throws it away at the counter.
What to do with it
Three things, in order:
- Offer it to everyone, but count on the newer customers. Hand a card to anybody who wants one. Just do not treat your existing regulars as the return, because they were already coming.
- Set the reward where the math works, which is usually further out than feels comfortable. The arithmetic for card length is here.
- Use the notification sparingly. One well-timed message beats four ignored ones, and the fourth is what gets the card deleted.
None of that needs a POS integration or a tablet on the counter. It needs a card people actually keep, and a reason to send them something worth reading.
It is also worth knowing how concentrated this usually is. Toast and Resy, looking at restaurant transactions in the first quarter of 2026, reported that up to half of order volume can come from about 7% of guests. That is US restaurant data and your own split will differ. The shape holds in most trades though: a small group accounts for a large share of the till, and they are the group a card lets you identify.
What happens after the reward
The common failure is treating redemption as the end. The customer gets the tenth coffee, the card resets to zero, and the shop goes quiet again.
Redemption is the moment to act on, and not because motivation is at its peak. Research on goal gradients finds the opposite: customers tend to slow down immediately after earning a reward, before picking up again toward the next one. That dip is the reason to start the next card while they are still standing there, and it is the single easiest thing to fix in most programs.
The longer-term version of the same idea is that the card was never the relationship. Once someone has joined, you can tell them the kitchen is open late on Thursdays, or that there is a table free tonight. Loyalty gives you the relationship. Communication keeps it alive.
Frequently asked questions
Is a stamp card just a discount by another name? Only if you give it to people who were coming anyway. The difference is whether the visit would have happened without it. A discount reduces the price of existing demand; a working card buys additional visits at a known cost.
What margin should I use in the calculation? Gross profit on a typical visit, not revenue and not the menu price. If a £4.20 flat white costs you 90p in coffee, milk and cup, your margin is around £3.30. The ~£4 used in the tables above assumes a slightly larger average order, which is why it is worth running the sum on your own figures rather than borrowing ours.
Should the reward be the cheapest thing I sell? Generally not. A reward nobody wants does not get carried, and a customer who does arithmetic on a 40p reward tends to stop. The practical test is whether you would say it out loud at the counter.
Does this work for a business where people visit rarely? The stamp mechanic struggles once visits are months apart, because the card outlives the customer’s memory of it. For salons and barbers the useful lever is usually rebooking rather than a long card.
What if my regulars complain that new customers get the card? Give everyone the card. The point is about where you expect the return to come from, not about withholding it. Counting your regulars as program growth is what produces the disappointing first row of the table.
Where to go next
The next decision after “should I run a card” is what kind and how long. Stamps, points or vouchers covers which mechanic suits your trade, and how many stamps a card should have covers the length. If the goal is filling the slow shifts rather than adding to the busy ones, the quiet Tuesday problem is the one to read. Getting the card onto the phone in the first place is what to put next to the till.
Running the card from a wallet pass, so it can carry a message as well as stamps, is what Passumo does. There is a worked example for cafés.
Sources: Malika Chaudhuri, Clay M. Voorhees and Jonathan M. Beck, “The effects of loyalty program introduction and design on short- and long-term sales and gross profits”, Journal of the Academy of Marketing Science, 47(4), July 2019; 322 publicly traded firms introducing a loyalty program between 2000 and 2015. Toast and Resy, The Regulars Report 2026, published 4 June 2026; aggregated Toast POS transactions, Q1 2026, US restaurants only. The margin figures in this article are illustrative examples, not measured averages.