Almost every stamp card in the world says ten. Not because ten is right, but because ten is a round number and the card had ten boxes on it. It is a printing decision that somehow became an industry standard.
The threshold is the only setting on a loyalty programme that touches your margin, and it is worth ten minutes of arithmetic. Here is the arithmetic.
The two numbers you need
Everything else is noise. You need:
- What the reward costs you to make — goods, not menu price. For a flat white, the milk, the beans, the cup and the lid.
- How often a regular comes in — visits per week for the customers who would realistically fill a card.
The first number sets what the programme costs. The second sets whether anyone ever finishes a card, which decides whether the programme does anything at all.
What the reward really costs
There are two honest ways to price a free coffee, and the truth is between them.
The floor is what the cup costs you to produce. If the reward drink is one the customer would not otherwise have bought — they took the free one on top of a normal week — you are out the goods and nothing more.
The ceiling is the full menu price. If they would have bought that coffee anyway, giving it away costs you the whole ticket, margin included.
Run both. Call a flat white £3.50 on the menu and £0.90 in goods, and the discount rate over the spend it takes to earn one looks like this:
| Stamps required | Customer spends | Cost floor (goods) | Cost ceiling (full price) |
|---|---|---|---|
| 6 | £21.00 | 4.3% | 16.7% |
| 8 | £28.00 | 3.2% | 12.5% |
| 10 | £35.00 | 2.6% | 10.0% |
| 12 | £42.00 | 2.1% | 8.3% |
Two things fall out of that table. The gap between six stamps and twelve is about two points of real cost — small. And even the most generous row sits well under what a delivery aggregator takes off the top of a single order, which is 14–30%. Loyalty is a cheap acquisition channel by comparison, and it buys repeat visits rather than one-offs.
The number that actually matters: time to reward
Cost is the boring half. The half that decides whether the card works is how long a customer has to wait before anything good happens.
| Visits per week | 6 stamps | 10 stamps | 12 stamps |
|---|---|---|---|
| 5 (daily commuter) | ~1.5 weeks | 2 weeks | ~2.5 weeks |
| 3 | 2 weeks | ~3.5 weeks | 4 weeks |
| 2 | 3 weeks | 5 weeks | 6 weeks |
| 1 | 6 weeks | 10 weeks | 12 weeks |
A reward more than about a month away stops being a reason to come back and becomes a thing people forget they are collecting. If your typical customer comes twice a week, ten stamps is already a five-week commitment — defensible. If they come once a week, ten stamps is two and a half months and the card is decoration.
Set the threshold from your own visit pattern, not from the card design. A café next to a station and a café in a residential street have the same menu and completely different answers. Aim for a first reward inside four weeks for the customer you want more of.
Pick the reward before you pick the number
The threshold and the prize are one decision. A few rules that hold up in a small venue:
- Reward the thing they already buy. A free coffee for a coffee drinker is instantly legible. A voucher toward a food item they never order is not a reward, it is homework.
- Use a high-margin item. Drinks, not the pastry you buy in at 60% of retail.
- Avoid conditions. "Free coffee, before 11am, excluding large" is three chances to disappoint someone at the counter and one argument for your staff to have.
- One reward, not a tier system. Tiers are a chain's game. A single venue with one card and one prize is understood by every customer on the first explanation.
Bonus stamps: the cheapest lever you have
The threshold is not your only dial. A one-off extra stamp costs you a single cup of goods and moves someone meaningfully closer to a reward, which is exactly when a loyalty programme gets its grip. Two that work:
- A birthday stamp, once a year. It costs nothing to give and lands as a gesture rather than a promotion.
- A stamp for completing their profile — a name and a contact detail. You are buying a customer record for the cost of a flat white, which is cheaper than any advertising you will ever run.
Do not attach a bonus to leaving a review. Google's policy prohibits incentivised reviews, and the risk to your listing is not worth a stamp.
Changing your mind later
Be careful here, because this is where owners quietly annoy their best customers. Progress on a stamp card is measured against the current threshold. Raise it from eight to twelve and somebody who was one stamp from a free coffee is suddenly five away — and they will notice, because they have been counting.
If you have to change it, lower it rather than raise it, do it once, and honour anything a customer was close to under the old rule. A threshold is a promise with a number in it.
A sensible default, if you want one
- Work out your goods cost for the drink you would give away.
- Estimate visits per week for your typical regular — not your best customer, your typical one.
- Pick the threshold that puts the first reward three to four weeks out.
- Sanity-check the cost floor against the ceiling. If the ceiling is under about 10%, you are fine.
- Leave it alone for six months, then look at how many cards actually completed.
For most cafés that lands somewhere between eight and ten. The difference from the standard ten is not the number itself — it is that you now know why it is the number, and what it costs.
In Coffeeloc you set the reward and the stamps required yourself on a digital stamp card, the counter screen prompts staff automatically when a card is full, and the first 100 customers are free.