The reward threshold is the most consequential setting in a loyalty programme and the one most often chosen arbitrarily. Too high and customers never believe they will reach it, so it changes no behaviour. Too low and you give away margin to customers who were coming anyway.
The rule: two to three months of normal visits
Set the threshold so a typical customer reaches it in roughly two to three months at their existing visit rate. That window is long enough for the programme to pay for itself and short enough that the customer can see the finish line from the start.
| Business | Typical visit rate | Suggested threshold | Time to reward |
|---|---|---|---|
| Café | 2 to 3 per week | 10 stamps | ~4 weeks |
| Bakery | 2 to 4 per week | 8 stamps | ~3 weeks |
| Juice bar | 1 to 2 per week | 8 stamps | ~6 weeks |
| Restaurant | 1 to 2 per month | Points, ~200 | ~3 months |
| Hair salon | Every 6 to 8 weeks | 6 visits | ~9 months |
| Barbershop | Every 3 to 5 weeks | 6 visits | ~5 months |
| Yoga studio | 1 to 3 per week | 10 classes | ~6 weeks |
Suggested thresholds from the table above.
Why the reward matters more than the number
A common mistake is setting a sensible threshold and then making the reward the cheapest item on the menu. A free filter coffee for a customer who always orders a flat white reads as a downgrade, and it teaches customers the programme is not generous.
Make the reward the customer's usual order. The margin difference is small and the perceived value difference is large.
Start customers above zero
Giving new members a head start (two stamps at sign-up on a twelve-stamp card rather than zero on ten) measurably improves completion even though the remaining effort is identical. Progress already made feels like something worth protecting, while an empty card feels like a long way to go.
When to change the threshold
Review it after three months against one number: the share of enrolled customers who have reached at least half the threshold. If that figure is below about a quarter, the threshold is too high and customers have disengaged. If nearly everyone is completing cards quickly, you are discounting rather than retaining and should raise it.
Changing the goal does not require reissuing cards on a wallet-based system: existing cards update in place.
Should I use stamps or points?
Choose the mechanic by how much your ticket sizes vary. Stamps suit businesses where every purchase is roughly the same value, points suit businesses where they are not, and visits suit appointment businesses where what matters is that the customer came back at all.
| Mechanic | Use when | Typical threshold | Weakness |
|---|---|---|---|
| Stamps | Every purchase is similar in value | 8-10 stamps | A €2 pastry earns the same as a €9 lunch |
| Points | Ticket sizes vary widely | ~200 points | Harder to explain at a counter |
| Visits | Appointments, infrequent but high value | 5-6 visits | Ignores how much each visit was worth |
The failure mode with stamps is a customer buying a coffee to earn the stamp they would otherwise have earned on a full lunch: you are rewarding the cheap transaction. If a third of your customers are doing that, move to points.
The failure mode with points is that nobody can remember what a point is worth. If you use points, anchor them: "100 points is a free main course" is legible, "you have 4,350 points" is not.
What does the reward really cost me?
Calculate the reward on its marginal cost, not its menu price, and the number is almost always smaller than it feels. A €4 flat white given free on a ten-stamp card costs you the milk and the coffee, typically under a euro, not four euros.
Spread across the card, that cost is under ten cents per visit. Against an average ticket of €4 to €6, the programme consumes roughly two per cent of the revenue from customers who, by definition, came back ten times.
This is why setting the reward to the customer's usual order is usually correct and rarely as expensive as owners fear. The margin difference between a filter coffee and a flat white is small; the difference in how the reward is perceived is not.
Should the reward expire?
For an independent business, no. Expiry protects against a liability that a café or salon does not meaningfully carry, and it costs you the goodwill of exactly the customers who engaged most with the programme.
The exception is a genuine seasonal offer where the reward only makes sense within a window. State the window clearly on the card in that case, and send one reminder before it closes. An expiry a customer discovers after the fact is worse than no programme at all.
Inactivity is a different matter. Removing a card that has had no activity in eighteen months is reasonable housekeeping. Doing it at three months is a way of losing a customer who was going to come back.
Should there be more than one reward?
One reward is the right answer for almost every independent business at launch, because the programme has to be explainable in a single sentence at a busy counter. A second tier can be added later, once you know the first is working.
If you do add one, make the second reward meaningfully larger and much further away: a free coffee at ten stamps and something substantial at thirty, for instance. The purpose of a distant second tier is to give your best customers a reason to keep collecting after the first reward resets the card.
What does not work is several small rewards close together. It reads as complexity rather than generosity, and staff end up explaining the scheme instead of serving.
Why does starting above zero work?
Two effects from behavioural research explain most of what makes a threshold work, and both are worth knowing because they point in practical directions.
The first is endowed progress: people pursue a goal more persistently when some progress has already been made for them. A card with twelve slots and two pre-filled outperforms an identical card with ten empty slots, even though both need ten more purchases. The customer in the first case has something to protect.
The second is the goal-gradient effect: effort accelerates as the reward approaches. Customers visit more often in the final stretch of a card than at the start, which is an argument for thresholds that are visibly reachable. A customer who cannot see the finish line never enters the acceleration phase.
Together they suggest a simple design: a slightly longer card, pre-filled at sign-up, with progress shown prominently. That is what a wallet card does by default, since the stamp count sits on the customer's lock screen rather than in a drawer.