How to choose the stamps and the reward

Start with the sentence the person at the till will say. Decide which purchase earns a stamp and what a completed card buys. Write both parts before choosing the number of squares. If the reward is still described as “something free”, the cost calculation is unfinished. If the qualifying purchase is still “a visit”, decide whether that means a purchase, an item or a transaction.

Treat this as a small operating decision with a written budget. You should be able to explain the offer to a guest, calculate the cost of honouring it and tell staff what to do at redemption. This article works through that decision using invented inputs for arithmetic only. It describes no real shop, customer, result or industry average.

Define what earns a stamp

For the example, let an eligible paid drink earn a stamp. The reward will be an eligible drink after eight paid drinks. A free reward drink earns no stamp, and extras are outside the reward. This means eight paid drinks followed by a free drink. Keep that wording precise: “the eighth drink is free” would describe a different offer, with only seven paid drinks beforehand.

Before using the rule, choose the eligible drinks and write down the treatment of sizes, extras, refunds and purchases of several drinks together. If you intend a stamp per transaction instead, redo the model around the qualifying transaction. Do not calculate with the price of a large order and then let a much smaller purchase earn the same progress without checking the result.

Build a completed card from the bottom up

Use these deliberately simple inputs: revenue of €3 per paid drink, a variable cost of €1 per drink, and a reward that also costs €1 to fulfil. Here, “variable cost” means the costs you include because you serve that additional drink. Put ingredients, packaging and any other relevant incremental costs into your own estimate. Do not copy the example’s cost because the selling price looks familiar.

For this model, contribution means revenue minus the variable costs included in the calculation. A paid drink therefore contributes €3 − €1 = €2 before fixed costs and programme overhead. We are not calculating net profit. Rent, any labour cost outside the variable estimate, equipment and other overhead still need to be covered. List exclusions alongside the inputs so you can see what remains outside the model.

1. Revenue from the completed card: eight paid drinks × €3 = €24.
2. Variable cost of the paid drinks: eight × €1 = €8.
3. Variable cost of the reward: one free drink × €1 = €1.
4. Contribution after fulfilling the reward: €24 − €8 − €1 = €15.

Without the reward, those same eight paid drinks would contribute €16. Adding the free drink reduces that contribution by €1, or €1 ÷ €16 = 6.25%. Its fulfilment cost is €1 ÷ €24, approximately 4.17% of paid revenue. These percentages answer different questions. Label the denominator whenever you show a percentage to yourself, a colleague or a supplier.

There is another percentage you could calculate. Nine identical drinks at the assumed €3 price would total €27; the card collects €24. Relative to buying all nine at that price, the saving is €3 ÷ €27, approximately 11.11%. That describes the assumed menu-value saving. It does not replace the calculation of the shop’s costs or tell you whether the programme increased profit.

Calculate the cost of keeping the promise. Then ask which paid purchases the promise actually adds.

Change the threshold without changing the story

Keep the same price and costs and compare alternative thresholds. After six paid drinks, revenue is €18 and contribution after the reward is €18 − €6 − €1 = €11. After ten paid drinks, revenue is €30 and contribution is €30 − €10 − €1 = €19. These are different-sized purchase cycles; the larger contribution is not evidence that the longer card will perform better.

The comparable fulfilment-cost shares are €1 ÷ €18 = approximately 5.56% for six stamps, 4.17% for eight, and €1 ÷ €30 = approximately 3.33% for ten. Under these fixed assumptions, spreading the same reward over more paid drinks lowers its cost per paid drink. Whether the longer journey is acceptable is a separate question to investigate in your shop.

For your own worksheet, call the threshold N, revenue per qualifying purchase P, variable cost per qualifying purchase C, and reward fulfilment cost R. Then contribution for a completed card is N × (P − C) − R. The reward fulfilment-cost share of paid revenue is R ÷ (N × P). Add programme expenses separately and use a weighted purchase mix if your qualifying items differ.

Test a reward that costs more

Now change only the assumed reward cost to €2. Keep eight paid drinks at €3 revenue and €1 variable cost each. Contribution becomes €24 − €8 − €2 = €14. The reward cost is €2 ÷ €24, approximately 8.33% of paid revenue. You have doubled that cost share while keeping the same stamp threshold. This is why the permitted reward matters as much as the number of stamps.

Price the full reward you promise, including allowed extras. If you offer a choice, run the calculation for the most expensive eligible option as well as your expected mix. If the expensive case is unacceptable, narrow the choice or redesign the offer before launch. Describe any limits where people join and where they check their progress, so staff can apply the same rule.

Keep a dated copy of the inputs beside the offer. When you change a selling price, ingredient, portion or reward choice, reopen the calculation and replace the relevant assumption. Decide how any revised terms will apply to cards already in progress, and have that approach checked before announcing it. Include the cost of honouring the existing promise in your decision about the next version.

Check what happens if the free drink replaces a sale

The first calculation assumed eight paid drinks plus an additional free drink. Consider a different, explicitly hypothetical comparison: without the programme, the same purchase horizon would contain nine paid drinks. At the example’s €2 contribution each, that would contribute €18. Eight paid drinks and a free one contribute €15. The difference is €3, rather than the €1 cost of serving an additional free drink.

That €3 difference consists of the lost €2 contribution from the displaced paid drink and the €1 cost of fulfilling the free drink. Do not add another €3 of lost revenue on top; that would count part of the loss twice. Compare complete revenue and cost totals under each scenario whenever the shorthand becomes confusing.

Under the narrower assumption that an extra paid drink contributes €2 and creates no additional reward cost within the comparison, €3 ÷ €2 = 1.5 extra paid drinks would cover the difference in aggregate. This is arithmetic, not a prediction. Extend the comparison to include newly earned rewards, programme fees and any extra operating costs before using it as a break-even target.

Translate stamps into a plausible journey

Use visit schedules as assumptions to test, not as facts about your audience. Starting from an empty card and assuming a paid qualifying purchase twice per week, eight purchases take four weeks. At one qualifying purchase every fortnight, they take sixteen weeks. Those are illustrative schedules, not forecasts. Replace them with observations from your own programme and make the starting point consistent.

There is relevant research, but keep its claim narrow. Kivetz, Urminsky and Zheng reported that participants in the café programme they studied bought more frequently as they approached a reward. Their 2006 Journal of Marketing Research paper supplies evidence of purchase acceleration in that setting, not an optimal stamp count for your shop. The full paper is available from Columbia.

My practical recommendation is to show the remaining requirement plainly and test whether your chosen journey makes sense. Do not insert a claimed uplift from that paper into your budget. Keep the cost model viable under conservative assumptions about extra purchases. Then use the trial to learn whether the offer changes behaviour enough to justify its cost.

Write the redemption rule before the launch

• Name the eligible paid purchase and whether the stamp belongs to an item or a transaction.
• Name the reward, its allowed size and extras, and whether it can be exchanged for anything else.
• State when the reward becomes available and whether the reward itself earns progress.
• Decide how refunds, accidental stamps and a missing reward item will be handled.
• If you choose an expiry policy, explain it before enrolment and have the terms checked for your market.

Run through the rule with staff using test cards. Include a completed card, an ineligible item and a request for an extra. Ask them to explain the result in their own words. If their answers differ, revise the wording and repeat the demonstration. Treat a rule that needs a manager to interpret every redemption as unfinished.

Review the offer using paid activity and actual costs

For the trial, record eligible paid purchases, rewards earned, rewards redeemed, actual reward cost and programme expenses. Keep the observation period and starting conditions written down. Track outstanding rewards separately, using their fulfilment cost for planning rather than assuming they will disappear. Agree the formal accounting treatment with your accountant; this worksheet is an operating model.

Choose a review point before launch. Compare the actual cost mix with your assumptions and examine paid purchasing before and after redemption. Where practical, use a comparable group or period to test your explanation for any change. Do not credit every purchase by a member to the programme. Keep the rule if the observed economics and the counter routine support it; otherwise change the specific assumption that failed.