Stamp programs: rewards and repeat visits

Start with four written answers: which purchase earns a stamp, how many stamps unlock a reward, exactly what the reward is, and how you will know that a customer returned. Choose the rule using your product costs, purchase cycle and staff's ability to apply it consistently. No single stamp threshold has been shown to be best for every cafe or market.

This guide focuses on designing the offer and interpreting its results. All figures are educational examples in hypothetical currency units, not BlackCards prices, customer results or promises of more visits. For wallet setup and device requirements, follow the wallet guide linked at the end.

An example rule is: “Earn one stamp for each eligible order, then exchange six stamps for a specified drink.” Before announcing it, define the eligible order, drink size and exclusions, and confirm that your actual program settings can implement it. Six is an illustration, not a general recommendation.

Decide whether eligibility is based on the receipt, an item or a visit. Do not leave each employee to interpret it differently. Three drinks on one receipt may still represent one purchase even if your rule awards several stamps. A loyalty stamp alone is not evidence of a separate visit.

After the first stamp in a six-stamp reward example, five eligible purchases remain. If each subsequent purchase is seven days after the previous one, completion takes another 35 days; if the interval is three days, it takes 15 days. This is arithmetic under a regular-cadence assumption, not a prediction that the customer will actually return.

Behavioral research has studied progress toward rewards in specific loyalty programs. It can inform hypotheses to test, but it does not establish the best threshold for your cafe or prove an effect for GCC businesses or BlackCards customers. Keep progress and reward terms truthful and easy to understand; avoid creating a misleading impression of a balance or benefit. [1]

Separate the price a customer sees from the cost to your business. An item with a selling price of 20 currency units might have a variable cost of 8. Use the actual cost of delivering the reward, including relevant ingredients, packaging and other variable costs, rather than treating its whole retail price as your cost.

Consider a simplified educational example: one stamp per eligible paid order, a selling price of 20 per order, variable cost of 8 per order, and a reward costing 8 after six paid orders. Assume one redemption. The example excludes taxes, payment fees, subscriptions, administration time and fixed costs. It is not a net-profit calculation.

Under the same assumptions, rewarding four paid orders gives a reward cost ratio of 8 ÷ 80 = 10%; rewarding eight gives 8 ÷ 160 = 5%. The amounts remaining before other costs are 40 and 88 units respectively. These are completed reward cycles with different numbers of paid orders, so the totals should not be compared without that context.

A higher stamp threshold reduces the example's reward cost ratio, but also lengthens the journey to the reward. The arithmetic does not tell you which option will produce better return behavior. A reward may also replace an order the customer would otherwise have paid for; that opportunity cost is outside the example. Review staff time, errors and any additional purchases you can substantiate as well.

In the scanner workflow, the customer presents the card, an authorized employee reads its code and checks eligibility, then chooses to award stamps or redeem a reward. Reading the code alone does not mean that a purchase has been recorded or a stamp automatically awarded. The cafe page explains the customer and staff journey, and the card demo lets you explore the experience before setting up your business program. [4]

Redemption deducts the configured stamp requirement from the balance under the published program rules; do not assume it clears the entire balance. For illustration, a balance of eight stamps and a reward requiring six stamps leave two stamps. Confirm the reward, balance and branch before submitting, and use the authorized correction process and activity record when resolving an error.

A cohort here means customers sharing a defined starting event, such as their first eligible purchase during launch week. Choose an observation window for each customer and include only those whose window is complete in the finished rate. Cohort analysis separates inclusion, return criteria and observation periods, but it cannot create store-visit data that was never collected. [3]

For an educational example, suppose 160 people registered a card, but only 100 made a first eligible purchase and completed a full 30-day observation window. If 35 of those 100 made a later eligible purchase within their window, the cohort's return rate is 35 ÷ 100 = 35%. Do not divide by all 160 registrations or mix people with incomplete follow-up into a group compared with this result.

Define a repeat visit or purchase in advance: must it occur on a later day, or can a second purchase on the same day qualify? Keep an internal customer identifier, a transaction identifier and the time of the transaction. Several stamps on one order must not become several supposed return visits. If those records are unavailable, report loyalty operations only, rather than labeling them verified visits.

Registrations describe enrollment. A return rate describes behavior in a defined cohort. Redemption records show that rewards were used. Claiming that the program caused additional visits requires an appropriate comparison: frequent customers may already be the most likely people to join. [2]

Marketing attribution allocates credit for an event across touchpoints under a model; that differs from counting returning customers. Assigning a visit to a message does not establish that it would not have happened without that message. [6] If you run a comparison, define the groups, assignment method, observation period and outcome before starting, and keep other promotions as stable as possible. Advertising experiment documentation explains the difference between attribution and additional impact using an exposed group and a control group. We use the principle here, without claiming that an advertising experiment tool is available inside BlackCards. A before-and-after comparison across different seasons is not causal evidence on its own. [5]

You can fix a card-reading problem or inconsistent staff explanations on day one. Evaluating a 30-day return outcome requires the cohort's window to finish. If enrollment runs for an entire month, the last participant may need another 30 days before contributing to the completed result. The end of launch month is not automatically the right reporting date for every metric.