Bring Coffee Regulars Back More Often with a Stamp Card

Raj hands Amira her usual latte at his coffee shop in Subang. She stops here on her way to work roughly every fortnight, likes the coffee, and sees no particular reason to come sooner. Raj recognises the same pattern among his 400 regulars: about two visits a month, when three would make a difference.

Raj, Amira and this shop are fictional; all figures are illustrative. A RM2 discount would make Amira's next planned latte cheaper without changing her routine. Raj wants a reason for another visit, not another discounted bill. Could a card that already looks started bring her back sooner?

1. Join with progress. Amira joins the pilot and staff immediately issue one courtesy stamp before any purchase. Her card shows one of six.
2. See the next step count. Her next RM12 latte earns another stamp. Two of six feels a third done.
3. Choose the familiar counter. At four stamps, Amira starts choosing Raj's shop over the nearer alternative. In this fictional journey, completing the card makes that small detour worthwhile.
4. Finish, then keep returning. Amira completes six stamps, claims her free latte and maintains the higher frequency for the following weeks. Raj measures whether this actually happens; the pilot offers no second card.

flowchart LR
    A[Join with a courtesy stamp] --> B[Buy and see progress]
    B --> C[Choose the shop again]
    C --> D[Enjoy the reward and keep returning]

Raj randomly divides 80 comparable regulars into 40 offered the card and 40 continuing normally. He keeps an eight-week baseline and follows both groups through the eight-week pilot plus four weeks afterwards, so an earlier purchase is not automatically counted as an extra purchase.

Amira receives her joining stamp from staff. They then check a paid bill of at least RM12 and issue one stamp per visit; all items qualify in this pilot, and split bills earn no extra stamps.

Raj sets one completion per customer and leaves automatic spend stamping off for this capped test. Staff control the participant list and 14-day gift deadline, checking that the configured reward remains usable throughout that window; total reserved campaign cost is RM280.

• Paid-visit gap: compare intervals in the first and second halves of each card, excluding the courtesy stamp.
• Incremental paid visits: compare baseline-adjusted changes between groups through the follow-up window; target at least 53.
• Post-reward frequency: compare paid visits after collection with prior frequency and the comparison group.

Common mistake: calling every completed card new revenue. Amira could finish through purchases she already intended to make.

Safeguard: reconcile one staff register of participants, qualifying bills, refunds, completion dates and gifts daily. Keep exposure within 40 participants and preserve the RM160 reserve for promised gifts.

Keep if at least 53 additional paid visits materialise with positive contribution and the safeguard holds. Improve a promising result with one bounded timing change in a fresh test; stop expansion if contribution is non-positive, the safeguard fails or the completed test misses its threshold.

Does the customer buy six coffees?

Not on this first card: one courtesy stamp leaves five earned stamps. Describe that clearly rather than advertising six required purchases.

Is eight weeks enough for occasional customers?

Five qualifying purchases may be a stretch at Amira's old pace. Check the real purchase cycle and test a longer window separately if customers cannot plausibly finish.

What happens after a refund or void?

This manual pilot needs staff reconciliation of the affected stamp and any completion reward. Do not promise that reversing a stamp automatically withdraws a coffee already served.

Can we reduce the target halfway through?

Active cards recalculate when the target changes. Keep this test unchanged and evaluate a different target with a fresh group.