SipPass Coffee
Ordering

In Gaza today cash is scarce and the internet is weak, so people pay by app transfer: cheap, but slow to send on weak Wi-Fi and slow for the shop to check. At one real takeaway café (renamed SipPass for this case study), I timed 10 to 13 minutes in line every morning. The cashier named the bottleneck in my interview: 20+ orders in the morning rush, and minutes to confirm a single transfer by hand.

An interview with the cashier and timed mornings, turned into two composite personas: Laila, who wants her coffee before an 8:00 appointment, and Sami, alone at the counter. Seven ordering products in five groups were checked against what this morning needs, and four constraints were fixed up front: no mobile data, no cards, checks by hand, prices that change often. Four ideas were dropped because each failed a constraint (an app with accounts, QR with pay at the counter, a cloud web app, a self-order kiosk). Seventeen grey wireframes settled the flow before any color.

The shop’s own network does the work: a small local server on the shop’s Wi-Fi. Laila scans a QR and orders with no app and no sign-up, then pays by transfer with her order code (A27) in the note. Sami confirms with one tap, Received or Not received. The bar board shows paid orders in order, and the pickup screen calls A27. Only the payment needs the internet. Honest screens cover the mornings that go wrong: sold out, payment not found, shop offline, and too late for the hold.

Tested with AI personas in a simulated usability test, not with real users, over three rounds. In round 1, 5 of 5 personas missed the latte; after the fix, 4 of 4 finished without help and misclicks went from 14 to 8. Time in the shop goes from 10–13 minutes (measured) to a target of 5 minutes or less, which is not measured yet. Sami’s check took about 35 seconds in testing, against a 30-second target: finding the transfer in the bank app is the slow part, and the next thing to fix. The next step is a pilot in the café.

Built for one café in Gaza, the pattern travels: local-first ordering, a code in the transfer note, and a one-tap check. None of it needs a card terminal or a strong signal. It could fit displacement camps, rural markets, festivals and towns after a disaster, though it has not been tested there.