Redesigning a school payments app

A school payments app for parents in the UAE, redesigned in English and Arabic at once. Registration, home, adding a student and payments were all rebuilt and tested with parents in both languages.

Overview

Parents in the UAE use Kidly to pay for school in one app: fees, excursions, uniforms. As competitors grew, it started to feel dated. I redesigned it end to end, from research to dev QA, as the only designer on a team with 1 PM, 2 iOS and 2 Android developers, and 2 QA. We started with about 2,000 active users and a goal of 10% growth in Q1 2024.

14→3%Registration dropout

+18%Payment completion

35%

Active users on ArabicReached within the first month.

The problem

Parents ran into friction at every step. 14% dropped out during registration, home was a static page with no quick actions, adding a student produced duplicates and wrong schools, and payments repeated the same confirmations.

A significant share of them also prefer Arabic. We tested with 30 UAE-based parents who actively pay for school, with children aged 6–18.

The approach

Design Arabic as its own interface

The fastest route was to translate after the English launch, but that ships a second-class experience. Mirroring every screen automatically was cheap, but it doesn’t decide what should flip and what should stay.

I designed both versions in parallel instead and validated each flow separately with Arabic speakers, because navigation, swipe direction, icons and reading order all change. Not everything should, though: the back arrow points the other way, while the search icon, Latin names and numbers keep their shape and direction. The cost was more work on every flow, and both mobile teams had to build for it from the start.

The same screens side by side, one decision at a time.

Building it

Registration and home

Registration was where parents dropped out. I replaced the password flow with Google, Facebook and Apple sign-in, and compared forms in a 2-week A/B test. Google became the dominant choice, and dropout fell from 14% to 3%. Sign-in now lands on a home rebuilt around what parents actually do: students, offers, school features.

Names are placeholders.

Adding a student

Duplicates and mismatched schools were a top complaint. Validation catches typos but not the wrong school, so we connected the UAE school and university API: parents pick the school, then find their child by name. There were no lookup failures in testing.

Names and schools are placeholders.

Payments

Payments had redundant confirmations and too many form fields. I removed the first and consolidated the second: fewer steps, same trust signals. Payment completion grew 18%.

Cards are placeholders.

What we cut

A customisable home with draggable widgets tested well, but the backend wasn’t ready and half-built wasn’t an option, so it stayed a concept.

Impact

35%Active users on Arabic within the first month. It shows adoption, not proof of what caused it.

+13%

Users in Q1 2024From about 2,000 to about 2,260 active users, above the 10% target. New features can now be added to home without a structural redesign.

Takeaways

What I’d changeCheck feasibility before testing a concept. The customisable home tested well, then stopped at the backend.

Localization wasn’t a translation step, it was a design problem.

Treating it that way from the start gave Arabic-speaking parents a complete product, not a late add-on.