The problem
An Arabic-first learning app with more than four million users, featured as App of the Day and App of the Week, cannot afford to crash on a five-year-old's iPad. Crashes were the product problem, not a technical detail.
What we built
Features for a four-million-user base
New features and redesigned screens shipped into an app already installed on millions of devices, in Arabic and English.
Crash rate cut 90%
The trending crashes were fixed first, in order of how many children they actually affected, rather than in the order they were reported.
A/B testing introduced
Product questions started being settled with data instead of opinion, which changed how the team chose what to build next.
A release pipeline
Builds and App Store releases automated on Fastlane, GitHub Actions and CircleCI.
Screens
How we worked
Our founder joined as an iOS developer and worked inside an existing codebase with a large installed base, where every change reaches millions of devices and rollbacks are expensive.
Results
users on the App Store
Featured as App of the Day and App of the Week.
crash rate
Trending issues fixed first.
A/B testing introduced, so feature decisions were settled with data
What's next
The Arabic and right-to-left experience from this work carries directly into the bilingual products we build now.
