Migrant workers in Malaysia were excluded from mainstream banking — not because they lacked the need, but because every existing solution was built for someone else. Existing fintech apps assumed smartphone fluency, reliable internet, and English literacy. This audience had none of those things reliably. They used older Android devices, often on 2G connections, in environments with poor light and background noise, sending money home to families who depended on it arriving correctly. The design challenge wasn't feature parity with competitors — it was building something that worked the first time, every time, for a user who couldn't afford a mistake and wouldn't ask for help if they made one.
Research happened in dormitories and work sites, not a lab. Real conversations with migrant workers about how they currently moved money — through agents, through friends, through informal networks — and what made them distrust digital alternatives. The consistent finding was that trust was the barrier, not technology. Users weren't afraid of apps; they were afraid of losing money to something they didn't understand. That shifted the entire design direction. Every screen had to confirm what was about to happen before it happened, and confirm again that it had happened after. Visual language had to carry meaning that text couldn't, because language itself was unreliable across this audience.
Icon-based navigation replaced text labels entirely for primary functions. The assumption was that universal visual metaphors would cross language barriers more reliably than translated copy. Real-device testing with actual users confirmed this — task completion improved significantly when users stopped having to read and started recognising.
The entire frontend was hand-coded in HTML, CSS, and vanilla JavaScript with no frameworks. This was a deliberate performance decision — framework overhead on a 2G connection on a three-year-old Android phone was not acceptable. Complete control over load behaviour and offline capability required building from scratch.
The butterfly brand concept came from the client, representing the slow, careful movement of money and connection to family. Rather than treating it as a logo brief, I built it into the visual language throughout — motion patterns, confirmation states, transition behaviour. It gave the product an emotional coherence that purely functional design wouldn't have achieved with this audience.
Buttons were made significantly larger than standard guidelines after real-device testing revealed that users with calloused hands from manual work were consistently mis-tapping standard touch targets. Outdoor contrast levels also required a colour and type weight rethink. The lab version looked simple. The field version had to be rebuilt around what simple actually meant in context.
active users in Malaysia within three months of launch
MyCash reached over one million active users in Malaysia within three months of launch — an adoption rate that reflected genuine utility for an underserved audience rather than marketing spend. The product won the Seedstars World Award in Malaysia and received regional recognition across Singapore, Hong Kong, and UAE. Users who had never used a digital financial service began conducting significant transactions through the platform. The design system built for the initial launch scaled across the mobile app, web platform, management portal, and partner integrations without structural rework.
The first version was simpler than anything else in the market at the time, but in hindsight it still wasn't simple enough. I was benchmarking against other fintech apps when the real benchmark should have been someone with no prior experience of digital financial services at all. A few of the early flows had two or three steps that existed because convention said they should be there, not because this specific user needed them. I would remove those earlier and more confidently — the instinct to simplify further was right, and I took longer than I should have to trust it fully.