Good product design starts with watching rather than guessing. That matters even more when you're building for a market unlike the ones most tech comes from.
Design for the conditions here
Translation is the easy part. The harder work is designing from the start for the infrastructure that is really there: patchy internet, power cuts, older phones. Your app has to run on a $50 Android over weak 3G.
Then there is how people pay. Cash is still the default here, QR payments are growing fast, and credit cards are rare outside tourist spots. Language matters too, which means Khmer script, mixed Khmer and English screens, and voice for people who don't type much.
The last one is the easiest to miss. Decisions get made by families and groups, on trust between people, not by the lone Western user most software assumes.
KOOMPI and Baray
The KOOMPI laptops had to hold up in heat and humidity on unreliable power, so we spent time on things like a power supply that survives voltage swings. Cheap was never the whole specification.
For Baray we started by watching how small shop owners handle money. They juggle several apps, keep paper records, and settle up at the end of the day. We built for that habit rather than trying to correct it.
Watch people work
We spend time in the field, watching people work instead of only interviewing them. Ideas get tested on paper before anyone writes code. We build for the hardest case first, on the assumption that if it works on a slow phone with bad internet it will work anywhere. Then we test with real users rather than friends and family.
Common mistakes
The usual ones are assuming everyone has a smartphone and fast internet, designing for a single user when families decide together, ignoring offline entirely, and building for behaviours people here don't have.
Local knowledge takes time
Understanding a place deeply takes time. A foreign company can copy your features in a quarter. Knowing how people here actually live and work takes years, which is the advantage worth protecting.