3 min read
B2C and B2B, plus content-driven commerce with no data yet to validate it.
Stakes
Daraz dominates online shopping in Nepal. ChinaBazaar couldn't beat it on catalog, price, or trust. Its only bet was content-driven commerce, meaning live video and shorts, a way of buying the market had never standardized. If that bet failed, the platform had no reason to exist. This wasn't a feature. It was the whole point.
Constraint
Two months to design. Seven for the whole system to go live.
No PM. No BA. No other designer. Every judgment was mine, and it carried real weight: if anything went wrong, I had to take full ownership.
The decision
Where the effort went. Right after reading the BRD, before touching design, I made the call. B2C was a proven model, so it got the least attention. B2B got a measured share. Content commerce, the bet, got the most creative energy. With two months, this couldn't be decided later.
How to design with no data. Content commerce had no local playbook. Daraz had tried it and stopped. If the market leader couldn't make it hold, ChinaBazaar had even less room to be wrong, with no users, no catalog, and no brand identity yet.
So I stopped waiting for local data. I studied how Amazon and Chinese platforms run the mechanics and designed from that logic, not their screens.
Then I made the call: the first version would be the data for ChinaBazaar, and whatever it showed would become the benchmark.

The mistake
I misread the BRD and took a B2B flow in the wrong direction. The client caught it, not me. Three days gone, out of a two-month window.
I didn't pause. The moment that meeting ended, I reset and went after the gap, recovering as much of the lost time as I could.
Trade-off
I put the most weight on live commerce, knowing Nepali shoppers are money-anxious and slow to trust it, and with no way to test that.
B2C and the dashboard paid for it. B2C ran on proven patterns with nothing of its own. The dashboard was only designed for its core flows, and developers built the rest from those while I guided the UX.
Some of it shipped with UX holes I could see and never got back to fix. I accepted that because I was busy designing the content commerce, the bet ChinaBazaar stood on.
Outcome
Shipping that first version turned content-driven commerce from a guess into a system with real signal behind it: live sessions, the shorts feed, and purchase behavior all instrumented from day one, feeding a working set of metrics across every surface instead of a single vanity number.
The bet wasn't that the first version would be right. It was that it would finally be measurable, and once it was live, it was.
Reflection
The three days weren't a design failure. They were an assumption failure, mine.
I read the BRD assuming I already knew the flow. Now I confirm it line by line before anything gets built. A familiar-looking flow is no reason to assume.
A familiar-looking flow is no reason to assume.