What I learned building Africa's largest home services marketplace

Context why · POV · RFI
Why I’m interested

Kandua was the crucible. A marketplace and business tools platform for home service businesses in South Africa, built from scratch and sold to the country’s largest insurance company. This thread is a retrospective — what I learned about micro-entrepreneurs, technology adoption, and what it actually takes to build infrastructure for informal workers.

Final POV
  • 5-10% of people doing “entreprenuership” want to do it. The rest would prefer a job but can’t or don’t know how to find it. A more centralised approach could cater for this dynamic.
  • Small business SaaS is hard. But really cool
  • “Design” can make itself small for no reason. If you can give people money - they will forgive a lot of UX/UI things. Design for giving people money then get better from there
  • Aka, offtake is everything. consumer can have a ceiling especially in a naturally low frequency use case
  • Online to Offline is hard. Collect your money at the point of value you control and make that the business. Or try to control the entire interaction
  • The more you control, the less breadth you should have. A “marketplace” only makes sense when you can outsource properly.
  • Venture culture is poison. It’s not relevant for most things. It’s very important to know why you need that kind of money
  • Culture is everything. Very hard to get right but when you do - it sings. When you don’t - it wails.
  • Mergers are tough. Culture is very much contextual and very difficult to reset
  • It’s not enough to read the fine print. Every promise should be written down
  • Some businesses are inherently ugly. You can’t change that. But you can protect against it.
  • Be mission driven

I’ll keep adding things as I remember them

Archived — no longer updated
20 Feb 2026 What building Kandua taught me — the full retrospective
The conclusions I'm carrying forward from years of building infrastructure for informal workers — the super agent hypothesis, narrow technology adoption, and credit as the biggest unmet need.