Episode 4 On Personalized Search (Part 1)

Bo and I have been working on Umi Search, a personalized search related product, for about half a year. It didn’t turn out well in the end, but we realized it would be helpful if we can talk about the idea from the very beginning, and visualize how it could be in the future.


05:00 How Bo started the idea from four to five years ago. He sold everything to start business in China, and met Umi, his wife there. Without phone or Internet connection, she became the source of information to him.

13:00 The problem can also be rephrased as how to do mobile search more efficiently. Umi Search try to solve it by searching within your social accounts, like Twitter, Facebook, Instagram or Wechat.

21:00 If the search results are most often UGC, some results might be redundant or of low quality, how could Umi Search solve that?

33:00 Technical problem might be one aspect that we were not proceeding with Umi Search, but it seems from the conversation, there are more issues to it, like the definition of personalized search.

49:00 We talks about Slash the keyboard as an interesting approach to the problem. It seems interesting if some shortcuts can be applied in the mobile search process, just programmers use code lines to create actions.

53:00 If the search product actually comes out, the user group might be pretty limited.