A Long-tail case study

This morning when I opened google analytics, I was a little bit surprised that CAIMAI‘s daily unique visits increased by 30%, and 90% of visits came from Baidu search results. The interesting part is the 90% of leading keywords are very infrequent, such as:

  • The best Peking Roasted Duck
  • The best movie about mummified corpse
  • The most popular blood type in Taiwan
  • The best way of oral sex
  • Why people like panda
  • ….

Since CAIMAI is a Chinese web2.0 site for sharing the best stuff and best memories (similar to Bestuff.com) , everything could be found on it. That’s why so many wacky keywords work.

I immediately told myself:”WOW! This is a good case for long tail theory.”

But I was suddenly frustrated when I saw a 70% high bounce rate. Damn, Netflix is not like that. So What’s the difference?

There are official sites for Roasted Duck, and movie, and blood type analysis, and sex position, and panda. They are more professional & detailed than CAIMAI in this specific area. In contrast, there are scarcely any substitute for those hard-to-find DVDs which could be only found on Netflix ( Those add up to 21% of all sales).

Yeah. I guess substitutability is fatal weakness for a failed Long Tail model.