netactr

Jason Calacanis, the beleaguered founder of Mahalo, gave the best keynote at the SES New York. Mahalo is a start-up search engine which tries to reinvent what Yahoo pioneered in the 90’s – a directory for internet maintained by humans. Mahalo pays $15+ to anyone who is willing to build a wiki-like page with links relevant to a specific keyword. An automated project management system allows students or bored professionals pick up work items and deliver the equivalent of Search Engine Result Pages. Apparently 400 people are already part of this distributed work force.
Calacanis talked about many things in his speech, but one particular comment stroke me the most. Answering the existential question whether his business model will survive in the long-term, Mahalo’s founder submitted that frequency of searches aligns with the best search technology for compiling search results. For some fundamental and high-frequency search terms (e.g. Spain or Cars) pages created by experts (e.g. Wikipedia or Mahalo) deliver better user experience; recommendations from friends are good for medium-volume terms; automated machine results are great for infrequently conducted queries or text search.

I get his point that an Encyclopedia Britannica-style answer to a query is the ideal presentation of a result page. However, Yahoo already tried to use humans to classify the internet, and the shier volume of new web content made their directory non-scalable. Besides, Mahalo is soon going to face smarter machine technology (termed Universal or Blended Search) which creates SERPs that combine images, links, definitions, domain-specific information etc. Moreover, many search engines already list Wikipedia result pages as their top results for high-frequency encyclopedia-style searches. I do wish Mahalo good traction with users, but they won’t have it easy, I am afraid.