Human powered search has been hitting the headlines again recently after former Weblogs Inc. owner Jason Calacanis launched the latest and most controversial human edited search engine, Mahalo.
Mahalo is run by a team of editors whose task is to hand pick a selection of websites, videos and other resources to form the search results pages for thousands of the most popular search queries on the web.
A number of popular search terms are listed although the site is clearly targeted at a US based audience. The results for queries such as iPhone and cell phones are fairly useful and most of the other
queries I tried weren’t much different to the choices I might have made if I was editing the pages. This is clearly the key aspect of a human powered search engine – it has to return the results that the user was expecting to see. The sites must be both relevant and offer users the information they were searching for. Google has spent years perfecting the positive user experience and human editors have a lot to live up to.
The reason human edited engines will fail is that they can only cover a tiny proportion of queries. Google sees over 3.6 billion US search queries per month and 20% of these are brand new, never seen before, search terms. When you consider 50% of the search market is outside the US you end up with the number of unique search queries growing by 1.5 billion every single month. Human powered search can’t even hope to compete with a normal search engine in terms of variety and depth of queries. Even Wikipedia only has 7.5 million articles.
Mahalo would argue that they backfill results with Google search results pages when they haven’t yet written a page about your query. That’s all very well but you can’t expect to gain market share by copying a market leader.
Another major problem for Mahalo is that they won’t be able to follow sites like Wikipedia and get billions of visitors per month from Google. Matt Cutts has recently indicated that Google doesn’t regard other websites search results pages as useful to its users and would like to see them removed from their index.
The points made above don’t even touch on the fact that if Mahalo was to take off there would be a huge incentive for the editors, who are currently being paid relatively low wages, to accept incentives and bribes to rank some sites higher than others.