Just in time to make me go “d’oh!”

Yesterday, I griped about Mahalo not fitting my definition of social search. Today, the latest release of Mahalo Follow is announced on the Mahalo Blog. In my previous blog post, I asked:

Is there a product out there that combines all this data and uses it to enrich search results in realtime?

Wouldn’t you know it, but that’s exactly what the latest version of Mahalo Follow tries to deliver, in a way. It will now rewrite a Google search results page (SeRP), injecting links from Mahalo data. This is exactly the kind of search engine enhancement that the algorithmic search engines really need. Well done, guys!

Of course, there’s still a huge gap of opportunity for improvement, here. Mahalo still isn’t leveraging the Web 2.0 network effect that is possible through social software. You can recommend a particular page to your friends and possible inclusion in the Mahalo data set. But, the real victory is when I can vote/indicate when I believe any particular search result is relevant or not for the search query I just performed. Then, to complete the circle, when I perform searches, highlight and/or bury the results based on what my friends have voted on.

In a way, the Spock folks have already implemented this, but have limited it to just people (for now?) … on a SeRP for a person on Spock, there’s a section called “Other Results from the Web” which is populated with results from Google. Then, as a registered Spock user, I can vote on each result as being relevant to me or not. Other users can similarly vote and the premise is that “wisdom of crowds” will enable the most relevant results to rise to the top. Mahalo could do the same through Mahalo Follow, by enabling voting/recommendations directly off the Google SeRP, and using that collected information to improve future search results.

What about the problem of people trying to “game the system”? That’s where the “social” part of social software combined with the network effect come into play. As long as the software either only uses my friends’ recommendations–and perhaps second-degree friends, but definitely not third-degree friends or further away–I can avoid being affected by people gaming the system by being selective about who I make friends with. Some folks will friend everyone they can possibly find, some will friend no one at all, but those who fall in between will receive the most benefit and that class of users should be the majority.

This is also where reputation and trust come into play in social software: a person who has a reputation of recommending good links and is otherwise trustworthy will attract followers as they will want to benefit from that person’s activity in the system. A person who tries to game the system will eventually be self-selected out by losing friends in the system, until they receive no benefit from trying to game it at all.

Overall, the latest change to Mahalo Follow is a great first step in the right direction, and I’m sure the Mahalo team is busy working on the next set of changes already. I just hope they keep pushing in the direction of making Mahalo a real Web 2.0 social search service. It could be game-changing for the Internet and the way we search.

Tags: Mahalo, social search