Wikia Search is well worth watching.

I've been playing around with Wikia Search quite a bit lately and have enjoyed the experience immensely so far. My initial reaction to the new project was rather negative -- I don't know about you, but I honestly expect to see good results immediately whenever I use a search engine. If it's a new engine I'm checking out, then I basically expect to see two things: relevant results for my queries (they don't necessarily have to be the best to begin with) and my own web projects in the index. The latter expectation seems not to be realized more often than not, especially since the SEO guru guys have made me terrified of submitting my own stuff anywhere. "You'll end up in the sandbox, man!" Wikia Search didn't impress me at first because my first searches didn't yield me relevant results. Where was the algorithm? Where was the automated sorting through the chaff that would help me find the needles in the haystacks of the Net? It felt a lot to me like using one of the early search engines when you really never did know what you were going to get, especially for obscure searches.

I've changed my tune after spending more time actually using Wikia Search. It is similar to Wikipedia in that it depends on contributions from people in order to work. A Wikipedia without people doesn't have articles; a Wikia Search without people doesn't have good search results. The search engine is still in alpha, but as it develops and grows I feel confident that the search results will get better. What surprised me the most about my experiences with Wikia Search was how fun it was to use it. Wikipedia lets everyday people play the role of encyclopedist; this project lets them play the role of Internet librarian. I loved going through my bookmarks and adding what I thought the best pages in various categories were to the Wikia index. It was really interesting to think from a query-level perspective and to decide what pages answer a given query best. It was also interesting deciding what description to write or quote for each site. People are going to really have fun with this after they give it a chance. In time, People Search Power could perhaps outperform most machine-driven algorithms. For now, though, Wikia Search is a small-scale affair and it doesn't yet have the community manpower to give consistently good results every time.

It must be acknowledged that spam and overly aggressive self-promotion could greatly damage Wikia Search's results. I've already encountered some of it, in fact. At the moment, the search results are very easy to game: one high rating will take a page to the top of the listings for many queries, for instance. There needs to be an active community of searchers to keep this under control; it'll be a big problem if spammers discover Wikia Search before the rest of the Internet community does. On the other hand, the very fact that a page can rise to the top so quickly should drive people to Wikia Search. I can totally imagine webmasters arriving in droves to claim their sites' long-tail keywords; that could become a required ritual for all SEO types eventually. So long as the pages are relevant to the query, then this behavior can actually make the search engine better. Spam unrelated to the original query is just bad news, though -- there's no way to put a smiley face on that.

Should Google, Yahoo!, and Microsoft be concerned about this new kid on the search block? To be honest, I think they should. This is a new way to handle search that has some real potential. There is no Google, Yahoo!, or Microsoft search community to compete with what Wikia is building, and it'll be difficult for any of the big search engines to outsource their search results to the public without those results suffering in quality for some time. Just look at Wikia Search right now: a lot of the results are really bad. Google couldn't get away with delivering such bad results and still keep its position atop the search engine charts, but since Wikia Search doesn't have a position to maintain the bad results are OK for now. A set of early adopters are already building up Wikia Search; by the time other people start noticing it it'll probably be much better than it is now. It should be noted, however, that a Google Experimental Search project already has been done which allowed test users to play around with the order of search results and allowed them to rate results positively or negatively. So Google is at least thinking about either community search or personalized search; knowing Google, they're probably thinking of both things. Whether it be a Google killer or not, Wikia Search is quite a cool project that people interested in search should definitely keep watching.