Wikiasari: Wikipedia success applied to social search?

Wikia Search icon Wikia will release a new search engine early next year according to an interview with Jimmy Wales in today’s Times of London. The new Wikia search engine project is named Wikiasari and will apply wisdom of the crowds features to search engine results, letting individual users rank sources of information and their relevancy to a particular query.

Of course the article takes a “gunning for Google” angle, citing the PageRank algorithm used since Google’s was founded in 1998. Search engines grow over time, and incorporate multiple ranking factors beyond the math of inbound links and source authority. Google (synonym for big search engine for simplicity’s sake) can assign domains of trust to highlight trusted content beyond their PageRank calculation. The Mayo Clinic might be a trusted source of health news. Government domains could be an authority for government searches. External links found in Wikipedia could carry additional weights as curated sources.

Google also incorporates click through rates into its advertising algorithms, relying on the preferences of a crowd to select the most relevant result. Google presents a searcher with a title, contextual summary, and domain name to help him or her select the result best matching their query. Click-through tracking can be grouped on a personal level (search results you previously visited), geographic level (popular in San Francisco), network-specific (other people on your corporate network liked these results), within an affinity group (search originates at Sierra Club or through a Google Co-op group), and much much more.

Wikia and its investor Amazon may have an edge incorporating a user’s purchase history, news preferences, and other profiling data into each search. You could place a set of eyes on the billions of web pages currently in existence, hoping that new stem cell review center achieves appropriate annotations for discovery, but I’m skeptical. Sites such as Google, Yahoo!, and Windows Live already have the crowds clicking on search results every day, submitting bookmarks, and, in some cases, flagging spam. Wikia would need a critical mass of users to maintain a useful search index and query analyzer to supply Britney Spears’ fans, medical research, and the many many other search queries submitted every day. The same same search engine pickpockets wandering through Google’s search index will continue to target any significant source of traffic and unlike Wikipedia, you can’t just lock down a contested (or heavily profitable) area and still maintain balance.

  • Posted
  • Updated at
  • Comments [6]

6 comments

Commentary on "Wikiasari: Wikipedia success applied to social search?":

  1. useAPI Search on wrote:

    Hope that Wikiasari will be free of any bias with its editors …

  2. Greg Linden on wrote:

    Chris Sherman at Search Engine Watch had a good critique of these kinds of social search efforts a few months ago.

    http://searchenginewatch.com/showPage.html?page=3623153

    In my opinion, the biggest problem is that social search by itself does nothing to help filter spam. If anything, it makes the problem worse by giving users more ability to manipulate rankings to drive traffic to their own sites.

  3. Jimbo Wales on wrote:

    The Wikia Search project homepage explains… Amazon has nothing to do with this. And Wikiasari is just the name of the software, not the name of the search engine. :)

  4. Joe Agliozzo on wrote:

    The biggest effect social search can have on search as currently constituted is to “bring out” pages that are way down the Google index but are nevertheless highly relevant. Think about how few search users EVER get past the first page (or at most two pages) of results. Sure, algorithmic search is supposed to make sure that the most relevant results appear on page 1, but we all know that it is typically the most “commercially” relevant results that appear there. The sites that can afford to “play the game” and optimize for the flavor of the month (or week) will show up there, usually selling something as well. If social search, through tagging or other human elements can highlight these otherwise “lost” relevant pages, it will be a great benefit for users and if it is a benefit for users, these efforts will succeed.

  5. Sarah on wrote:

    At the very least, it should be an interesting experiment. I’m curious to see how well it will work.

  6. Keith Eysmun on wrote:

    I don’t think developers are going to want to work on an open source search engine. This whole premise just sounds absurd to me.