Google PageRank complexity still on the rise

Angus Kidman
Tuesday, 24 April 2007 20:28

Your IT - Home IT

Generating more links to a site to improve its visibility in Google's search results is one of the most widely discussed online promotional tactics, but the problem of how to do that could be more challenging than anyone previously suspected.

Google likes to keep the exact methods used to rank search results a closely-guarded secret, and constantly tweaks its systems to try and thwart sites which manipulate their content purely to achieve a high visibility.

One of the better-understood of Google's techniques is PageRank, which in simple terms gives greater prominence to sites which themselves are linked to by large numbers of other sites. To stop site developers from creating lots of links to their own sites through other sites built purely for that purpose, PageRank also considers issues such as the "importance" of those linking sites.

Working out how to manipulate these results has spawned an entire sub-industry of search-engine optimisation, as well as being responsible for large volumes of comment spam on blogs and other sites.

During a presentation at Hyperion's Solutions 2007 conference in Orlando, Google technical program manager Chris Schulze pointed out that PageRank these days is a considerably more complex beast than such spammers give it credit for. More than 200 signals are used to determine the overall ranking of a page, he said.

Schulze didn't give details of what the 200 elements were, but their mere existence suggests an even more complex algorithm than the basic methodology for PageRank proposed in Wikipedia's entry on the subject. With PageRank having jumped from considering one basic factor (inter-site links) to more than 200 in ten years, future ranking schemes seem certain to be even more complex.

Disclosure: Angus Kidman visited Florida as a guest of Hyperion.