Via an excellent article by Bill Slawski at SEO by the Sea, a patent has been filed by Microsoft for a search engine ranking algorithm that uses new technology to identify whether or not a page being crawled is a blog. If it is indeed a blog, Microsoft’s algorithm will reward any outgoing links to a greater degree than non-blog linksm, increasing the PageRank of sites receiving these links. Therefore, Microsoft’s new algorithm will make any links originating in blogs more valuable for websites than non-bloglinks.

We already know that Google likes blogs, and that when it comes to SEO, developing bloglinks is one the most reliable and easy way to raise a site’s search engine ranking. What Microsoft is developing only underlines a thought process that has been going on for quite awhile now: skewing PageRank in favor of links originating from blogs gives the searcher more relevant results.

Microsoft isn’t doing this for for the heck of it. Their goal is to create a search engine that gets the best results. To paraphrase the semi-annoying Volkswagen ads, “It’s what the people want.” As the patent application itself states:

This disclosure is directed to a modified scoring technique whereby the score is biased toward web pages linked to blogs. The notion is that blogs still represent human created content, and that links from blogs are genuine endorsements of the target web page.

How will the algorithm go about identifying the blogs? Again, Microsoft’s patent app sheds some light:

These features are generally organized into four categories: (i) whether a page is hosted in a known blog hosting DNS domain; (ii) features from the non-HTML markup words and phrases contained in the page; (iii) the targets of outgoing links in the page; and (iv) whether the string “blog” occurs in the URL. In addition, human judgments of results from the web crawl can be used as input to the web crawler strategy to improve the probability that web pages could be identifies as blogs.

Microsoft is wary of “commercial” websites that generate a high volume of links with the intent of artificially increasing PageRank for the sites it points to. The patent also mentions a penalty for sites such as these, making a high volume of identical links coming from one location less potent than intended.

It can thus be seen that an agent or commercial operator could generate a very large number of web pages that reference a target page, thereby boosting the PageRank score of the target page. One countermeasure is to use a non-uniform reset vector b that is weighted towards pages that are perceived to be trustworthy. The key question is how to appropriately choose the non-uniform reset vector b.

So will search engines continue to reward links from blogs? It appears that the technology is headed in that direction, and the impact on SEO will big.

Thanks again to Bill Slawski at SEO by the Sea for picking up on this story.

This entry was posted on Monday, October 20th, 2008 at 12:08 pmand is filed under Search Engine Marketing. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.