SEO Master Class: The Mathematics and Operation of Google PageRank
The following is an excerpt (with some recent modifications and editorial comments) from our book, WordPress Search Engine Optimization. You can buy the book at Amazon.
The Mathematics and Operation of Google PageRank
Google’s PageRank is part of its search algorithm; the other search engines’ ranking algorithms work similarly. Yahoo and Bing, while they obviously measure inbound link counts as a ranking factor, do not disclose to web users any measure of page value equivalent to PageRank. PageRank works through complex mathematics. Understanding the mathematical intricacies is not vital, but can help illuminate how PageRank impacts your link building efforts. PageRank works the same on all platforms, WordPress or otherwise.
The PageRank Calculation
PageRank calculations works as follows: Google assigns a numerical value to each indexed page on the Web. When an indexed page hyperlinks to another page on the Web a portion of that numerical value is passed from the linking page to the destination page, thereby increasing the destination page’s PageRank. Inbound links increase the PageRank of your web pages and outbound links decrease PageRank. PageRank, often abbreviated as “PR,” is expressed as a number from 0 to 10. Google.com and Facebook.com, both of which benefit from millions of inbound links, enjoy a PageRank of 10. In common parlance, a PageRank 10 site is referred to as a “PR10 site.” Remember though that PageRank refers to pages on the web, not just sites themselves. A PR5 site simply means that the site’s front page is a PR5.
So how is PageRank specifically calculated? Every indexed page on the web enjoys a small amount of PageRank on its own, a PageRank score of 1. This inherent PageRank is the original source of all PageRank on the web; it is only through linking between pages and sites that some pages accumulate higher PageRank than others. However, a page can never send all of its PageRank to other pages—this is where the damping factor comes into play. The damping factor is simply a number between 0 and 1 (but think of it as zero to 100 on a percentage scale); it represents the amount of PageRank that can be sent away from a page when that page links out to other pages.
If a search algorithm’s damping factor were set to zero, no page would ever send PageRank away, and the entire PageRank calculation becomes pointless. On the other hand, if the damping factor is set to 1, then 100% of a page’s PageRank is sent away through outbound linking, and any page with any outbound links retains no PageRank. In this case, the algorithm also fails—the internet would be populated entirely sites of either PR0 or PR10 with no sites in between. As it happens, the damping factor employed by Google is widely believed to be .85. This means that 85% of a page’s PageRank is available to be passed to other pages through linking, while 15% of a page’s PageRank will always be retained. It is believed that Google can alter the damping factor for particular sites.
Consider for a moment that Google manages PageRank calculations for billions of web pages. If that wasn’t daunting enough, consider that Google undertakes the even more staggering task of managing the mathematical calculations of immeasurable numbers of links between those billions of sites.
PageRank, Diagramatically
You Have to Share Your PageRank
Also bear in mind that the amount of PageRank available to be passed by a page will be equally divided among all the outbound links on that page. So, if a webpage has a total of six links: three internal links and three external links (links to outside websites) then the PageRank passed away by that page will be shared equally among the six links on that page.
What does that mean for the link builder? Well, it means that if you have secured a link on a great PR4 page, but that page has 200 outbound links, then you’ll be sharing the available PageRank with 199 other sites. That’s why you want to seek out pages with low numbers of outbound links. When there are fewer outbound links, your link will enjoy a much greater percentage of the available PageRank.
The Logarithmic PageRank Scale
If the mathematics underlying PageRank weren’t complicated enough, there is another facet that you must consider. The PageRank scale of PR1 to PR10 isn’t linear, it is logarithmic. Therefore, it takes ten times as much linking power to rise from a PR2 to a PR3 page. Expressed another way, a PR4 page has 100 times the linking power of a PR2 page. As each level of PageRank is reached, it becomes harder and harder to reach the next level. There are only about 120 to 150 PR10 pages at any given time, and generally this elite class of pages and sites includes Google.com, Microsoft.com, WhiteHouse.gov, and other sites of equivalent popularity and character.
PageRank Is Historical
PageRank is historical and only updated every three months or so (although sometimes much longer periods pass between PageRank updates, it’s really up to the whim of Google)—when you check the PageRank of a page, you aren’t seeing the current PageRank, you are seeing the PageRank reported as of the last PageRank update.
Leave a Reply
Want to join the discussion?Feel free to contribute!