PageRank Demystified

Matt Cutts of Google published a very interesting article on his blog titled “PageRank Sculpting” on the 15th of June 2009 that reflects some of Google’s latest views on PageRank. As of today the post has received 352 comments and if you want to print it all out get ready to read through 99 pages!

What follows is a quick summary of some important points raised by Matt. I have also included some illustrations to help clarify certain points.

But, before we get into it…

Let’s just quickly have a look for the benefit of new webmasters what PageRank is all about and what role it plays in Google’s algorithm.

“In general, webmasters can improve the rank of their sites by increasing the number of high-quality sites that link to their pages.”

Source: Google Ranking

 “Make sure that other sites link to yours. Links help our crawlers find your site and can give your site greater visibility in our search results. When returning results for a search, Google combines PageRank (our view of a page’s importance) with sophisticated text-matching techniques to display pages that are both important and relevant to each search. Google counts the number of votes a page receives as part of its PageRank assessment, interpreting a link from page A to page B as a vote by page A for page B. Votes cast by pages that are themselves “important” weigh more heavily and help to make other pages “important.””

Source: Google-friendly sites

“We use more than 200 signals, including our patented PageRank™ algorithm, to examine the entire link structure of the web and determine which pages are most important. We then conduct hypertext-matching analysis to determine which pages are relevant to the specific search being conducted. By combining overall importance and query-specific relevance, we’re able to put the most relevant and reliable results first.

PageRank Technology: PageRank reflects our view of the importance of web pages by considering more than 500 million variables and 2 billion terms. Pages that we believe are important pages receive a higher PageRank and are more likely to appear at the top of the search results.

PageRank also considers the importance of each page that casts a vote, as votes from some pages are considered to have greater value, thus giving the linked page greater value. We have always taken a pragmatic approach to help improve search quality and create useful products, and our technology uses the collective intelligence of the web to determine a page’s importance.”

Source: Google – Technology Overview

From the above quotes it should be clear that PageRank plays an important role in Google’s algorithm and it should come as no surprise that many SEO experts have been trying for years to figure out exactly how PageRank gets calculated by Google.

Ok, let’s get on with Matt Cutt’s blog post on “PageRank Sculpting”, shall we?

The illustration below is a good example of a classic PageRank structure. However, as Matt points out, even when he joined Google in 2000 they were already doing more sophisticated link computations than the classical PageRank structure can account for. But, although we don’t have all the details we do have a basic foundation that’s much better than having no understanding at all on how Google calculates PageRank.

However, please keep the following points, that are not not reflected in the illustrations below, in mind:

i) Larry and Sergey (the founders of Google) introduced a “decay factor” in the past. It basically means that 10% – 15% of the PageRank on a page disappears before the PageRank flows out to other pages.

In other words, in the illustration below Site A has a PageRank of 100 points and two external links to Site B and Site D. In our example both these sites get 50 points each from Site A. However, thanks to the decay factor the actual points will be less.

ii) Not all links will pass the same amount of PageRank. For example, returning to the illustration below, Google may decide (based on various factors) that the link to Site B should carry more weight than the link to Site D. One reason may be that Site B’s content is more relevant than that of Site D in relation to Site A.

1. Position before the introduction of the rel=nofollow attribute in 2005.

In this “classical” PageRank model, Site A has a PageRank of 100 points. Site A has two (external) links to Site B and Site D meaning that Site B and Site D will each receive 50 points from Site A.

2. Position after the introduction of the rel=nofollow attribute in 2005.

In 2005 Google introduced the rel=nofollow attribute. Basically, by using the nofollow attribute you are saying to Google that although you are linking to another site that you are not vouching for the content on that site and that they should not see it as a “vote” from your site to that site.

NoFollow links do not pass PageRank as far as Google is concerned.

In the above example Site A still has 100 points of PageRank and two (external) links but one of them is a NoFollow link. The result is that the DoFollow link that points towards Site B passes 100 points of PageRank to Site B and the NoFollow link that points towards Site D does not pass any PageRank.

3. Position since 2008.

A NoFollow link still passes no PageRank BUT has a direct impact on the PageRank that other DoFollow links pass on.

In our example above Site A still has 100 points of PageRank and two (external) links, a DoFollow link and a NoFollow link. However, whereas previously the NoFollow links were completely discarded for PageRank purposes they are now counted (as they were before the rel=nofollow attribute was introduced in 2005) except for the fact that they do not pass PageRank.

Why has this change in 2008 made many webmasters mad?

Well, many webmasters are worried about so called “PageRank leakage.” They are concerned that PageRank will “leak” from their sites if they have any or too many external links to other websites and that this may compromise their rankings in Google.

Once again it’s the easiest to explain this point by looking at the situation before 2005, the situation after Google introduced the rel=nofollow attribute in 2005 and the change Google made in 2008.

For the sake of example we shall assume that a webmaster has a home page and a blog on his own website…

1. Position before the introduction of the rel=nofollow attribute in 2005.

In our example the home page has a PageRank of 100 points with only one link to the blog. The home page passes 100 points of PageRank to the blog. The blog has four links, one back to the home page as well as three external links.

In this example the website is leaking PageRank through the three external links as the blog only passes 25 points of PageRank back to the home page.

2. Position after the introduction of the rel=nofollow attribute in 2005.

When Google introduced the rel=nofollow attribute in 2005 many webmasters welcomed it. As can be seen in the above example, the home page still passes 100 points of PageRank to the blog and the blog still has four links, one back to the home page and three external links. But, since the three external links are NoFollow the one DoFollow link back to the home page gets all the PageRank points and there is no PageRank leakage.

3. Position since 2008.

Many webmasters have build sites since 2005 on the assumption that NoFollow links do not cause  PageRank leakage and only found out in June 2009 that Google changed the rules in 2008 and that these NoFollow links do have a direct impact on the PageRank points that are passed by DoFollow links, as can be seen in the illustration below.

The website structure has remained unchanged but the one DoFollow link back to the home page now only passes 25 points of PageRank and no longer 100 points as was the case before. In other words, the three NoFollow links (although they do not pass any PageRank themselves) are taken into account when the PageRank that the single DoFollow link passes is calculated.

Is this really such a big deal?

One person who commented on Matt’s blog post used the phrase “what a nightmare will now be unleashed” to illustrate his frustration to which Matt Cutts replied that it has been “running quite well for over a year now.”

Matt Cutts also said that Google thought some people running tests would have noticed the change in 2008 but they didn’t. So is it really such a big deal?

Well, in many ways “PageRank leakage” is an unfortunate term… It nearly sounds as if a page with 100 points of PageRank will lose (“leak”) all of its PageRank and become less important in the eyes of Google even if it has only one external DoFollow link to another page. This is not necessarily the case!

It’s clear that Google may penalize a site that links to “bad neighborhoods”…

“Don’t participate in link schemes designed to increase your site’s ranking or PageRank. In particular, avoid links to web spammers or “bad neighborhoods” on the web, as your own ranking may be affected adversely by those links”.

Source: Google – Quality Guidelines

It’s also clear that Google has a problem with paid links that passes PageRank…

“Some SEOs and webmasters engage in the practice of buying and selling links that pass PageRank, disregarding the quality of the links, the sources, and the long-term impact it will have on their sites. Buying or selling links that pass PageRank is in violation of Google’s webmaster guidelines and can negatively impact a site’s ranking in search results”.

Source: Google – Paid Links

“PageRank leakage” is not necessarily a bad thing…

According to Matt Cutts, Google encourages and rewards sites that link to other high-quality sites!

He also suggests that webmasters should let PageRank flow freely within their own sites and does not recommend “PageRank sculpting” (trying to manipulate how PageRank flows within your site).

Trust you have enjoyed this post and I hope that it has helped to demystify the concept of PageRank at least a little bit. :) Please feel free to leave your comments. I welcome comments from my blog readers, read of all them and will try to reply to all comments when necessary.

All the best,

Francois du Toit