PageRank Explained

Understanding The Importance of PageRank

by John Hughes
of http://www.oyster-web.co.uk

Last updated: 12 Sep 2007

Ask anyone who knows (or thinks they know) anything at all about Search Engine optimisation, and they will have an opinion about Google PageRank. There are as many descriptions about what it is, and ideas about how to improve it as there are so-called SEO experts. How can one filter myth from reality? How indeed can one really improve ranking in Google and will better PageRank help?

Ignoring the Experts and Listening To Google

Go to the following page of Google and read the text - http://www.google.com/technology/. You will see that the second paragraph down categorically states The heart of our software is PageRank and PageRank continues to provide the basis for all of our web search tools.

So, Google’s claim is that whatever PageRank is, it is still the foundation of ALL of their search tools, which would presumably include news search, Froogle, and geographical searches as well as the original Google.com search engine.

Taking Google at face value for a moment, this means that PageRank is currently, and is likely to continue to be extremely important for search engine ranking.

Many SEO experts claim that "PageRank is dead" and "PageRank is no longer important". Personally, I’d rather believe Google.

So What is PageRank?

On the page of Google we have identified, Google describes how PageRank is essentially a conceptual method of using the way websites Link together to deduce relative importance. For example, many people might link to the BBC News website, meaning that it is a relatively important site. Meanwhile, Joe Bloggs’ homepage might only be linked to by a few close friends and family, reflecting that website’s overall importance relatively speaking.

What PageRank does not do, however, is determine what search phrases a website is relevant for, or even how relevant a website is for a particular phrase. PageRank doesn’t know whether a website sells red widgets or purple gizmos.

So How Can I Increase PageRank?

One simple word. Links.

Only by more people linking to your website will your site’s PageRank increase. However, remember that this only increases your site's importance, not its relevance. Increasing your PageRank does not necessarily equate to increasing your search engine positions, especially if Google can not determine what your site is about.

OK, Here Comes The Science Bit…

If you want to understand how PageRank is calculated, it helps if you understand algebra, as really it is a mathematical concept.

Mathematicians with time to spare and an interest in the subject matter can uncover the documentation that Sergey Brin and Larry Page first put together when building Google as part of their research project at Stanford University. Since this time, the principles, and therefore underlying maths of the PageRank algorithm has changed little. The only items that have really changed are the relative constants and damping effects, to adjust how important certain elements of the algorithm are compared to others.

If you are less interested in the advanced maths, but have enough school algebra to attempt to follow the principles, we’ll explain them here.

First of all, let's say that there are X number of links on page A. Page A has a PageRank of P. There is a universal damping effect of d (we’ll explain this later).

So the value of a link "i" on page A is

A_i = d.(P/X)

If a page B has n links to it, its PageRank is the sum of all the values of the links that point to it, plus 1 minus the damping factor.

So the PageRank of page B is

B = (1 - d) + A₁ + A₂ + ...  + A_(n-1)  + A_n

You can see from this that pages that have thousands of links to them, even from relatively unimportant sites, are likely to build a high PageRank for themselves because there are so many link values to add together that even though each is small, as a total, they are large.

Also, pages that have few links need these links to individually have quite a large value to make a significant Impression on their own PageRank.

Why Is There A PageRank Damping Factor?

There are two good mathematical reasons to add a dmaping factor.

Firstly, the damping factor is there to stop PageRank spiralling out of control. Essentially, PageRank accumulates, which means that as more links are created around the web, the PageRank circulates and grows eventually fuelling the Pages which first conceived it. This means that the PageRank values will continue to grow significantly on old web pages, and new and fresh pages would have no way of competing for importance. Therefore, the damping factor is applied to reduce the value of the links slightly so that by the time a portion of PageRank returns to it conceiving page, it is not significant enough to grow the importance of that page.

Additionally, looking at the PageRank algorithm from an applied mathematics point of view, the PageRank algorithm is actually a probability description of a Random Surfer. The mathematic above is often called the Random Surfer Model. It describes first of all the probability of a surfer being on page A, and then of them clicking a link to page B. The damping factor takes account of the probability that the random surfer closes his browser and starts playing Solitaire instead. Of course, that probability has to be assumed, and can not be calculated.

Although Google keeps such damping factors (and other values of constants used in its algorithms) pretty close to its chest, it is generally assumed that the damping factor for PageRank is 15% damping (in other words, each link is only actually worth 85% of its true value). In fact, this is the value that Brin and Page state is most often used in the original thesis.

How Does Understanding PageRank Help Me

Well, that all depends on what you want to achieve. If you just wanted to understand PageRank a little better, then I hope I have helped you. If you want to increase your positions in Google, then hopefully I have explained that the more links you have the better (and the better links you have the better).

However, PageRank is not the only factor affecting search positions. Come back later for articles about improving your Relevance to Google search.