Saturday, December 27, 2008

Google Page Rank Explained

One of the criteria used by Google when displaying search results is the Page Rank (PR). The higher the PR, the higher a page will be shown in search results. In this article we explain how Google calculates he PR and how you can optimize your website for achieving a high PR. Google also uses factors like relevance for the search - this is not discussed in this article.

This article is based mainly on an original article by Google founders Sergey Brin and Lawrence Page but also on hearsay, speculation and rumours. Google is not interested in allowing the public to know exactly how the PR is calculated since it is then easier to trick or cheat the search engine and since this would undermine its credibility. It does however appear that a variation over the basic formula is still used.

Google assigns a PR of between 0 and 10 to all pages which it indexes. It is possible to install a Google toolbar which shows the PR when you visit a page. The toolbar can be downloaded from http://toolbar.google.com/.

In reality, the PR is a value between 0 and "very large number". Algorithmic scale is most likely used to translate the value into the PR we know. A logarithmic scale works so values may be translated as follows:

Values 0.1 to 1 results in PR 0
Values 1 to 10 results in PR 1
Values 10 to 100 results in PR 2
Values 100 to 1000 results in PR 3 etc

What we'll look at in this article is how the value is determined. The actually scale used is not important to understand the concepts and it is not known to the public. It is likely that the scale is changed regularly.

Before looking at the actual formula - which may seem a bit scary - we'll discuss the basic idea by which the founders of Google wished to set importance - expressed in PR - to webpages.

The basic idea is that a link from a page A to a page B indicates that page A casts a �vote' for page B. The link from page A to B indicates that page B has something important on it which means that page A endorses page B. The basic idea of Google PR is that the more votes a page have, the higher the PR it should get. There are however different weights assigned to the votes cast. A vote from a page with a low PR is less important than a vote from a page with a high PR and therefore has less weight. If a page A casts many votes to many pages by having links to a number of pages, then the vote of page A is divided into small parts and each part is then assigned as a vote to each of the pages which page A links to. As you may have deduced from this explanation, the PR then expresses - to some extent - the likelihood that someone surfing randomly on the web will end up on the given page. A probability is therefore expressed in the PR.

Before looking at how this works out, we'll look at the exact formula used. You need not understand the formula as long as you understand the basic idea, but we will add it here for completeness. We'll look at what the basic idea and what the formula means in terms of getting a high PR a bit further down.

The basic formula is the following, where A is the page we wish to find the PR for and pages T1 to Tn are the pages linking to page A:

PR(A) = (1-d) + d (PR(T1)/C(T1) + PR(T2)/C(T2) + ... + PR(Tn)/C(Tn))

Where:
1) PR(A) means the PR of page A. Subsequently, PR(T1) means the PR of page T1
2) C(T1) is the number (count) of links going out of page T1.
3) PR(T1)/C(T1) - it follows from 1) and 2) that this is the page rank of the page T1 divided by the number of outgoing links from page T1. In other words, PR(T1)/C(T1) expresses the part of the vote of page T1 that is awarded to page A.
4) d is a damping factor which is probably set to around 0.80 to 0.85. We will not look into details as to why this factor is needed, but just state here that it has to do with probability distribution.

You may want to make a few examples and you'll see how it works out.

If you are bright, you may have noticed that you need to know the PR for all the pages linking to page A in order to work out the PR for page A. This may however not be possible as some of the pages are new and were never indexed before by Google. For this reason, Google assigns an initial value to all pages and after a few iterations where the PR is recalculated, an appropriate value is determined.

From the description of the basic idea and the formula we are now able to look at the factors which result in a high PR and thus give guidelines for achieving this.

The more webpages linking to your website the more votes will be cast upon your website, so the higher PR you will get. Links from pages which a high PR and with few outgoing links count higher than other links. How to achieve this is the subject of many articles on their own and we will not dwell into this further here.

If you have links to external sites from a page A on your own site, then you will be casting your vote to other sites. If however you also have links from page A to your own other pages, then you will be sharing the vote for page A with external sites and your own pages. This way, you will keep part of your own vote within your own site. For this reason, it is recommended that you link from each of your webpages to your main page or even link from each page to all your other pages.

Note that Google may penalize a website which tries to achieve a higher PR by what Google considers an inappropriate means. If for example you purchase links from a link exchange provider, Google may see this as a form of spamming and reduce your PR as a result. Playing fair seems to be safer than trying to "cheat"

In this article we have discussed how the Google page rank is calculated and what you can do to increase your page rank. Equipped with this information you may be able to improve your presence in search results.

The original paper by Sergey Brin and Larry Page can be found here: http://www-db.stanford.edu/~backrub/google.html

6 comments:

Bram said...

nice post bro !

tukeran link ama gue ya ^^

bramzzz.blogspot.com

1/2 Price Pro Winnipeg Painting said...

This article should be updated as almost everyone is well aware that google doesnt rank serps results by page rank =/ thats why those two pr4 guides on google for how to make money outrank bidvertizer whos got like pr8 =/ its all about the backlinks =/

Nandu Green said...

Thank you for your informative post - I am a newbie and it is all most confusing.

Mohsin Ibrar said...

Very Good Post
www.mnewsalert.blogspot.com
www.supirscience.blogspot.com

Unknown said...

too good buddy...:)

http://makemoneyonlinegoals.blogspot.com/

concarrier said...

Affordable SEO
Very Useful Information

Thanks Bro.........