PageRank and the Supplemental Index
 

Back | Print Version

Search Engine Optimisation

Oyster Web Articles

How PageRank Delivers Search Results

PageRank and the Supplemental Index


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

Last updated: 12 Nov 2007

I just finished reading this excellent article by Dan Thies. It is a brilliantly lucid explanation of the way PageRank is used by Google.

For years, the SEO industry has been plagued by misinterpretation of what PageRank actually is and how it is used. As it is relatively complex probability mathematics that underpins the supposed toolbar scores that people bandy about, far too many people ignore its logical basis and jump to wild assumptions.

For example, I have heard people say such things as PageRank is different depending on what keyword you search for, PageRank is affected by the age of your domain and PageRank is not used any more. All of these things are absolutely untrue.

In summary, PageRank uses probability maths (read my other articles about PageRank here and here) to give web pages a score of "importance". I hate using that word for it, but it is difficult to find anything more appropriate. Google then uses that score in two ways when handling searches.

Firstly, in combination with a few key indicators of Relevance (entirely secret, but I guess Meta tags and flat text content of a page are the most likely candidates), Page Rank is used to generate a limited number of documents that use a search phrase. Think of it as Google asking its database "Show me the top (perhaps) 40,000 documents that use the word 'test', sorted by highest PageRank first." That is a relatively simple query to undertake. If it doesn’t get enough results here, it will top up the list by asking a similar question of the supplemental index.

Once Google has its relatively small list of relevant sites, it is then much easier to run its complicated search algorithm on the small data set, giving more accurate and high quality search results.

These principles raise important issues for websites.

Number one – you must have a reasonable PageRank to be in the mix for competitive searches regardless of how relevant your content is.

Number two –once you are in the mix for a search phrase, your PageRank (while still counted) is much less important.

Number three – Supplemental Results are only used for low yield searches, so you need to keep your sites as high a quality as possible to keep in the main index.

Dan's article explains the relationship between the main index, the supplemental index and the use of PageRank in determining search results probably better than I could hope to, so if you are still confused, why not give it a read.

At last, an SEO blogger who understands databases (and therefore the technical aspects of running a search engine)!

Also, if it helps, I have design a handy diagram, which I have included below!

pagerank explained diagram

Search Engine Optimization