Author: Bob Wakfer (2004)
PageRank: The Background
Google's PageRank (PR) is one of the most sought after, and yet misunderstood, web page attributes. PageRank, named after one of the founders of the Google search engine, Larry Page, was the innovative foundation that the Google search engine was built on.
The theory was that a link from one web page to a web page of another site
was in essence a vote for that page. The reasoning was that webmasters
would only link to pages that they thought were interesting and of value
to their
viewers. Google used the number of inbound links (IBL) to a page to judge
the importance and relevance of that page, and based on this calculation,
and other
factors, decided where to place that page on the search engine results
page (SERP).
They devised a scale of measurement for PageRank from 1 to 10. Then for
the information of webmasters and interested people they produced a
toolbar that
can be deployed in Internet Explorer that will indicate the PageRank
value of any page being viewed in the browser. These values have become
known
as PR0 to PR10. Note: A reader has pointed out an apparent contradiction
between
a scale of 1 to 10 and a reporting of PR0 to PR10. In fact PR0 indicates
no PageRank. One cannot be said to have PageRank until the site achieves
a PR1
ranking.
Since PR values are a result of IBLs, Google decided to give them their
own name and refers to inbound links as backlinks. As part of the
toolbar there
is a quick lookup of the number of backlinks that Google reports
for the page that is currently being viewed in the browser. This search
can also
be done
without the aid of the toolbar by simply typing link:http://www.yourURL.com" into
the Google search box.
The one trick to this link search is that Google does not display
all backlinks. At one time it was thought that they only listed
pages with
a value of
PR4 or greater. Today however, you will find backlinks reported
from pages of lower
PR values. So, at best, Google's backlink search seems to present
some sample of pages linking to the site. Suffice it to say that
this search
is not a reliable
measure of all IBLs to a page.
How is PageRank calculated?
In simplest terms PR is calculated by the sharing of PR from all
the IBL links to your page. This is not strictly accurate because
Google
also uses
the internal
links within a site in the calculation of PR. Each link to
a page carries with it and passes PR value to the target page. The PR
points or value
passed depend
on the PR value of the page they come from, and the total outbound
links from the page. It is generally agreed that a page will
only pass about
85% of its
value to the page it links to. So a PR5 page with a single
outbound link will pass 85% of the value of a PR5 page to the page it
links to.
But virtually no page has only a single link -- remember internal
links are also used in the total outbound link count -- so
the value passed
to any page
is 85% of the PR, divided by the total number of outbound
links.
The question now becomes what is the PR point value of the
different PR levels. Most observers believe that the relationship
between
PR levels is logarithmic
rather than linear. In other words PR5 is not worth 25% more
than a PR4, but may be worth 4 to 6 times more.
It is also understood that a PR value is not a single number,
but is in fact a range of values. So not all PR6 ranked
pages are equal.
As
the chart
below
shows a PR6 maybe just on the upper boundary of a PR5 or
it maybe just short of the entry point for a PR7.
The chart that follows shows the range for each PR value.
It also shows how much PR value or PR points a page with
50 outbound
links
will pass
depending
on of its own PR rank. From this I have calculated the
number of links required from each value of PageRank
necessary for
a page
to attain
a desired page rank.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The assumptions and the mathematics:
For those who are interested I have used logarithmic values of base 5.5.
In other words the value range for a PR1 lies between 5.5 to the power
of 1
and 5.5 to the power of 1.99, and PR2 lies between the value of 5.5 to
the power of 2 and 5.5 to the power of 2.99 etc. The rest of the chart
is fairly
straight forward. It assumes that there are 50 links per page and that
85% of the PR value is passed to the recipient page.
The number of links required to attain any ranking is based on the median
value of the donor page and the entry threshold of the desired PR value.
In other
words to achieve a PR5 you need 5,033 points and the average points available
from a PR6 page with 50 links is 1507.
The chart was calculated with an Excel spreadsheet and it can be downloaded
if anyone wants to play with the calculations and assumptions. It might
be interesting to work with a different base number for the logarithmic
calculation.
And it is also interesting to see the impact of more or less outbound
links from a page.
Are these findings valid?
Nobody knows for sure how Google calculates PR. I have shown this chart
to a number of knowledgeable people and they have all agreed that
my calculations look reasonable. One search engine optimization (SEO)
guru from a major
firm
said the results were very similar to independent research that his
firm had conducted.
So take it or leave it. It is probably a fair reflection of how PR
is passed and accumulated.
The lesson that can be drawn from this explanation and chart is
that if you want to increase your PR you need a few links from
pages with
equivalent
or higher PR, or a great many links from sites with lower PR.
For more information on methods for actually accumulating links
see part 2 of this article.
About the Author
Bob Wakfer is the owner of Computer Partners and a long time
student and practitioner of SEO. Computer Partners offers web
site design,
hosting and search engine
optimization for its clients. If you would like to discuss
any of these services with Bob you can email him at bob@compar.com.