Wednesday, October 8, 2008

This is the only chapter of the book that deals pretty much with theory only. However,
because of the general confusion and hype over Google PageRank, it is important
that you understand how PageRank works and what role it plays in achieving
rankings.



Note: In mid-November 2003, Google introduced a major algorithm change during
the so-called “Florida Update” that changes the way Google ranks certain websites.
It is believe these changes mainly involve how PageRank is computed – more
accurately that PageRank is being phased out by a new “LocalRank” system of
computing PageRank dynamically on the fly. At this writing, this new algorithm is still
changing, with some of the effects not well understood. For more information, see
the applicable section Appendix D – About the Florida Update”.

PageRank, or PR, is a numeric value that Google places on how important a page
is on the Web. PageRank is determined by how many incoming links there are to a
page. Google figures that when one page links to another page, it is in effect
“casting a vote” for the importance of the other page. The more incoming links
(votes) there are for a page, the more important the page is to Google, generally
speaking.

Many people obsess and over-hype the importance PageRank and therefore
introduce worry and confusion that is not warranted. It is but a single factor in
the overall ranking algorithm. This may be because it is the only part of the Google
ranking algorithm that is made public – well sort of.

You can see what the “public” PR value of a page is in the Google Toolbar:







PageRank vs. Search Result Ranking

Some people confuse Google PageRank (PR) with their page’s ranking for a certain
search result for a certain keyword. Google PR is just one factor that is used to
determine your page’s actual rank (position or placement) on a search results page
for a given search query.

It is not uncommon to see a page with a lower PageRank value that is positioned
higher on a search results page than a page with a higher PageRank. This shows
that PageRank is not the most important factor in Google’s ranking algorithm. A
properly keyword-optimized site with a lower PageRank can outrank a non-optimized
site with a higher PageRank.

This is a common scenario for large corporate sites. The corporate site may have a
high PageRank as a result of the large number of other sites that link to it, but they
end up being outranked due to their lack of keyword optimization for their pages.

Toolbar PageRank vs. Actual PageRank

So how does one find out what the PageRank (PR) value is for a given page? Well,
Google has a tool, called the Google Toolbar, which allows you to see a crude
approximation of PageRank value. Download and install the free Google Toolbar at
http://toolbar.google.com/. Once installed, it should look like this in your browser:






What most people fail to realize is that the PageRank values shown in the Google
Toolbar are not the actual PageRank values that Google uses to rank web pages.
The Google Toolbar is divided up into 10 equal linear ranges from 0 - 10. These
linear divisions correspond to a logarithmic scale that Google uses. The actual
scale is estimated to be anywhere from log base 5 to log base 10. This public
Toolbar PageRank is however what people talk and agonize about.

The Toolbar PageRank value only indicates that a page is in a certain range of the
overall scale. One PR=5 page could be just above the PR=5 division and another
PR=5 page could be just below the PR=6 division, which is a really vast gulf between
them.


Increasing PageRank

Each page of your website has a PR value, and as such you can simply add up the
individual PR values of each page to arrive at the total PR that your site has (Bear in
mind however that someone speaks of PR, it applies to a page). How you structure
your internal links can influence to some extent what the PR value of a page will be,
as will external links pointing to a page on your site. Although page PR value is
important, you should really be trying to increase your total site PR value.


There are only two ways to increase your “site PR” value:

1. Get more incoming links that point to pages on your site.


2. Add new pages to your site (which was discussed in a previous chapter).

The actual PR value of each page indexed by Google on the Web is in constant flux.
All over the Web, new pages are added, old pages are removed, more links are
created – all of which over time decrease the “value” of your incoming links.

As the number of websites (and web pages) in Google’s index increases, so does
the total PageRank value of the entire Web, and so also does the high end of the
overall scale used. This is kind of like the top student setting the “curve” for an exam
at college. The top student gets 100% and everyone else gets correspondingly less.
Therefore, the top-ranking site (or handful of sites in actuality) gets the maximum,
perfect PageRank score (which is a 10 in the Google Toolbar) and everyone else is
scaled down accordingly. As a result, some web pages may drop in PageRank value
for no apparent reason. If a page's actual PR value was just above a division on the
scale, the addition of new pages to the Web may cause the dividing line to move up
the scale slightly and the page would end up just below the new division.

What this means is that you should always strive to get more links that point to your
site, otherwise your site can naturally start slipping in rankings due to this decay of
PageRank value for incoming links – both from other pages on your site as well as
from other websites. This is also why you should add new pages to your site on a
regular basis, as additional pages will increase your site’s total PR score too.


The PageRank Equation

OK, I know you’ve waiting for it, so here is the official PageRank equation. It is
calculated by solving an equation of 500 million variables and more than 3.3 billion
terms (web pages):


PR(your page) = 0.15 + 0.85 [(PR(page A) / total links (page A) ) + (PR(page B) / total links (page B) ) + …]


There are a couple of observations to note about the PR equation:

• PR values for based on individual web pages.

• The PR value of each page that links to your site in turn is dependent on the
PR of the pages that link to it, and so on backwards.

• A link’s value (amount of PageRank or “voting power” forwarded to the linked-
to page) is at most only 85% of the linking page’s PageRank value, and this
value is diluted (decreased) by the number of other links on that page.

• PR has nothing to do with keywords or text in links - it is purely dependent on
link quantity and link strength, as discussed previously.

Some may incorrectly conclude that a link from a page with a PR=4 with only a few
outgoing links is worth a more than a link from a page with a PR=7 with 100 outgoing
links because for the latter, the “voting power” or value is divided up among 99 other
links.

However, you must remember the logarithmic nature of the true PageRank. This
means that a link from a PR=6 page that has lots of outbound links may be worth
more than a link from a PR=4 page that has only a few outbound links. Whether this
is true is dependent on the actual log base used for the PR equation, which is a
secret.

Do not get caught up in the minutiae of determining whether a site is worth
exchanging links with. Barring link farms, Free-For-All (FFA) sites, sites with a true
PR of 0 (which either aren’t indexed or have been “blacklisted” by Google), and sites
that have nothing to do with your theme, you should strive to get more links that point
to your site - period.

Google performs multiple iterations of the PageRank equation for each page in it’s
entire index to determine the new PageRank value each month.


PageRank Being Replaced?
Google has done a major reweighing of the ranking factors it uses in the recent past,
in particular how it analyzes links. Links are being scored in a new way and some
links just may not count for as much or not at all.
PageRank (PR) is calculated once a month for every page in the Google index. The
PR score for a page is pretty static until the next time around that PR is calculated,
and is computed based on ALL links that point to a page.
There is evidence to suggest that Google will be replacing PageRank with
LocalRank, which computes a link score based only on incoming links from pages
that are returned from a given search result set that matches the search query
(whether the result set is 100, 1000 or 10,000 pages is not yet known).
Which means that a flower site only gets links counted from other sites that are
related to flowers - not from sites that are about mortgage loans for example. This is
similar to how the Teoma search engine does it.


Think of LocalRank has a dynamic PageRank score that is computed on the fly each
time for a given search query
PageRank has been abused as more people try to get as many links as possible to
their site – regardless of whether the other site has anything to do with the topic of
the linked-to page. By using LocalRank, Google hopes to filter out irrelevant links
that has thrown the value of PageRank off in the past.

0 comments: