|
|
|
Improve Google Page Rank
s Page Rank Calculation To justify our assertions more
still following the rigorously, let's first quickly
"BackRub" Algorithm? An introduce the Page Rank
examination of the PR algorithm, Calculation.
with examples from real world Page rank measures the importance
sites. of a page using the Google page
rank formula:
Many a time and oft, we hear the
speculation that the Page Rank PR(A) = (1-d) + d (PR(T1)/C(T1) +
Calculation has changed. ... + PR(Tn)/C(Tn))
Disappointingly, it is voiced by The page rank shown on a tool bar
people who should know better. is actually a scaled number
Perhaps the green monster of between 0-10. The scaling is
jealousy at seeing a spammy logarithmic to accommodate the
(competitor) site or a "google widely different number of IBL's,
bomb" taking a site to PR-7 gets where a linear scale would not
the better of their judgement. We convey the information
will proceed to shatter a few appropriately.
urban legends.
While we are not privy to the log
1. Home page is necessarily a factor "lf" that Google uses for
higher PR than the rest of the it's Page Rank tool bar
site. normalization, we take a stab at
2. Older pages get highest PR. the number by making a few
3. Google considers "on page" assumptions. We assume that
factors or contents to judge PR. Google is the highest ranking
4. Pages with most external link PR10 page. We use the number of
backs get the highest PR. links incoming to Google reported
by Google itself. On March 29,
2006 it shows it to be 3,750,000 PR-5 < 24,248
links. We further assume that the
average Page rank of the links Naturally, some pages are a very
pointing to the Google home page strong PR-5(i.e. almost a PR-6),
is PR 1. or a very weak PR-6. Therefore,
we allow a user to set the
Therefore, to obtain the upper strength of the page. The default
bound of the log factor("lf"), we is 0.5(i.e. it's the equivalent
just take the appropriate root. of a PR5.5 page).Now, if the
(e.g. ((Incoming Links to above holds, then "1", "2" or "3"
Google)^(1/(Google are irrelevant. Our contention is
PR-Average_Incoming_PR)) ) ). that Google has not changed the
Page Rank calculation in any
This boils down to significant way, aside from
(3.75*10^6)^(1/(9)) or lf=5.38. tweaking the constants.The proof
Naturally, as the size of the web is obviously in the pudding.
increases, or the number of back Let's take three sample sites and
links that Google exports to see how the update of April 6,
outside world increases, "lf" is 2006 has affected them.
likely to increase. Our tool
allows you to set a different We turn to our Event Tickets
"lf". Website and examine two pages on
the site:
We use the following formula to
scale numbers to the "RPR"(real 1. Tickets Website's Home Page ,
non normalized PR) and obtain the PR-4
lower and upper bounds. 2. New Tickets Events Page , PR-6
((lf)^(n)) =< RPRn < ((lf)^(n+1)) The comparison is indeed telling.
As an example when n=5, 4507 < A page that was added a few weeks
ago attained the highest PR in 2. New Tickets Event Page , PR-5
the site. While the home page
with most of the back links Coincidence, you say? Let's look
stayed at 4. The reason is quite at a third site
obvious if you examine the actual
pages. The home page is donating 1. Click Fraud Website's Home
it's PR to a lot of pages, while Page —PR-4
the new events page has only 3 2. Click Fraud Website's Internal
outgoing links(in addition to the Page —internal pages like click
common links). This behavior is fraud with no link to Zaralyzer
predictable, if we apply the but lots of links from external
"backrub" algorithm. sites, PR-5
Happenstance, you say? Let's look In conclusion, even the most
at another site, with less trivial changes in site
contents and a different navigation structure can cause
"offer"/e-commerce pricing model. major changes in how the search
The common feature is that the engine view your pages. For a
navigation structure is exactly free Google Page Rank Improvement
the same. Consultation on how you can
improve your site structure.
1. Tickets Website's Home page ,
PR 3
About the Author:
Ron Arthur is a Search Engine Marketer working for Carlsbad, CA based web-metrics company Sofizar. He is a member of the team developing a click fraud detection software, ZarTective. While not writing expose's on the darker side of the web, he plays with his cat "Mano" and watches "Rocky Horror Picture Show" for the 17th time. Or maybe 117th. For More Information visit http://www.sofizar.net/improve-google-pagerank.php
Read more articles by: Ron Arthur
Article Source: www.iSnare.com
|
|
