Google PageRank Algorithm -
carefully explained and what you can do
with it - written by top SEO experts
PageRank
or PR(A) can be calculated using a simple iterative
algorithm, and corresponds to the principal eigenvector
of the normalized link matrix of the web.
What that means to us is that we can just go ahead and
calculate a page’s PR without knowing the final value of
the PR of the other pages. That seems strange but,
basically, each time we run the calculation we’re
getting a closer estimate of the final value. So all we
need to do is remember the each value we calculate and
repeat the calculations lots of times until the numbers
stop changing much.
Lets take
the simplest example network: two pages, each pointing
to the other:

Each page
has one outgoing link (the outgoing count is 1, i.e. C(A)
= 1 and C(B) = 1).
Anybody who monitors their
rankings with the same vigor that we in the SEO
community do will have noticed some fairly dramatic
shifts in the algorithm starting last Thursday (July
5th) and continuing through the weekend. Many sites are
rocketing into the top 10 which, of course, means that
many sites are being dropped at the same time. We were
fortunate not to have any clients on the losing end of
that equation however we have called and emailed the
clients who saw sudden jumps into the top positions to
warn them that further adjustments are coming. After a
weekend of analysis there are some curiosities in the
results that simply require further tweaks in the
ranking system.
This update seems to have revolved around three main
areas: domain age, backlinks and PageRank.
Domain Age
It appears that Google is presently giving a lot of
weight to the age of a domain and, in this SEO's
opinion, disproportionately so. While the age of a
domain can definitely be used as a factor in determining
how solid a company or site is, there are many newer
sites that provide some great information and innovative
ideas. Unfortunately a lot of these sites got spanked in
the last update.
On this tangent I have to say that Google's use of
domain age as a whole is a good filter, allowing them to
"sandbox" sites on day one to insure that they aren't
just being launched to rank quickly for terms. Recalling
back to the "wild west days" of SEO when ranking a site
was a matter of cramming keywords into content and using
questionable methods to generate links quickly I can
honestly say that adding in this delay was an excellent
step that insured that the benefits of pumping out
domains became extremely limited. So I approve of domain
age being used to value a site - to a point.
After a period of time (let's call it a year shall we)
the age should and generally has only had a very small
influence on a site's ranking with the myriad of other
factors overshadowing the site's whois data. This
appears to have changed in the recent update with age
holding a disproportionate weight. In a number of
instances this has resulted in older, less qualified
domains to rank higher than newer sites of higher
quality.
This change in the ranking algorithm will most certainly
be adjusted as Google works to maximize the searchers
experience. We'll get into the "when" question below.
Backlinks
The way that backlinks are being calculated and valued
has seen some adjustments in the latest update as well.
The way this has been done takes me back a couple years
to the more easily gamed Google of old. This statement
alone reinforces the fact that adjustments are
necessary.
The way backlinks are being valued appears to have lost
some grasp on relevancy and placed more importance on
sheer numbers. Sites with large, unfocused reciprocal
link directories are outranking sites with fewer but
more relevant link. Non-reciprocal links lost the
"advantages" that they held over reciprocal links until
recently.
Essentially the environment is currently such that
Google has made itself more easily gamed than it was a
week ago. In the current environment, building a
reasonable sized site with a large recip link directory
(even unfocused) should be enough to get you ranking.
For obvious reasons this cannot (and should not) stand
indefinitely.
PageRank
On the positive side of the equation, PageRank appears
to have lost some of it's importance including the
importance of PageRank as it pertains to the value of a
backlinks. In my opinion this is a very positive step on
Google's part and shows a solid understanding of the
fact that PageRank means little in terms of a site's
importance. That said, while PageRank is a less than
perfect calculation subject to much abuse and
manipulation from those pesky people in the SEO
community it did serve a purpose and while it needed to
be replaced it doesn't appear to have been replaced with
anything of substantial value.
A fairly common belief has been that PageRank would be
or is being replaced by TrustRank and Google would not
give us a green bar to gague a site's trust on (good
call Google). With this in mind one of two things has
happened; either Google has decided the TrustRank is
irrelevant and so is PageRank and decided to scrap both
(unlikely) or they have shifted the weight from PageRank
to TrustRank to some degree and are just now sorting out
the issues with their TrustRank calculations (more
likely). Issues that may have existed with TrustRank may
not have been clear due to it's weight in the overall
algorithm and with this shift reducing the importance of
PageRank the issues that face the TrustRank calculations
may well be becoming more evident
In truth, the question is neither here nor there (as
important a question as it may be). We will cover why
this is in the ...
Conclusion
So what does all of this mean? First, it means that this
Thursday or Friday we can expect yet another update to
correct some of the issues we've seen rise out of the
most current round. This shouldn't surprise anyone too
much, we've been seeing regular updates out of Google
quite a bit over the past few months.
But what does this mean regarding the aging of domains?
While I truly feel that an aging delay or "sandbox" is a
solid filter on Google's part - it needs to have a
maximum duration. A site from 2000 is not, by default,
more relevant than a site from 2004. After a year-or-so
the trust of a domain should hold steady or at most,
hold a very slight weight. This is an area we are very
likely to see changes in the next update.
As far as backlinks go, we'll see changes in the way
they are calculated unless Google is looking to revert
back to the issues they had in 2003. Lower PageRank,
high relevancy links will once again surpass high
quantity, less relevant links. Google is getting
extremely good and determining relevancy and so I assume
the current algorithm issues has more to do with the
weight assigned to different factors than an inability
to properly calculate a links relevancy.
And in regards to PageRank, Google will likely shift
back slightly to what worked and give more importance to
PageRank, at least while they figure out what went awry
here.
In short, I would expect that with an update late this
week or over the weekend we're going to see a shift back
to last week's results (or something very close to it)
after which they'll work on the issues they've
experienced and launch a new (hopefully improved)
algorithm shift the following weekend. And so, if you've
enjoyed a sudden jump from page 6 to top 3, don't pop
the cork on the champaign too quickly and if you've
noticed some drops, don't panic. More adjustments to
this algorithm are necessary and, if you've used solid
SEO practices and been consistent and varied in your
link building tactics - keep at it and your rankings
will return. Once
you know how Google measures or evaluates your links
then you will be in a better position to conduct a more
effective link campaign. Before I explain the current
pattern of link evaluation let's take a stroll down
memory lane to see how far things have come.
In The Beginning A link Was The Equivalent Of A Thumbs
Up Vote!
The original premise of the internet was to have a bunch
of related resources (websites) linking to one another
in a natural pattern of progression. Thus someone on a
website focusing on gorillas for example would then
naturally migrate from that site to one that listed
gorilla safaris. Thus quite soon links became (and still
are) the standard with which to measure a website's
relevance.
The thinking behind this was that the more popular a
website was the greater the likelihood that people would
link to it; and since people would only tend to link to
sites that offered useful information then it naturally
followed a large number of links signified quality...at
least that was the idea.
The Birth of Link Spam!
Soon enough wily webmasters realized that they could
manipulate the linking concept to their own advantage
and make a handsome profit while they were at it.
Instead of waiting for people to link to their new sites
(a process that could take forever) why not form link
directories whereby one, for a certain fee, could amass
a large number of links in next to no time at all!
Booming Link market
Getting a new website indexed by Google used to be a
task of Herculean proportions. Quite often new websites
would languish in obscurity for ages, but those in the
know realized that the process could be accelerated. By
linking a new website to an established website or
webpage with a high page rank (typically page rank (PR)
7 and above) that new website could be indexed by Google
within days or a matter of hours depending on the PR
value of the webpage on which the link was located.
However getting a link from a high PR webpage did not
come cheap and it was not uncommon for webmasters to
fork over $800 for a one month link on a PR8 webpage!
Debut of The Link Farm
The link market trade flourished for a while, making a
mockery of Google's system of assessing the importance
of any given website or webpage. However, a lot of
webmasters in a hurry to see their websites scramble up
the search engine indexes could not afford the hefty
price commanded by a link from a high PR webpage and
thus link farms came into being.
Link farms were the poor man's solution to the hefty
priced, high value links typical of high PR webpages.
The premise of the link farm was as follows: instead of
paying an outrageous sum of money for one link from a
high page rank website why not pay a small fee for
thousands of low PR links? In essence the link farms
abided along the principle of the sum amounting to
greater than its individual parts!
The Obsession With The PR Band!
Google thoughtful as ever, made it possible for anyone
to determine the relative importance of a website at a
glance. This they did by the introduction of the PR
band, a small line atop one's browser fittingly entitled
with the word "PageRank." The pagerank band scaled from
a low of zero (PR0) to a high of ten (PR10). The greater
the value that Google attributed to a given webpage the
higher its designated page rank. In little to no time at
that little green slash (PR band) became the highly
effective unofficial publicist for the flourishing link
market trade.
Webmasters obsessed over the PR band even though in
truth the page rank value of a webpage plays little to
no part in determining the position any webpage will
attain on the SERPs (search engine index results pages).
However savvy marketers used the PR obsession of the day
to great advantage...using it as a tool to establish
credibility and ultimately sell their wares to the less
savvy.
However somewhat on par with the abuse of the linking
system, the page rank concept soon became equally
defunct. In no time at all the Google index was awash
with spam sites topping its front page! Something had to
give; and it did!
Fast Forward To Today...New Link Evaluation Parameters!
Oh how times have changed!
The search engine algorithms have gotten so much smarter
that even blackhat SEO (shady search engine optimization
techniques) these days is more trouble than its worth.
With the continued abuse and manipulation of the system
Google furiously tackled the issue of how to maintain
website relevance and quality in its index in an
increasingly spammy world.
Since its embracement by the public, the internet has
seen its fair share of SEO manipulation from blackhat
techniques such as multiple-blog-creation software
following the debut of blogs (principle behind this
system was that one could create thousands of instant
one-way backlinks through the instant generation of
thousands of blogs) to whitehat techniques such as the
widespread dissemination of articles through article
distribution software or services.
Some of these techniques still have value today while
others have been effectively nullified such as the mass
generation of backlinks through the creation of instant
blogs.
How Google Currently Evaluates Links.
It used to be that if the PR band of your
website/webpage was gray then this signified that said
webpage had been banned. That is no longer the case. All
new websites/webpages start off with a grayed out PR
band. This band will remain that color depending on a
number of factors before it turns white (indicating that
the probationary period is over) and eventually
transforms to green with establishment of page rank
above zero.
Link Aging Filter / Link Probation Period
To counter the widespread habit of acquiring links for a
short-term period, say a month, so as to get a website
indexed or attain page rank quickly, Google now subjects
every link to a probationary period in which time the
link is identified by Google but not accorded any value.
Although such a link is recognized by Google, typically
it will not be registered as a link associated with the
website to which the link points for a subjective
period; for that to happen the link needs to mature, and
the rate at which a link matures is dependent on a
number of factors.
Factors That Hasten Link Maturity Or Reduce Link
Probationary Period
1. Keywords: The keyword incorporated in any given link
is going to determine how long that link will be under
probation. The more competitive the keyword the longer
will be the link's probationary period. In reality this
link probationary system has superseded the Google
Sandbox, which was/is Google's technique of ensuring
that super-optimized new sites do not zoom
uncontrollably up the SERPs, quickly overshadowing long
established websites. With respect to the Google Sandbox
concept, if a new website was/is targeting a very
competitive market already saturated with millions of
sites, then that new website would spend a longer period
cooling its heels in the Google Sandbox.
2. Volume of Traffic Across A Link: The more trafficked
a link is, the quicker it will attain full SEO-value
recognition by Google. This is a relatively new system
in play and basically what it means is that a heavily
trafficked low PR link will bestow far more SEO
influence to the webpage it points to than a little
trafficked high page rank link. In essence Google has
pretty much nullified the system whereby webmasters
could purchase links from high PR webpages in hopes of
improving the SEO status of their own site. The
application of this system explains why some newer
webpages attain PR before older more established
webpages on the very same website.
3. Links From Topically Related Sites: A links that
comes from a website or webpage that has a
comprehensively topical relationship to the destination
webpage will have greater value than a link that does
not. For example, say your website is about internet
marketing, getting a link from another website that
focuses on online marketing software would definitely
fall under the category of topically related sites.
4. Utilizing Different Keywords In Links: Having the
same keyword incorporated within all the links pointing
to your site will tend to set off red flag alerts, which
means that those links are not going to attain full SEO-value
recognition by Google for a longer duration. Another
link strategy you should employ is to have your links
pointing to some of your inner pages as opposed to all
your links pointing to your home page. Having links
spread across multiple webpages registers as a more
natural link strategy to the Google algorithm and the
premise of this algorithm tweak was to nullify the
system where webmasters would purchase links from link
directories.
Remember anything that appears suspicious to the
Googlebot is ultimately going to decrease the
effectiveness of your SEO campaign. There're many more
factors that play a role in determining how high and how
fast your website will make it up the SERPs (Google of
course doesn't spill all the beans) but if you conduct
your link campaign with these few tips in mind you won't
go far wrong.
|