Yet another evil trick to work-around Google getting smarter

I just had this thought and typed it up (including a wee bit of Perl
code) before thinking it all the way through.

Please review and let me know what you think.



Problem Definition:

Google has become much smarter about what I will call
"footer links".

These are links which are placed in the header or footer of a web page, so
that they occur on every page on the web site.

It should be obvious to anyone who is looking that 100 identical links
from the same web site are NOT one hundred times better than one link from
that web site.

The question which naturally arises is: Does Google only look at the URL
or does Google look at the URL and the link text?

I had guessed that Google only looked at the URL.

Webby's report from this years SES conference leads strongly in the
opposite direction -- and that creates an opportunity.

Quoting Webby:

"It was also hinted that Google finds suspect a large number of links
pointing to the same site with identical link text. It looks (and 9/10
times is) a sign of manipulation of its pagerank algorithm though the
purchase of links and artificial linkage. Natural links don't always use
the same link text, Google is looking for natural linkage. In other words
alternate link text for your inbound links."

Reference: http://forums.seochat.com/t11609/s.html



Solution:

The solution is to vary the text of footer links. This isn't exactly
rocket science.

The most simple approach technically and the hardest to implement would be
to manually put different links on each page. That's a whole lot of work!

A slightly more complex technical solution that would be easier to
implement would be to create a server-side script that displayed random
outbound links. You could write each link variation by hand, and simply
have hundreds of variations.

Fortunately, the structire of the English language gives us a third
option. We could write a script which dynamically builds simple
sentences. This way, the link text could vary and we would still have
very little data entry to do. :-)

As a demonstration, I have such a script running at:
http://www.internet-search-engines-f...est-page.shtml

The script is amazingly simple. All it does is prepend one of the
following phrases onto the names of a number of web sites:

.. Learn more about
.. Discover
.. Check out
.. Visit
.. Learn about
.. Explore
.. Find out about

The script output is randomized every second, so don't hit <RELOAD> too
often.

An inproved variant of the script would display several links while
ensuring that no duplicate links were displayed. Or, for a small number
of outbound links, only the prepended phrases could be randomized.



Thoughts?

--
Will
Webmaster: http://www.internet-search-engines-faq.com

"Will Spencer" <will.spencer (AT) internet-search-engines-faq (DOT) com> wrote in
message
newsan.2004.06.09.19.16.26.77000 (AT) internet-search-engines-faq (DOT) com...

I just had this thought and typed it up (including a wee bit of Perl
code) before thinking it all the way through.

Please review and let me know what you think.
That's amazing - almost like SearchEngineCloaker ..??
Tony

I implemented this concept in a slightly different manner for my footer
links at http://www.fortliberty.org.

I made the links constants and am only randomizing the beginning of the
link text. I also made it work for every link in the footer.

(Footer is on the right menu)

The difficulty is... I don't really know how I am going to test the
results of this.

It's not like *anything* else is constant.

Will

On Fri, 11 Jun 2004 03:40:51 +0000, John A. wrote:

Sounds to me like they're concerned with intersite links, not
intrasite. If Google expects "organic" variation in link text within a
site and flags it as possible spam otherwise, they're nuts. They'll be
flagging every site on the web.

I am pretty certain that I have already seen this affect intrasite links.

I am far less certain of the effectiveness of my little work-around.

The obvious question being "How dumb could Google be?" ;-)

--
Will
Webmaster: http://www.internet-search-engines-faq.com

On Fri, 11 Jun 2004 05:16:15 GMT, Will Spencer
<will.spencer (AT) internet-search-engines-faq (DOT) com> wrote:

On Fri, 11 Jun 2004 03:40:51 +0000, John A. wrote:

Sounds to me like they're concerned with intersite links, not
intrasite. If Google expects "organic" variation in link text within a
site and flags it as possible spam otherwise, they're nuts. They'll be
flagging every site on the web.

I am pretty certain that I have already seen this affect intrasite links.

I am far less certain of the effectiveness of my little work-around.

The obvious question being "How dumb could Google be?" ;-)

The obvious answer being, "Extremely dumb!" Remember that
press-release they put out years back, the one including the advice
that site owners should demand an unconditional money-back guarantee
from any SEO compoany they hired?

BB

i cant see how this would work from googles perspective, what about sites
that have text links (sitemap) across the bottom of the pages, this would
devalue the entire site

as pages become increasingly css based, there is little code between primary
menu and the top of the page, same for the footer

mark

On Fri, 11 Jun 2004 09:19:53 +0100, mark | r wrote:

i cant see how this would work from googles perspective, what about sites
that have text links (sitemap) across the bottom of the pages, this would
devalue the entire site
as pages become increasingly css based, there is little code between primary
menu and the top of the page, same for the footer

That's an excellent example of why some dampening algorithm is absolutely
necessary.

Let's say that you have a web page called clay-aiken.shtml. Your link
text is "Clay Aiken".

There are quite a few other web pages on Clay Aiken out there on the
Internet. Heck, they may even be better than your Clay Aiken page.
However, your web site has 50,000 pages on it (Most with average PR,
ket's say 4.) You put a footer link in an .shtml include and suddenly you
have 50,000 links saying that your page is relevant for the phrase "Clay
Aiken".

You would get ranked pretty well.

The #1 result (not counting Froogle and Google local search results) for
the search term "Clay Aiken" only has 77 backlinks showing at Google.

"Clay Aiken" was the #1 search term on the Internet this week, on a list I
just checked. (Your Lists May Vary)

Suddenly, your page looks pretty relevant for the phrase "Clay Aiken"!

Unless Google dampens this algorithmically.

I always assumed that Google used some sort of URL-based dampening. That
is what made sense to me. Others forwarded IP-address based dampening
theories, but I never bought into those.

When Matt Cutts implied that the dampening was based upon link text, the
possibility of an easy work-around became apparent. And here it is.

Except instead of 6 variable sentences, I need to think of hundreds.
Time to hit M-W.com!

--
Will
Webmaster: http://www.internet-search-engines-faq.com

"Will Spencer" <will.spencer (AT) internet-search-engines-faq (DOT) com> wrote in
message
newsan.2004.06.11.13.45.32.882000 (AT) internet-search-engines-faq (DOT) com...

On Fri, 11 Jun 2004 09:19:53 +0100, mark | r wrote:

i cant see how this would work from googles perspective, what about
sites
that have text links (sitemap) across the bottom of the pages, this
would
devalue the entire site
as pages become increasingly css based, there is little code between
primary
menu and the top of the page, same for the footer

That's an excellent example of why some dampening algorithm is absolutely
necessary.

Let's say that you have a web page called clay-aiken.shtml. Your link
text is "Clay Aiken".

There are quite a few other web pages on Clay Aiken out there on the
Internet. Heck, they may even be better than your Clay Aiken page.
However, your web site has 50,000 pages on it (Most with average PR,
ket's say 4.) You put a footer link in an .shtml include and suddenly you
have 50,000 links saying that your page is relevant for the phrase "Clay
Aiken".

You would get ranked pretty well.

The #1 result (not counting Froogle and Google local search results) for
the search term "Clay Aiken" only has 77 backlinks showing at Google.

"Clay Aiken" was the #1 search term on the Internet this week, on a list I
just checked. (Your Lists May Vary)

Suddenly, your page looks pretty relevant for the phrase "Clay Aiken"!

Unless Google dampens this algorithmically.

I always assumed that Google used some sort of URL-based dampening. That
is what made sense to me. Others forwarded IP-address based dampening
theories, but I never bought into those.

When Matt Cutts implied that the dampening was based upon link text, the
possibility of an easy work-around became apparent. And here it is.

Except instead of 6 variable sentences, I need to think of hundreds.
Time to hit M-W.com!

--
Will
Webmaster: http://www.internet-search-engines-faq.com

Last week http://hosting.fast-trak.next had PR=0 despite having 92 inbound

PR4+ links, all from the another single specific site. Is this an example
of being marked down because you have paid to have 92 links from one site
all pointed at yours ?
Are there other examples like this ?
Best regards, Eric.

I would also like to know this as we pay to advertise on Scriptwiz for the
traffic but do not want to harm our Google PR. There are lots of other
places to advertise but only one Google.

"Eric Johnston" <eric.johnston (AT) blueyonder (DOT) co.uk> wrote

"Will Spencer" <will.spencer (AT) internet-search-engines-faq (DOT) com> wrote in
message
newsan.2004.06.11.13.45.32.882000 (AT) internet-search-engines-faq (DOT) com...
On Fri, 11 Jun 2004 09:19:53 +0100, mark | r wrote:

i cant see how this would work from googles perspective, what about
sites
that have text links (sitemap) across the bottom of the pages, this
would
devalue the entire site
as pages become increasingly css based, there is little code between
primary
menu and the top of the page, same for the footer

That's an excellent example of why some dampening algorithm is
absolutely
necessary.

Let's say that you have a web page called clay-aiken.shtml. Your link
text is "Clay Aiken".

There are quite a few other web pages on Clay Aiken out there on the
Internet. Heck, they may even be better than your Clay Aiken page.
However, your web site has 50,000 pages on it (Most with average PR,
ket's say 4.) You put a footer link in an .shtml include and suddenly
you
have 50,000 links saying that your page is relevant for the phrase "Clay
Aiken".

You would get ranked pretty well.

The #1 result (not counting Froogle and Google local search results) for
the search term "Clay Aiken" only has 77 backlinks showing at Google.

"Clay Aiken" was the #1 search term on the Internet this week, on a list
I
just checked. (Your Lists May Vary)

Suddenly, your page looks pretty relevant for the phrase "Clay Aiken"!

Unless Google dampens this algorithmically.

I always assumed that Google used some sort of URL-based dampening.
That
is what made sense to me. Others forwarded IP-address based dampening
theories, but I never bought into those.

When Matt Cutts implied that the dampening was based upon link text, the
possibility of an easy work-around became apparent. And here it is.

Except instead of 6 variable sentences, I need to think of hundreds.
Time to hit M-W.com!

--
Will
Webmaster: http://www.internet-search-engines-faq.com

Last week http://hosting.fast-trak.next had PR=0 despite having 92 inbound
PR4+ links, all from the another single specific site. Is this an example
of being marked down because you have paid to have 92 links from one site
all pointed at yours ?
Are there other examples like this ?
Best regards, Eric.