The Sekhet Semantic SEO Experiment

[morestar]

Hello ladies and gentlemen of the SEO and Web World. I'd like to share a little experiment "Professor" Webnauts and I did and check out the results related to CTAGs and the meta keywords tag.

We created a subdomain on one of my domains called SekhetFeist. On the main page, we created five links to sub pages which were almost duplicates of each other but with subtle differences.

Simply search for SekhetFeist and you will see the results - the pages that Google has displayed.

Here are the search results.



The experiment was the following: we created a home page plus 5 more pages that were 99% identical. The only difference was that we added only numbers to the sub page names, to the title of the pages, made distinctions to the H1 tag of the pages and to the targeted key phrases in the document.

We tried to avoid the pages being rejected by Google by making each of the previously said elements distinct.

What happend was that Google picked up the home page and the page that was the third link on the home page (not the first).

That page that Google picked up was the page that the exact keywords were in the CTAGs and only visible to machines.

Please check out and tell us what you think were the factors that made Google index and display that page over the other 4 pages (not including the index page obviously).

[williamc]

The fact that you said you changed the titles and h1 tags already shows the test is worthless as you have just changed enough that google simply chose the page(s) it found more relevant. It does not show that ctags played any part at all in the result. When you have that many variable factors in motion there is no way to gauge any results as being anything but random. John should already know this.

[kgun]

We created a subdomain on one of my domains called SekhetFeist. On the main page, we created five links to sub pages which were almost duplicates of each other but with subtle differences.

The experiment is setup with the kind support of Professor Webnauts.

I assume that is a joke, so be careful. You can mislead some visitors.

[Webnauts]

William are you claiming here that that only because the difference on that page was the number 05, and not 03, 04, 06, 07?

And what about the position of the link within the content?

And who is talking about the keyword itself? The test is about indexing and relevancy.

And as you have noticed, Google picked up the page which CTags where only serving machines and not users.

Anyway I will now ping on Skype Andy Beard and see what here thinks too.

I will let you know what we discussed after all presented their opinions here.

[morestar]

The fact that you said you changed the titles and h1 tags already shows the test is worthless as you have just changed enough that google simply chose the page(s) it found more relevant. It does not show that ctags played any part at all in the result. When you have that many variable factors in motion there is no way to gauge any results as being anything but random. John should already know this.

Please name out all the differing variables William, their effects and why the third link became the most relevant if you can.

[crankydave]

I've only glanced, but the different titles alone skew relevancy, as well as the URL's which are really very different.

[Webnauts]

I assume that is a joke, so be careful. You can mislead some visitors.

You are correct. But since you felt offended or misled, I asked Morestar to move that. And I was expecting a more constructive post from you, since you are the most popular semantics guy on WPW. But maybe you are planning to share with us some of your experience.

[williamc]

William are you claiming here that that only because the difference on that page was the number 05, and not 03, 04, 06, 07?

And what about the position of the link within the content?

And who is talking about the keyword itself? The test is about indexing and relevancy.

And as you have noticed, Google picked up the page which CTags where only serving machines and not users.

Anyway I will now ping on Skype Andy Beard and see what here thinks too.

I will let you know what we discussed after all presented their opinions here.

John, you were talking about a particular keyword SekhetFeist which is what the results are based upon. You varied the title, the h1, and the content of each page. I am assuming that the links are also on pages which may pass more or less ranking juice which also plays a part in relevancy with the aforementioned factors. The fact that google picked the page with the ctag on it still simply means google found that page more relevant. It does not mean that the ctag caused any relevancy score to be attached in any way. There are too many other relevancy factors that varied among the pages for the test to be close to conclusive or even remotely conclusive IMO.

I too would like to hear Kjells thoughts on the test itself.

[Webnauts]

I've only glanced, but the different titles alone skew relevancy, as well as the URL's which are really very different.

Homepage URL: http://www.sekhet.rdfa.ca/

Subpage 1: ./sekhet-03.html

Subpage 2: ./sekhet-04.html

Subpage 3: ./sekhet-05.html

Subpage 4: ./sekhet-06.html

Subpage 5: ./sekhet-07.html

The page indexed was: ./sekhet-05.html

Am I missing something? Can you please explain what you mean?

[crankydave]

John... the only page I see in the results for a search is the friendfeed