webby2001 - 11:00 am on Aug 29, 2004 (gmt 0)

I was actually specifically refering to major changes in traffic on Wednesday 25th in particular the travel booking theme sites were hit with a very big filter stick.
I researched possible reason because it wasnt across the board on all terms and compered pages which had been hit with those that wernt. My post on page three outlined my findings. That was my hypothesis which btw requires one amendment and that is that you can have too many internal links with the same link text. I noted one page had over 2000 internal links and also dropped. Those pages with around 80 internal backwar dlinks actually rose. This was fact not hypothesis. I also believe footer links are also not weighted anywhere near what they used to be so the tests I am doing are increasing the internal backward links to a page (destination) that dropped 20 places from 1 backward link to 50 and I am embedding the link in a paragraph rather than link ¦ link2 etc.
It is only through these kind of tests that can provide evidence of a hypothesis/ theroy (call it what you will). The big problem with testing is by the time the new links are weighted and a ranking change noted Google has moved on and tweaked their algo some more. However I believe the drop in serps on Wednesday is down to the fact some internal pages simply didnt have enough internal links with the right link text. I'm also adding an external backward link to the page I am using for testing as it is most likely not just an internal linkage problem but a pr prob as well.

I stress my concern is really only for those sites that dropped on Wednesday not before, as on Wednesday something changed that has affected many travel sites and from what I am hearing other competitive keywords as well.
I'll keep you informed of the tests.

I agree however that it is very easy to come up with a hypothesis. Too often hypothesis posts on this forum remain that because no one actually bothers to put them to the test and publish results. It all ends up in speculation which helps no one.
Because urls, specific examples aren't allowed here you cannot really show/prove your hypothesis anyway.

Alan