bakedjake - 11:30 pm on Sep 29, 2004 (gmt 0)

Dave, I think you're half right. I propose the idea that the "Google Lag" is not an active evaluation of the site in question, as in "If site posseses attribute X, accept it, else do not".

related sites queries show that Google often groups sites together that have links from multiple, unrelated sites pointing to all of them.

Example: I run a site about widgets. A webmaster halfway around the world runs a site about naked cows.

We each go out and get links from well known web directories. 2000 of them, let's say. And pretend they're one page directories with a bunch of links on them.

Visualize: directoryApage points to widgets and naked cows. directoryBpage points to widgets and naked cows. directoryCpage points to widgets and naked cows. Wash, rinse, repeat 2000 times.

Suddenly, Google thinks our sites are related, just because we appear on so many pages together.

What if the "Google Lag" is not evaluating attributes of individual sites, rather, the evaluation is happening on "groups" of multiple links?

Most SEOers here are reporting problems with the "Google Lag". Most SEOers here are getting links from the same sites/pages, using the same linkdev techniques.

Why would they do this? If they did, think about how easy it might be to categorize sites, without the need for a human edited directory.