(2)
Google does track clicks, but do they use this data to place page A over page B? Proof that they track links, well, I did a search on rustybrick and then moused over the first result. I saw the following target URL, http://www.google.com/url?sa=U&start=1&q=http://www.rustybrick.com/&e=10012 - see for yourself.
So, does this click tracking affect a site's rankings? Many people I meet that are "computer savvy" ask me if Google determines page rank by user clicks. Of course I laugh, but its a solid question. Direct Hit was an engine that based a page's rankings on click data, but that didn't last too long.
In all honesty, does the click data affect relevancy? You have to think so, on some level. Google is probably storing this data in order to do UI testing, relevancy testing and other kinds of testing. So on some small level, one can argue it affects relevancy and possibly a page's rank. But it is known, that clicking on your page in Google over and over again will not shoot your site up to the number one slot.
Forum discussion at WebmasterWorld.
Comments:
Jon Payne
02/23/2005 12:39 pm
If Google is user CTR on the SERPs as a factor I would have to believe its extremely minor. Otherwise for virtually any search term you'd just have the strong get stronger, as the top 10 will undoubtedbly get more clicks than those on the second and third pages, thus negating updates to Google's algorithm that otherwise might push those second page sites up. This whole idea seems counterproductive. For CTR the case could be made that this would not matter, as those on page 3 get less views and thus even though dramatically less clicks their CTR is not necessarily negated. However, for all sites in the top 10 this issue still exists. Site 10 is always going to have a lower clickthrough than site 1 in any large study... If anything I think they might use it to analyze what they display as the description, and how that impacts CTR.
SP
10/11/2009 03:47 pm
Jon, I don't think Google is stupid; if CTR plays a role Google is likely measuring deviations of CTR from what would be expected, not raw CTR.