diberry - 3:52 pm on Oct 12, 2012 (gmt 0)
Tedster, the Adjusted Bounce Rate sounds like an even better metric than Exit Rate for determining whether a user got what they wanted. Clicky (what I use for stats) has this built in - the only count as "bounces" users who both hit the back button AND stay less than 30 seconds.
Could Google measure the lag between someone clicking a result and hitting the back button to Google? If so, that would give them some sense of adjusted bounce rate to feed into the algos without using Analytics data. Would they be able to track things like that I click a result on Google, look at the webpage for 10 seconds and then exit the page by clicking a non-Google (say, Federated Media) ad? If so, this could be a pretty powerful indicator of user satisfaction.
If not, it would probably still work well on large sites, where lots and lots of incomplete data can still show patterns. And didn't a lot of us note that Panda seemed to start with bigger sites? So if claarky's right that Panda is mainly about one or two metrics, this one might be a biggie.
Checking this theory against my own sites: I'm pretty sure my Penguinized site just got Pandalyzed on the Sep 27 update, even though it's adjusted bounce rate is pretty decent. But the pages that have been hardest hit are the ones that have a number of links to other websites. Users tend to visit these pages, click a link, come back to my page, click another link and often end up subscribing/bookmarking me... it may be that Google isn't recognizing that as site engagement because it can't track everything I can.)