I used to follow the browser metrics from a couple dozen popular blogs that have public Site Meter stats. After a while, I realized that the average of those blogs was trending right along with boingboing.net's browser breakdown so I stopped looking at Site Meter. Well, boingboing seems to have dropped their stats page so I went back to Site Meter to look over the stats for the couple dozen or so blogs I used to track and something looked really wrong.
The stats are clearly backwards. There's just no way that Firefox 2.x is in the 1-3% range and Firefox 1.x is in the 30-70% range.
I've heard about this kind of problem with other stats packages but it's surprising to see SiteMeter, the service used by so many top blogs, getting something like browser share so completely wrong. I sure hope none of those blogs made browser support (or any other) decisions based on this data.
Posted by: Robert Accettura | December 11, 2007 7:45 PM
You should try to get some total average statistics from Statcounter, that would be very interesting. They have one of the best statistics services around.
Posted by: David Naylor | December 12, 2007 1:13 AM
I find that www.webreference.com/stats/browser.html shows some good detail of the mix of Firefox versions in use. While they unfortunately don't seem to have archives of their data, if you check daily you can watch people transition from release to release (I love how Mozilla has gotten people to keep up with the web rather than the way Microsoft does things... IE 6 must die :)
Posted by: Limulus | December 13, 2007 10:12 AM
Who cares about the version, at least people use Firefox :-)
Posted by: Nathan | December 13, 2007 11:27 PM
I think this is a SiteMeter problem. I haven't heard of or seen this in other analytics products (Google Analytics, Omniture). Strange. A search on Google as well as their site shows no indication of this problem. I wonder if this has been a longer term problem, or if it's simply a bad code deployment over the past few days?