> Something must be very wrong with their methodology. I don't believe the total installed base is so volatile that the Mac share jumps from 7% to 9.57% then do

> Something must be very wrong with their methodology. I don't believe the total installed base is so volatile that the Mac share jumps from 7% to 9.57% then down to 6% within a single year. Agreed. And it's by no means the first such example. However, the numbers are based on unique visitors to a range of websites. The variation might reflect a wide variation in web accesses from different operating systems (though I doubt it), but it doesn't directly measure the installed base.