This should already tell you that your traffic is not representative.
The point being, maybe IE users don't actually make up more than 10% of traffic. Maybe the situations in which people use IE are such that they use the web less. He did say 10% of traffic, not 10% of unique users.
Not that I have evidence to back this up; however, the kind of analysis that gets done on market share in computing doesn't seem to be able to accurately reflect how much something is actually used.
I still don't believe this. I don't think that IE users are equal to or less than 10% of web traffic in the US or the world. That would be pretty huge if it was the case. Even a drop below 25% worldwide would be pretty huge news.
Quickly flicking through my analytics reports, I'm seeing IE percentages ranging from 84% (on an investment management site) down to 11% (a designer furniture site).
It wouldn't surprise me at all to see a youth oriented and/or computer gaming oriented site even further skewed away from IE than my furthest outlier.
Now, I too would be very surprised if that observation brings 45% of web users down to under 10% of web traffic, but I don't doubt that web traffic from IE is lower than 45%.
