If I visit ten websites, total the traffic and divide by ten, is that closer to mean or median? (Too early for me to math it up.)
Anyway, the mean and median didn’t diverge that much imo. For the “average” user I’d say site selection matters quite a bit (do you prefer gizmodo or Engadget, etc.?). But the numbers look pretty heavy either way.
There might not be an average page, but that doesn’t change the fact that many websites are horribly bloated for what amounts to some text and pretty pictures.
That’s the problem, not the imprecision of commentary.
There is a reason that the author linked to When Bill Gates Walks Into A Bar. It is an intro into exactly these two things – mean and median – and what they can tell you about the data. The first two sentences:
The mean of a data set is obtained by summing the data elements in the data set and divided by the total number of data elements (what most people think of as average). The median of a data set is the middle value of the data elements when the data elements are sorted.
The average page.. is a home page?
I might be wrong here, but it looks like they’re only pulling down homepages (i.e. GET /). I wouldn’t be surprised if those were the largest pages on a given domain. Besides, if I’m just opening up links I see on lobsters/reddit/hn/etc, I’m never even going to see most homepages; I click straight through to the article.
Also, I don’t think there’s much to learn knowing a site like www.thebestshemalevideos.com is a page weight outlier due to images (or videos, for that matter). It’s like telling me YouTube loads lots of videos, or imgur loads images. I guess when I read “average page”, I think “average blog post” or “average news article”.