Re: Is A2K Losing Some of its Luster?
nimh wrote:
Yeah, just thought it was odd that the top sites wouldnt grow along with the total volume - you'd think that the new internet users use the top sites in roughly equal proportion to the old users, and that they would thus grow along in comparable measure with total internet traffic.. (unless like I was speculating, its a language thing, but the top Chinese sites werent showing a systematic increase on Alexa either..)
Alexa's calculations are all ratios. For example, the reach ratio is a number per million internet users.
So yahoo, for example with about 28% reach of internet users would have an alexa reach rank of about 280,000.
This is why I said that you won't find any evidence of a growing internet on Alexa. Think of it as a market share measurement and for sites to go up in the ranking they not only have to be growing in traffic but growing in the total share of internet traffic.
So a site's traffic can actually grow while the alexa rank diminishes even if the alexa statistics are accurate.
Quote:
Craven de Kere wrote:As to the internet growing, I don't get why you are trying to measure that through the alexa stats. Nothing they have in their free statistics will show you that growth.
Silly question perhaps, but is there actually any kind of place where they'd collect some sort of stat on err, total internet volume/site visits? Even just extrapolating from sample data or something? Or is that simply technologically impossible because of the sheer scope?
It's about as possible as much other statistical calculations you reference in political research.
I don't know of a single easy resource to get what you want. When I need data I've used agencies like Nielsen and comScore but I've never needed to know that particular statistic.
Extrapolation of some kind would be needed to get a total user number, as the internet's protocols and architecture really only can count computers visits and such (so you could, in theory, find ratios of human traffic vs machine traffic and ratios of unique humans to traffic and use the total).
Thing is, most such extrapolation from sample data would need to already know the total. Alexa and Neilsen use sample data and get the total externally as far as I know.
The only place I know that they may be getting it from would be the itu (
http://www.itu.int ). I see media outlets and sites like this
http://www.internetworldstats.com/stats.htm quote statistics all the time, but I haven't run into any such citation that tells exactly how they reached their numbers.
[Reyn runs away, as he's not on the list] 
