JD on EP

Many sites turn away a certain percentage of users by using Flash, so it's no surprise that they might use another technology to do the same.

Personally, I don't think I've ever seen a reported percentange lower than 3-5% for users with no JavaScript. That's purely from memory, but I think I'd actually be surprised to see only 1-2%.

for what it's worth (since it's only one site: my own), awstats reports that only 0.5% (one half of one percent) of my visitors have javascript disabled. awstats doesn't seem to report on javascript versions, though.

the lion's share of my traffic comes from myspace, so you might be able to extrapolate that cyberpet-loving myspace users have javascript enabled 99.5% of the time ;)

What I suspect is happening is that stats packages are counting 'fake' or unknown userAgents that ignore javascript. By 'fake' I mean, bots that identify themselves as real browsers or are not as yet recognized as spiders and such and thus subtracted from visitor numbers.

One clue might be to compare the number, if given, of unknown browsers and OSen to the number of users reportedly surfing sans-javascript. A quick look at my blog's AWStats reports for the last two months show a suggestively close relationship between the numbers:

Dec 8.1% disabled - 3.8% unknown browser - 6.3% unknown OS

Jan 3.7% disabled - 3.7% unknown browser - 2.9% unknown OS

Over the holiday break I implemented some blocking measures to the more aggressive link/trackback farmers, which may help explain the difference between Dec and Jan.

A client's high-traffic site also uses AWStats (as a backup for SiteCatalyst) and before the site became Akamaized, consistently showed 0% disabled(!) with below 1% numbers for unknown OS and browser. (now shows 20% unknown for both browser and OS, which prob is Akamai; still 0% Javascript incompatible, though). The site offers nothing of use to link farmers and so may be avoiding unidentified bot traffic.

Just a guess.

Occurs to me that another variable to consider might be RSS feed presence. Aggregators and consumers may not all be recognized UA and OS, and if RSS feeds are counted as visitors, this could skew results. I also upgraded AWStats over the break (includes new RSS UA in the bots db) and this could further factor into the Dec/Jan differential.