Wednesday, April 09, 2008

Browser Upgrade Campaign

As I move into the redesign of redesign, I'm using Dean Edwards IE7.js to get older versions of IE to behave almost like IE7. This means people using older versions of IE with scripting disabled will get a sub-par experience. I'll do my best to allow for graceful degradation, but things will probably look noticeably "off".

I'd like to offer a quick paragraph explaining why things look odd and take the opportunity to gently encourage them to upgrade to a modern browser, displayed only for those users who need to see it, like the old WASP Browser Upgrade Campaign. If that modern browser happens to be IE7, at least that's an improvement over previous versions of IE. I figure not providing a plug for IE7 in some way will result in an overall lower conversion rate. We're also dealing with a population that already use Microsoft products and may be paranoid about security (if you'll allow me the luxury of stereotyping people based on their browser settings). Sending them somewhere like may just get me dismissed as a “fanboy” or “cultist”.

I figure my options are:

  1. Link to the IE7 download page and be done with it.
  2. Link to both the IE7 download page and Browse Happy.
  3. Link to Save the Developers and let the various browsers fight it out there.

It may be useful to look at some recent browser info for our site in Google Analytics. IE users make up 89.57% of our visitors over the past month. Of those, 57.59% are using IE7. That leaves 42.41% using some other version of IE (including 17 visitors using IE8 somehow, but we can ignore numbers that small at the percentage level). That means that roughly 38% of our visitors are using older versions of IE. Getting accurate figures on people with scripting disabled because by definition these folks don't show up for javascript based analysis tools (like Google Analytics). Other sites report this info based on server log analysis but have skewed audiences and probably don't reflect a representative sample.

In the absence of any scientifically reliable data, let's make something up!

Ok, let's at least make something up in a range. We'll use 2% as our lower bound and 35% (the highest claim I've found so far) as our upper bound. That would mean somewhere between 0.8% and 13.3% of our visitors fall into the category I'm dealing with. Anything below 1% might not be worth troubling myself over, but anything that exceeds 10% certain is. I have now used fuzzy math to justify taking the time to write this post.

No comments: