Eric Meyer has recently posted a fairly cross broadside aimed at those who would utilise generic Browser Share stats when designing how to target a site’s design. Whilst I agree with 95% of what is said (i.e. that browser stats are useful only when taken within the context of the site they’re from and even then should not be relied on too heavily) I figure I should stick my oar in and add a couple of thoughts.
Firstly, I agree that using global statistics is a poor way to determine which browsers your site should support but I have to say that a little careful thought will allow you to derive useful information from the stats of similar sites. So whilst Eric is fundamentally correct in stating that his stats are of no use to any other site I say that they could be of use to someone planning a site with a similar demographic. Of course this information could be gleaned by applying a little common sense. For example:
A Website dedicated to web design and development
Such a site or blog is going to have a high share of Firefox users simply because people who are up on web technologies will more rapidly embrace such a browser.
A website discussing Mac software
Again, a little CS tells you to aim for Safari, FireFox and IE 5 for Mac first and all else after.
My only other concern is simply this - no-one should be using browser stats to determine which browsers should be able to view their site purely on the basis that all browsers should be able to view it. That’s the whole point behind XHTML, CSS, Semantics and all the other lovely things that have been keeping us web designers busy over the past 2 years. If you are picking and choosing between target browsers based on statistical analysis, even if your stats are the most accurate in the world (which they rarely are) then you have really missed the point. Keep your code valid first and ensure browser compatibility second. Make sure your design degrades well and not only will you not have to worry about browser share changing but your work will be effectively future proof.
Anyway, enough web talk for one day. Besides which it makes me more than a little nervous to comment on the writing of a guy who is responsible for 2 of the most thumbed books on my desk!
No Comments on "Browser Stats - Irrelevant or what?"