I feel some empathy for the folks behind SiteMeter. As has been widely reported, they did an update to their service, and said update broke fairly catastrophically in IE. The reason I have some empathy is that I've executed similarly botched updates to this blog site - the day I first used "future posting" I was on a plane when my update crashed the app server :)

Still, their problem sounds a lot like the one I had: a failure to test adequately. The big question is, how did something like this slip through? Having a Javascript flaw that breaks in IE seems like the sort of thing that should have been caught.

This story points out the dangers of not doing that testing, too - lots and lots of sites yanked SiteMeter, because the flaw made them inaccessible to a large body of readers. I expect that a not insignificant number of those sites won't get around to putting it back up, either due to irritation or inertia (there's almost always something better to do than go back and check some third party software).

Updates are very easy to roll out on a web app - but with that ease comes the ability to make an absolutely horrid impression.

Technorati Tags: PR, management