Sunday, September 14, 2008

A shipwreck at sea can ruin your whole day

Way, way down at the bottom of this blog is a little green and white icon that looks something like the Brooklyn Bridge (except that is not what it is meant to look like). That’s an indicator that I use a free service, called SiteMeter, to keep track of how many people visit the blog.

The free version of SiteMeter is what you might call web analytics lite. It isn’t robust enough to service the needs of a large ecommerce company, but it gives me interesting information and, more important, yet another way to waste time during the day, checking on my site stats.

SiteMeter makes their money from paid subscribers, of course; and it gives them more data than it gives me. That’s fine, and more power to them.

Well, for the last several months, SiteMeter has been promising a spiffy new user interface, and this weekend was when they decided they were ready to roll it out. Unfortunately, although the company’s management may have been ready to roll out the new software, the new software wasn’t ready to be rolled out. It has, um, performance issues, sort of the software equivalent of erectile dysfunction. It can’t stay up, in other words.

So now SiteMeter is rolling the interface back to the old software, which worked fine. I have no inside information on whether they’re going to try to repair the new software or abandon it and start over. My opinion, for what it’s worth, is that this was a really, really bad idea. Instead of continuing to serve data as HTML and just spiffing up the presentation, they got all artsy-fartsy and implemented the interface in Flash. This led to three completely predictable problems:
  1. Graphic artists got hold of the process and introduced all kinds of pretty but completely illegible presentation formats.
  2. Because the presentation is controlled by Flash, the browser cannot reach in and make the fonts larger (more legible)—and the design does not include Flash buttons to do the same thing.
  3. Accessing the data and refreshing the display take way too long.
What makes software companies do this?
Why do companies invest months of expensive development time in a bad idea and then roll it out when they should already know it cannot possibly work? It seems to me that this is consistent with the magical thinking school of management: Logic be damned! If I keep saying something will happen, and I say it enthusiastically enough, by golly it will happen. “Magic words of poof, poof, piffles, make me just as small as Sniffles!”

Clear thinking matters.

No comments: