A thoughtful post on web analytics, re-emphasising the different needs of museums compared to e-commerce sites with regard to making sense of their stats, and the confusion and tail-chasing that result. We aren't especially sophisticated users of stats, although I'm the worst culprit on the whole: I don't get WebTrends and leave it to our web content manager to set up the profiles and deal with all that. However I do keep a fairly close eye on our Google Analytics data, although I'm not exactly a power user. We also don't have the script on quite all our pages, which means that I'm missing data on some of our most popular but older parts of the sites.
With GA it's a lot easier that with WebTrends to get to lots of revealing information about the sourc of our traffic (geographically and referral-wise), or stuff like entry and exit pages - you can hop from one statistic to another quite intuitively, and that's encouraging me to delve a bit deeper. But I'm not particularly methodical, it's more if I see something unusual (like spikes in visits from the BBC or games websites) that I dig into it.
The other form of web-related stats we gather (but which, again, I pay probably too little attention to) is on the exit surveys from the museum or from temporary exhibitions. These typically ask whether users discovered the museum on the web or used our sites for visit information or anything else. We've tried web surveys to little effect, and they're too self-selecting anyway, but the exit surveys are pretty informative.
There's lots more to say but I'm clearing out a backlog of drafts and this will have to do for now...