Web site statistics

We use Urchin 5.6 for our web site statistics. This works better for us than Google Analytics for a number of reasons, which is why it is somewhat irksome that a newer version of the Urchin software hasn't come out. I hear reports that Google, who bought Urchin a while back, is working on a new software based version of their statistics software but I haven't heard much.

I hope it comes out.

Google Analytics is unabashedly designed around advertising-related statistics. No surprise, since that's where the money is to be made. And for that, it works great.

What it doesn't do is tell me a few, very key things:
  • How many total bytes did this web-server serve in this time period? Network monitoring will give me the whole server, but this will give me the specific web-server itself.
  • What are the top 10 hit files?
  • What are the top 10 files generating traffic?
These are things I'm concerned about as a webmaster. This is stuff you can only get by parsing web-server logs.

Of the top 10 hit files on student MyWeb, 6 of them would be revealed with Google Analytics.
Of the top 10 files on student MyWeb generating traffic, which consists of 81% of total data transfer, not a single one would be revealed by Google Analytics.

The top file last week for student MyWeb is an MP3 file generating 31% of total data transfer traffic. After digging into the actual log-files to see what is referring that traffic, I learned that there is a new flash-based music search service out there. While Analytics would track the loading of the flash file itself on those not-WWU servers, it won't track the transfer from my server. That Flash prog definitely doesn't execute custom Javascript.

Google Analytics and server-log parsing programs serve different market segments. Google, understandably, is only interested in the ad-driven segment. I just wish they'd get off their butts and release a new version of the log-parsing Urchin software.