If you want to have a look, just check my website.

The other important thing to note about this method is that it’s a little lossy. We spent months trying to figure out why when we first implemented page tagging, moving from our older log-based model, that we seemed to be missing visits. I spent a lot of time comparing our access logs to the raw page tag log from the supplier and in the end, for mysterious reasons (and believe me I know and accounted for all the reasons page tag results are *supposed* to be different from log results), page tagging was dropping about 5% of our visits. (Page tags got some visits logging didn’t, but lost a bunch logging did, and this is the net difference.) In the end, the supplier admitted “Yeah… That happens.”

For marketing purposes, that’s usually just fine. But I think it’s important to note that what you see from page tagging isn’t 100% of what’s actually happening, so don’t make logic mistakes based on it (e.g. “I’m sure no one from Company X has looked at our Web site, because it doesn’t show up in our page tag based analytics.”).

You might like to have a look at the opensource jsHub project:

http://jshub.org/

which is attempting to deal with core problem of making all this stuff easier to understand and visible to the Web Analyst.

A key requirement of any web analytics product is getting accurate data into it in the first place, and there are so many things that can be misunderstood by a developer during the technical implementation that result in the permanent loss of data.