I've been using Google Analytics for many (!) years as a standard addition to pretty much any website I've been involved with. It seems to be widely accepted as the de-facto standard in free website analytics and as such I've generally taken the information it provides to be pretty accurate.
But here's the thing; what if it's not?
I can explain away many of the differences between what I'm now recording and what Google analytics provides, typically there's a 5-20% difference, which isn't massive in the grande scheme of things as typically it's trends we're interested in rather than precise numbers. However, what does worry me is the difference in average session length.
Google seems to think it's currently 1:19 per session, whereas I'm measuring a figure of 11:17 per session, which is a factor of 9 out, which is a pretty major disagreement. Assuming for a second that my number is the accurate one (which it is!) there are implications to be drawn which have some real potential. It means the site is probably being left open in a tab, by a notable number of clients.
Two things spring to mind;
- The page they are on is sufficiently important to them to leave it open over an extended period of time - so I'd quite like to know what that page is.
- While the page is open, there is a connection (in this instance) from the server to the browser, so there is the potential to interact with someone who has a specific interest in the site.
It would appear Google are missing a trick! Not to mention, I'm sat here comparing analytics and Google are at least 15 minutes behind, that's not going to fly for the next generation of real-time interactive websites.
Anyway, we're not far from an initial beta release now, everything literally in real-time, and hopefully completely accurate with all bots and search engines excluded .. :)