So now that I am doing the universally loved University of British Columbia’s ‘Award in Web Analytics’ I can really start to look at our analytics use seriously.
We have both Webtrends and Google Analytics running – the first by design the second purely out of curiosity. While the former is very powerful but with a steepish ‘WTF’ curve, GA is a lof friendlier and it’s hard to think of a public sector site that wouldn’t benefit from the 12 hour ‘go to stats’ setup time and the radically less steep WTF curve.
But, first things first – what are you trying to track and why?
Of course, like all sites, we want to know the number of visits. Actually thats about as far as most people here want to know – lots and lots of visits, website must be doing Ok then. They don’t want to know about bounce rate and all the various other metrics that show a lack of engagement. Luckily I do because I want to ’sell’ them our fantastic story of decades of excellence.
My course talks about the 22 basic metrics. Some of those are pretty obvious, like bounce rate, but there are a couple that stand out as being new (for me at least):
- Scanning visitor rates – number of visitors who do visits of one minute of less
- Committed visit rates – number of visitors who engage for long periods with high page views
I have included these on my new long-term summary spreadsheet and have tracked out the last 6 months on it. And what did I find? What is interesting is the consistency on our site – nothing moves more than %5 and given the difficulty of being exact with analytics, that’s pretty much like saying nothing moves at all. So, is that a good thing or a bad thing?
You know what, I don’t know yet. I suspect it’s mostly good because of the user-testing I did on the prototype last year. But I am pretty sure that there are some issues, so I am going to set up some campaigns and figure out what is happening on the micro scale – and here I am going to set some actually useful KPIs that can show us if we are achieving our stated mission.
More on that tomorrow.
