Like many other websites Europeana collects statistics on the number of visitors our portal receives, which pages they look at, links they click and from which sites they arrive to our site. For this we have basically two tools: a custom deep log analysis made available to us through the Europeana Connect project’s University College of London member and Google Analytics.
In this post I’ll share some thoughts on Google Analytics. First of all I should disclose that I’m by no means an expert on analysing web statistics. My experience is limited to search applications withing the heritage sector, but I guess I know the basics. My thoughts in this blog post were awakened when a Europeana Connect colleague pointed out to me that the Europeana portal is at its heart a search engine! Well, duh! I know this of course, but what he said and the context in which he said it (we were talking Europeana portal stats) made me realise that analysing the statistics of a search engine reverses many of the common wisdoms of analysing your webstatistics. Two examples:
Most site owners want the users spend a hight amount of on the site on average. This makes sense for sites that want to be immersive and where the user can be captivated and explore. A good example would be a e.g. a virtual exhibition. In a search engine however you’d want the user to quickly find what they’re looking for. People usually have something very specific in mind when using them.
Another metric is the bounce rate. The bounce rate is the proportion of users who arrive at a page on your site and promptly leave. Not good for most types of websites so many site owners strive hard to keep the bounce rate low. However, a high bounce rate can also mean that the user found exactly what they were looking for and then left. For Europeana this could be that a user has found an interesting object and clicks through to the original context on the data provider’s site – exactly what we would want! So in the case of Europeana the bounce rate must be analysed in context:
high bounce rate from our first page = not good.
High bounce rate from object views = not necessarily bad, could be exactly what we want.
The same would go for any search engine I would believe.
My take on Google Analytics is also that for a site like Europeana looking at total page views, number of visitors, unique vs. returning visitors and referral sites only tell you so much. If you’re a search site like Europeana looking at the Google Analytics Site Search statistics could well be more interesting! It’s here you can see exactly which keywords users have searched for and how often, the keywords they use to refine their initial results, and so on. Needless to say this tells a lot about what users are looking for on Europeana and what they are finding when they do so. Information that can be used not only in the refining of our search features and functions, but also inform content strategy!
So perhaps we should start looking a little bit more at what our users are searching for and how they naviage within our site to complement our existing knowledge on how many they are, where they are coming from and how many of them are regulars? Another new and interesting feature in Google Analytics is the In-Page Analytics, but I haven’t really had time to play around with that one.
Do you have experience with analysing the usage of search sites within the GLAM-sector? Which tools do you use and how systematic is your reporting and analysis of the statistics? What angles do you use to sift useful information out of all the statistics? Please share!