Going beyond basic web analytics

A key part of our Content Strategy is about being user driven and evidence based when it comes to improving our digital online presence. Russell has previously posted on the increasing trend of mobile visits and Matt spoke about the importance of responsive design as we move forward.

We are only really at the beginning of this and we know that we had and have some gaps in our feedback as we have only really used volume data to help us understand the performance of our website – e.g. how many people have visited a page. But that doesn’t go anywhere near telling us if the content or service is actually helping people.

The content strategy states:

User Driven (Evidence Based)

Using web-based analytics is essentially about looking at how your website is being used. It gives you a feel for how many visits your site gets over a given time period, how many of these visits are unique, what the most popular pages are, as well as a whole range of other information. In order to measure and evaluate our content effectively we need to adopt a blended approach which uses traditional web analytics along with additional sources of information/data. These would include

  • Heuristic Evaluations (end users and staff)
  • Data from Customer Service Centre
  • Data from Freedom of Information requests
  • Web Analytics
  • Feedback and complaints
  • Profiling and Content Engagement
  • Online Surveys

People can make a suggestion via this blog as well if they’ve got an idea…

One of the things we recently implemented on our main website in support of this plan was Govmetric, this is a tool which helps us understand what people think about our website through the use of faces (like the ones below).

A number of other councils are also using this, so you may have come across it on some sites you’ve visited (if you do actually visit local authority websites!), we also use it within our Customer Service Centre (CSC) and in a couple of council buildings (libraries). The platform is multi-channel and can help us better understand how our access arrangements are performing and we need to learn from the wider channel feedback to help us improve our web content.

As part of our commitment as a team to be more open and transparent about our work we thought we would share the first months stats from Govmetric.

We’ve not formally promoted the fact we have this yet (although will shortly) as we wanted to make sure that it worked and we actually got data which we could use, so here is our first months stats.

A total of 859 in the first month…this currently represents around 0.4% of our unique visitors to the site during the month of July – so it is as expected to be honest.

As you can see the site is “average” – which is pretty much what we thought before we started this process, but it is great be able to start supporting these views with evidence and feedback from users of the site – the stats do go a bit deeper but I’ve only decided to share the top level performance at this point in time.

By way of comparison in the same period we also had the following stats from google analytics with the most popular content area being libraries and in particular the library catalogue:

  • Visits: 293,206
  • Unique Visitors: 189,727
  • Pageviews: 989,566
  • Pages / Visit: 3.37
  • Avg. Visit Duration: 00:02:08
  • Bounce Rate: 56.07%
  • % New Visits: 51.40%

So what’s next?

We need to look at implementing this across all of our sites in order to get a holistic view of our web estate and to ensure that we can understand how content is performing and meeting users needs. We need to start providing a package of statistics instead of simply number of people visiting in a meaningful format for Heads of Service and colleagues to respond to.

Carl Haggerty

All things digital - content, strategy, communications, innovation, engagement, participation, data and people

More Posts - Website

Follow Me:
TwitterLinkedInPinterestGoogle PlusFlickr

2 thoughts on “Going beyond basic web analytics

  1. Ben Darlington

    We recently incorporated a web chat service into our platform and, as part of the accompanying analytics toolset, we provided a facility to output the content of chat sessions alongside page view history for the specific visitor. It’s providing unbelievably valuable research in terms of understanding how to improve a site. Once you know what somebody was specifically looking to achieve on your site (by reading the chat session) and then at the pages they viewed, you can quickly see how to improve things. About 5% of people offered an ‘in page’ chat session will use it, so we get a response rate of 10 times that of Govmetric.

    I’ve found that GA stats can be very misleading because they don’t filter out non-core visitors. So when you use GA stats, there is a danger that you allow the preferences and behaviour of non-core visitors (typically 70% of your visitors) to determine your understanding of site and web page performance.

    I like the idea of Govmetric but I think it needs to be combined with other analysis to actually produce real value? So we’re building a Govmetric equivalent service into the platform but we’re marrying the results with analysis of behaviour on the web channel and giving the facility to group respondents by profile-type (core, non-core, loyal, non-loyal, path-through-site, services used, services not used etc. etc). I think we’ll get some real value here.

    Reply
    1. Carl Haggerty Post author

      Hi Ben,

      That sounds very interesting, I still very much like the idea of what you outline and we are developing our approach to analytics so we need to stay in touch.

      Carl

      Reply

Leave a Comment

Your email address will not be published. Required fields are marked *

 

Comments are held for moderation. House rules