(A work in progress)
An early WIP II task was to investigate the statistics and data that is currently available that could usefully inform our work. The starting point was a conversation with Keith Bolland (Unlicensed) but more work will be required before we have useful information and it actually influences what we do.
I suggest that the ways in which it might be used include:
...
The following notes are based on my conversation with Keith, but include some thoughts I have had since and from looking around GA (as I now have access). I believe that we have enough data to act on (in ways described below) if there is a willingness to do so.
And this is a first draft and as such would benefit from updates and comments by others. Go for it.
Information obtained from the analytics could be used in WIP II in:
- Making high level IA decisions: Allow accurate drafting of templates and content, with the minimum chance that rework is required.
- Locating content: How far down is the most frequently used content located? Should we cull content if it is seldom/never accessed?
- TagsSelecting tags/search terms: How to improve the hit rate for searches.
- Validate Validating the key drivers for WIP II: Importance of mobiles, extent of duplication, etc
- Measure Measuring success: With KPIs the analytic data might help measure the impact of WIP II. This might lead to
- New Prioritising new functionality prioritised higher: We are unlikely to be able to measure success using the current tools/functionality.
The following notes are based on my conversation Key points my talk with Keith , but include some thoughts I have had since.
- GA is the main source/tool for analytics: GA contain a whole heap of data about most of the sites in question, but little analysis is currently done on the data beyond what Google provide in the reports.
- This means the team is "data rich but information poor"
- Reasons for the lack of analysis time pressures, a lack of reason or a work driver (i.e. ability to out-work insights gained from the analysis).
- Maybe in the University there is little willingness to have empirical data guide decisions (need to validate this with team).
- There are very few KPI's at present, so even if the technology was not a limiting factor it is impossible to determine if we are succeeding.
- And where people have attempted to set some they are qualitative and therefore less amendable to measuring by analytics
- Need to be honest on how challenging this will be and start by going back to user needs. Listen to our users.
- Although difficult to do the following is probably essential if (good) use of analytics is to In order to make better use of this data there would have to be:
- Define a reason: For example "analyticity will inform IA and content placement decisions"
- Draft a proxy for a web strategy: To provide higher level direction. For example "The purpose of the web channel is to ..." or "Our three main products are student recruitment, service delivery and public engagement."
- Agree some KPI or success measure: Reduced number of clicks for users to get to what they want, or Lower the bounce rate to X, or Start using goal conversion (or even completion) rates?
...
- Too many site exits that are not linked to a goal/conversion.