One of the key concepts used by story tellers in any medium is that of answering the obvious next questions. If a filmmaker shows a close-up of a previously unknown face the obvious next question would be ‘who is this person’, the filmmaker then proceeds to answer the question.
The obvious next question serves to create a flow in a narrative, maintaining the interest and the concentration of the viewer.
Data communication is primarily about the communication. As such techniques that are effective in other forms of communication apply just as well.
We’ve previously made the point that communication planning is important for reporting design; that reports should be designed to meet specific user needs rather than as a ‘one-size-fits-all’ approach.
Successful visualisations also take into account the obvious next question that showing a viewer some data will raise. The designer needs then to provide answers to those questions. Naturally this cycle might repeat.
As an example, say I show a visualisation of attrition rate across departments. The obvious next questions could include ‘why does this happen?’, ’is this serious?’ and ‘what do I need to do?’
These questions can be identified early in the design process. We tend to use lo-fi prototyping and testing with users as a way of validating and identifying omitted questions. It will be an area we explore again at final usability testing.
This links nicely with Ben Schneiderman‘s information-seeking mantra of ‘overview first, zoom and filter, then details-on-demand’. What designing for the obvious next question does is provide guidance on which of the many possible paths the user is likely to want to explore.
Another way of looking at this situation is through the use of information experience mapping. Looking at a visualisation of report is not an isolated activity. It’s important to understand what the viewer was doing before they go to the report and what they are likely to do after viewing it. Decision trees are useful to understand the possible paths and the designer needs to be as extensive and complete as they can be with the options.
Only when the visualization is rooted in the context of how it’s going to be used and the obvious next questions have been understand and provided for the visualisation will be seen as natural and valuable by users.