Tableau Software : Why design theory should apply to business intelligence

The second level of processing is behavioral. When someone wishes to use an everyday thing, such as a door at an entrance to a building, they have an objective (enter the building). Using that object, can he or she successfully complete the task at hand? Are the things the object can do apparent and obvious? Don Norman shares many failures in the book, including examples of poorly designed doors where a lack of hinges and push/pull signs thoroughly confuse people. It's just as important on your dashboards: Can the person understand how to interact with it? Can the dashboard answer the users' questions?

Check out the excellent dashboard on animal gestation and longevity below, designed by Eric Brown. Do you see how can you interact with this dashboard, or what you can click on?

Longevity and Gestation in Humans and Animals by Eric Brown

There are eight ways to interact with Eric's dashboard. Did you spot them all? I'm guessing that you didn't, but that's ok. Not all the 'affordances' (things an object can do) have a 'signifier' (a visual indication that it can be done). If there's nothing telling your audience what they can do, how do you expect them to know?

Whenever advising dashboard designers, I implore them to observe other people using their dashboards. This test always reveals whether the behavioral aspect of their dashboard is successful or not.

Also, note: You don't necessarily need to signify everything; training or familiarity can get over that problem, but it is something you need to consider. Can you sit and train your users or change the experience for first-time users?