Oftentimes you will hear a business analyst make a prediction, and then hedge their bet on it coming true by issuing a caveat that begins with the words “On the other hand…” It’s hard, if not frustrating, at times to extract the message from such elusive forecasts. In the world of business intelligence such predictions amount to what Gertrude Stein described as a situation where "There is no there there”. Well, putting the “there” back in “there” is exactly what Dashboards are meant to do. In this article I am going to discuss some high-level concepts that will go a long way in ensuring that the business message gets through in your dashboard design.
From a very high-level perspective, an effective dashboard design should:
- Display an organization’s key performance indicators.
- Filter out distractive business data noise.
- Present in an aesthetically pleasing way to the user.
There are other important functionalities that are required in good dashboard design, but those mentioned above are rudimentary to any effective design. In short, when users inspect their dashboards in their everyday use, they should not have to decipher the messages that lie therein. If there is a problem that requires the user’s attention, that problem should be immediately obvious to them. If the message is lost amongst bells and whistles, then the dashboard design is a failure. On the other hand, a dashboard should be pleasing to look at, and engage the user to continue to use it.
There is no denying that a good dashboard design requires some ergonomic and aesthetic qualities to support it. Nobody wants to stare at a boring, lifeless or awkward interface, regardless of its potentially superior technological traits. A solid dashboard design seamlessly fashions technological competence with a beautiful ergonomic layout. The message is conveyed with style and efficient delivery.
Where the “rubber meets the road” in this equation is in the choosing of the data visualization components that will be used to display the message. Specifically those charts, gauges, and maps used to amass and display the data needed to indicate the status of the procedure, department, or organization that is being monitored or measured.
Here are some tips on choosing your components:
- Using charts to display numerical or statistical data is an efficient means to rapidly visualize data.
- Use gauges sparingly, since they inherently take up a great portion of the overall interface, and present a limited range of information. However, keep in mind that when gauges are used properly, they can be a powerful component in giving the user a sense of progress, or a discrete measurement like a temperature reading.
- Geographically significant data should be tied to the use of a mapping component. No other data visualization component can convey a geographical message better than a good map.
- Design your interface to display the most important data within the most visible areas of your interface. Think of your interface as a piece of real estate -- certain areas will have a better location than others in terms of their immediately visibility. Generally speaking, components that are placed at or near the middle of the interface have the highest visibility, and are therefore the components that users see first. Next in descending order of visibility are those components that are placed at the top left and right corners of the display, followed by those that are placed in the bottom left and right corners of the interface display.
It is worthy to mention that a great deal of study is still ongoing with regard to the “screen real estate problem” and consequently a varying number of theories have and will continue to emerge. Simplistically viewing this problem, one would tend to believe that visibility is tied to the user’s line of sight. By this I mean, the areas where the user is likely to look when first presented with the interface. Logically, it seems that the middle of the display would be most visible, followed by the top left and right, then the bottom left and right sections of the interface. While many studies absolutely confirm this suspicion, others nevertheless tend to refute parts of it. Specifically the difference marked in those studies demonstrates that while the middle of the display still remains as the most visible area, in descending order of visibility the next most visible area is the top right moving across to the top left, followed by the bottom right moving across to the bottom left. This is the exact opposite of the previous findings.
In conclusion, in order to create a good design for an effective dashboard, one must be judicious in choosing the correct components to use, and then properly place those components within the interface to display their data to the user in a clear and concise manner. These are the basic ergonomics of a good dashboard. On the other hand, one should be mindful to create a dashboard that is aesthetically pleasing to the user without interfering in the overall business message. Unfortunately this all starts to sound like the predictive advice of those “On the other hand” economists that frustrated former President Harry Truman enough to say "Give me a one-armed economist".
Copyright 2008 - Dashboard Insight - All rights reserved.