Telling the Right Story
I have worked on several custom Dashboard Solutions and, in this article, I want to share with you a major pitfall that comes with the territory. The clients I work with range from the financial industry to the food industry; however, the same challenge arises – visually depicting data in way that can be intuitively understood at a glance. The problem stems from having a set of useful metrics but lacking the best way to tell the story. Here’s a simple example.
“I want to show our sales reps how they are doing against one other at the next sales meeting in order to get them motivated,” says the Director of Sales, Davis, from company ABC regarding his concern with his team. In the next sales meeting, he brings up this chart in a Power Point Presentation showing each reps YTD performance as depicted in Fig. 1.1.
“Jim and Sally, look how you compare to everyone else.” Jim and Sally stare quizzically at the graph and stay silent. The Director doesn’t understand their reaction, and exclaims, “Obviously, you’re not performing as well as the others! Look how low your peak is compared to everyone else’s!” Jim and Sally react with a combination of revelation and disappointment.
Although Jim and Sally finally understood what Davis is saying, it should not have required an explanation in this simple case. Graphs are a form of data visualization for easy consumption and to be effective they must be able to tell the right story with minimal explanation. Line charts are usually used to depict trends over time. A more appropriate chart for this particular example would be a bar graph as depicted in Fig. 1.2.
Jim and Sally would notice how small their bar is relative to the others and automatically supply Davis with the desired reaction from the start.
A more poignant example is the classic single metric value. Davis is at it again, and this time he is concerned with yesterday’s sales revenue. He decides that this information needs visualization to drive home the disappointment. He gathers the sales team and shows them the gauge as depicted in Fig. 2.1.
“Look at yesterday’s sales revenue!” he gripes. Based on his tone of voice, the team assumes this must be bad. Sally raises her hand and asks a very relevant question. “Is this bad?” Davis screams, “Of course it’s bad! It’s at least half as low as our daily average!” Wait a second Davis. Why didn’t you just show another needle depicting the daily average to get the point across as depicted in Fig. 2.2?
In the case of a dashboard, there won’t be Davis’s tone of voice to suggest the severity of the value. Therefore single values have to be put in context for the user to understand their significance.
Here is a recap of the two main points in making sure the dashboard is telling the right story.
- Choose the right graph visualization. A quick rule of thumb: line graphs to show trends, bar graphs to show comparisons, and pie graphs to show ratios.
- Put single values in context. Common things to use are: ranges of good \ bad values, averages to compare with, a previous value to compare with (e.g. last year vs. this year), or a combination of any of these three.
About the author
Alexander Chiang leads the Consulting Services Department for Dundas Data Visualization Inc. He has advised various Fortune 100 companies on the technology platforms and tools that best fit their dashboard requirements. His many years of user interface design and understanding of data visualization techniques have enabled him to facilitate effective software solutions for the broader BI community.