Guest Post: Get The Real Story From Your Dashboards
This is a guest blog post written and contributed by Tara House, Director at Bluewolf, a global agile business consulting firm and GoodData Solutions Provider. In this post, Tara shares 5 best practices for dashboard design.
“Not everything that counts can be counted, and not everything that can be counted counts.”
- reads the sign in Albert Einstein’s office at Princeton.
Early in my career, I worked in the art department of a newspaper and was lucky enough to be mentored in the art of creating effective charts. The main goal was to ensure the chart type chosen was effective in communicating the data, not obscuring it, and to ensure the reader could see the “story” of the data quickly. The same rules apply in creating dashboards.
I had the good fortune to work with GoodData very recently. This company helps convert big data into profitable insights and strategies for business executives by offering a disruptive cloud-based enterprise business intelligence platform. We collaborated to show how effective investigation and counsel when envisioning a dashboard will give the executive more actionable information. In newspaper vernacular, it tells a story.
At first glance, it looks like Janice is not working up to par with her colleagues. However, let’s look at the same information in a different way (see chart on the right). Notice that instead of a single color bar chart, it is now a stacked, colored bar chart. A time element was also introduced. The stacked bar chart quickly shows that a major product issue cropped up this week. And in fact Janice is not avoiding work. She doesn’t work on product calls, but her colleagues do. There is an issue that must be dealt with by management.
This illustration shows how by not choosing the proper report attributes and chart type, you can obscure the information that would be most meaningful and actionable to you, your staff and the business. Below are 5 best practices we at Bluewolf and GoodData counsel when we approach dashboard creation.
- Understand what attribute dimensions are most appropriate for your business or role, and which chart types will best visualize those dimensions.
- The best charts show comparative data. This can be historical, year over year, month over month, groupings or date progressions.
- Dashboards should be organized so that trends and Key Performance Indicators (KPIs) stand out and are easily identifiable. The top left chart is usually the chart most people look at first in countries where written languages are left to right. Choose this chart carefully. Also, think about what charts will be seen when the dashboard is first clicked on. The most important charts should be seen in that first screen.
- Too many colors in a chart will be overwhelming to the reader and will occlude the message. Try another grouping instead. Keep background colors to a minimum.
- Titles should be descriptive and any assumptions clearly explained. Attribute filters, such as date ranges or territories, should be applied at the dashboard level rather than individual report level whenever possible.
Albert Einstein had good reason to like the quote above. Data can be misleading, and you don’t want to be led to the wrong conclusions.