For organizations that are just starting to get their feet wet in business intelligence (BI) and data visualization, I sometimes have them look at dashboard applications centered on their favorite sport or hobby. By using subject matter that is familiar and exciting to a budding dashboard designer, I can convey sound dashboard design principles, especially with respect to the presentation layer. The newbie always comes away with a better understanding of how to best create visually compelling BI applications which are highly interactive and intuitive. Since many of my clients are golf enthusiasts, there are a few golfing dashboards that serve as wonderful educational tools, providing me with a “template” from which I can illustrate some ways to effectively track performance from a visual perspective.
Designing the layout for a dashboard is half art and half science. No matter the type of dashboard, there will always be a battle between form and function. While there is no standards committee that has formally endorsed a list of best practices for designing dashboards, there are certain rules that designers should heed. Here are some of my favorite dashboard design best practices:
When designing the overall layout for a dashboard, start from a high level of visual abstraction and then drill down to specific details, such as how to physically integrate data into the dashboard. It is not enough to know what data dashboard users require, the designer must understand how they viscerally think about their business. Just as a nuclear engineer would not feel comfortable in the cockpit of an airplane, or vice versa, not all dashboard users will be happy with an out-of-the-box layout that fails to provide a level visual customization that is suited to their business world view.
Decide early on the minimum font size for the text on your dashboard. More often than not, you will want to save the display of smaller fonts for reports. Why? Because the best business dashboards instantaneously convey business realities to users; after a few seconds of interaction, they are able to objectively assess the most mission-critical measures of their business. It is not possible to quickly convey an organization’s state of affairs if data is presented in too granular of a format from the outset.
When deciding whether or not to show more information on a dashboard screen, it is best to err on the side of not directly showing the data. Instead, enable a click-through so that the users can “drill into” desired results to gain more detail. If the dashboard is cluttered with too many KPIs or gadgets, the user’s focus is sure to wane and wander.
Learn to recognize whether or not your data visualizations are leading or lagging in nature. Leading indicators usually anticipate a future business condition. An example may be the “average daily water level of the Mississippi River”, which will directly influence a lagging indicator, such as the “monthly transportation cost per unit of crop yield”. It is far too common for even seasoned dashboard designers to misplace or improperly combine leading and lagging visualizations. It is important to recognize the reactive and proactive nature of each type from the outset of dashboard design.
As much as possible, limit the amount of text or content on the dashboard that is truly static. Labels for data and visualizations can be attached to supporting metadata that describes exactly what the label or text means from a business perspective. For instance, the name of a KPI (such as “number of units processed per hour”) would show a full explanation when clicked or when the cursor hovers over the label. Note that displaying full explanations may be come tricky, especially when users are allowed to create custom queries. Decide early in the design process how a solid metadata repository or metadata layer will lend support to the data and visualizations on the dashboard.
In the last few years, visual data mashups have become the rage in dashboard design. But creating data mashups that convey powerful messages to the user can be an extremely difficult proposition from a logical standpoint. Again, golf dashboards provide an outstanding case study in what can work with respect to mashups. Many commercially available golf dashboards will let the user integrate and match up the performance statistics with data and pictures taken from real-life golf courses all over the world, greatly enhancing the user experience, and providing a much clearer and three-dimensional crystallized picture of their overall golf game. Not only can golfers now track performance on an inter-club or inter-course basis, they can engage in rudimentary predictive analytics, anticipating the challenges of a new course based on past performance on similar courses. In choosing a golf dashboard, prospective users will be wise to check and see what golf courses are included in the package and how robustly they are conveyed via data visualization techniques. (Many of golf performance applications can be somewhat limited on the courses that they include out of the box.)
The best practices I have bulleted above should be fully understood before embarking on any dashboard design tasks. They will provide the foundation for a business dashboard with a robust presentation layer. Just as a good golf pro can help reduce a golfer’s handicap and give them an edge over their peers, a richly visual dashboard application, when designed with just the right amount of information, can increase competitive advantage—across all industries and business verticals. Such a dashboard will provide new insights about one’s business on a daily basis, with enough visual consistency and maneuverability to enable business managers to holistically assess their company’s performance across multiple lines of business. Having a huge functional palette on a business dashboard will not guarantee the success of the dashboard. In order for users to gain keen insights into their business operations and walk away with actionable knowledge, a tremendous amount of time must be spent on deciding the best way to display every piece of data that will find its way onto the dashboard.
But keep in mind that flashy dashboard layouts and data visualizations will do little to improve corporate knowledge management until business and data fundamentals are addressed: Business processes will need to have matured to an acceptable level and basic data integration infrastructure will have to be in place. Moreover, there must be a deep understanding and agreement on which facets of the business to track and report against. Key Performance Indicators (KPIs) will have to be clearly defined before forging too far ahead with initial dashboard application development.
About the Author
William Laurent is one of the world's leading experts in information strategy and governance. For 20 years, he has advised numerous businesses and governments on technology strategy, performance management, and best practices—across all market sectors. William currently runs an independent consulting company that bears his name. In addition, he frequently teaches classes, publishes books and magazine articles, and lectures on various technology and business topics worldwide.. As a Senior Contributing Author for Dashboard Insight, he would enjoy your comments at firstname.lastname@example.org
Copyright 2011 - Dashboard Insight - All Rights Reserved.