When you’re designing a dashboard to track and display metrics, it is important to consider the needs and expectations of the users of the dashboard and the information that is available. I have found that there are several aspects to consider when creating a new dashboard in order to make it a useful tool that your users will rely on. Each of the elements needs to be considered in the complete design.
The term “dashboard” comes from the dashboard in modern cars. When driving a car, you need to be able to quickly and easily monitor critical performance information. Speed, fuel, and engine temperature are considered critical and are almost always displayed most prominently.
Some cars include extra data, such as oil pressure, battery charge, and a tachometer, which, while informative, are in most cases not very actionable to the typical driver. Many cars have moved to using indicator lights to alert drivers to out-of-compliance standards for this kind of data, including low tire pressure, when an oil change is due, and other maintenance conditions.
This concept of a car’s dashboard is directly applicable to our customers: Show the most relevant information in real time in a format they can swiftly and easily understand. Monitor and offer access to additional information elements users can get to if they desire, but when a risk of compliance violation is detected, add that to the dashboard in a highlighted manner to gain attention.
For a mnemonic device to help you easily remember the qualities that make a good dashboard, just remember the acronym “VITAL.”
Of course, the data you share on the dashboard needs to be correct. But “valid” also means considering the correct units of measurement, rounding, and frequency of update.
Beyond the format the data is available in, also check with the users to see what they need and what makes sense to them. We could show car speed as FPS (feet per second) or MPM (miles per minute), but as US speeds are regulated in MPH (miles per hour), that is the format that will be most useful to American customers. But what about places where speeds are in metric?
The same is true for your data: Consider your audience in the design and use the scale and format they expect.
Beyond simply supplying the information, it is critical to supply a perspective. In cars, this is done by having areas on the gauge displayed in red to show danger areas. The same thing is needed in your data.
If you supply a value of 3.5, the user needs help to understand if that is good or bad. This can be done by color or supplying target details. You can also help your users if your data has control ranges (e.g., an X-bar). Show those with your data so a user can easily understand if the data is showing an issue or your system is under control.
Your users need a fast way to see how things are going and recognize if they need to take a deeper look. Help them by including a perspective on the data.
Nobody would tolerate a five-minute lag in their car’s speedometer, and the same is true for your dashboard. The value of information degrades exponentially as time goes by. It is often better to supply data that has a known limit of precision but is very current versus taking the delay needed to improve the data’s precision.
Work with users to find the correct balance between speed and precision. If you are tracking an emissions value for a byproduct, the need to be timely is extremely important. Showing a trend over a time period is also valuable, and timing is likely to have more priority over precision.
Be aware that not all users will have the same opinion, so it is important to build a consensus of the users of the dashboard and then communicate that to them, so they can understand any limitations or compromises made.
When selecting the information to share in a dashboard, it is important to make it actionable, meaning the user can do something to change the data when an issue is displayed.
In our car example, when we detect via the speedometer that we are going faster than planned, we slow down; when we are running low on fuel, we buy more. A dashboard that only includes data that can’t be changed by user intervention will lose its value and relevance to the users.
Not all data needs to be actionable, but it should be informative to help the user gain a better understanding of the status. In cars, the external temperature is data you can’t change, but it is nice to know while driving.
Okay, this quality is so named more because I had to make the mnemonic work, but aesthetics are a critical element in your dashboard. Making it easy to read and interpret helps the dashboard be more effective in its messaging. You can use graphs, charts, heat maps, dials, and all the other myriad data display methods, but you need to consider your users and the data in the decisions.
There are traditional design best practices for roadside billboards that are applicable to a dashboard—namely, that your users are in a hurry, and this is not their primary focus. See if you can apply these rules to your dashboard design.
- Keep it simple; tell me what I need to know and why I should care
- Make it short and easy to read; I don’t have time to analyze a book
- Design responsibly; a good design is the key to readability
- Use colors responsibly; only use green for good, yellow for risk, and red for bad
- Keep your format of good to bad consistent; not all your readers can see color
- Provide a way for your customer to get more detail if they want it; drill down enabled graphs to the data behind the graph
This is just the start of good dashboard delivery, but taking the time to develop a quality design will yield a much better project. In most cases, it also can be delivered faster than the “winging it” concept of just getting something published fast and working to improve it later, as putting thought into the design up front helps you avoid the scope creep common to this kind of project.
As part of my design, I build sample mockup pages and share them with the developers and the customer to make sure we all understand the deliverable. The old maxim “Failing to plan means planning to fail” is especially true in a highly visible solution like a dashboard.