What is a good practice for dashboard design?
Creating a dashboard seems simple with the drag-and-drop concept of Power BI or Tableau. However, many dashboards were created but not used because they can’t serve users’ objectives or are too difficult to understand. In this article, I will cover good practices for dashboard design from the concept of user-oriented design to visualization arrangement.
Contents of this article
- Who are dashboard users?
- Problems — Reasons — Actions
- Visualization selection
- Visualization arrangement
Who are dashboard users?
First of all, you must know the dashboard's primary users. Then, you can properly design a dashboard to match users' expectations.
These are the list of questions you should be able to answer.
- Are they executives or low-level staff?
This question can determine how frequently they are likely to use this dashboard. If they are executives, they may have limited time to use a dashboard, so the dashboard should be as simple as possible such as having only a small number of visualization per page. - Are users familiar with technology?
If users are familiar with technology, you can include graphs that can use drill-down or add more options for users to customize. On the contrary, If users are not tech-savvy, you should avoid hidden data inside drill-down. You can add new visualizations or pages, so users don’t need to drill down graphs by themselves.
Problems — Reasons — Actions
Now, you have users as the starting point.
Next, It is important to know the direction of your dashboard development by using the Problems — Reasons — Actions Framework.
Think of a dashboard as a direction sign to guide the users toward their goals. A great direction should specify the destination clearly and have an arrow to show how to go to that destination, so lost travelers know exactly what they have to do.
Problems are the current situation of the users that users aren’t satisfied with. For example, the profit doesn’t meet the target or the number of application active users is low. In another word, it is the destination where users want to go. The problem should be easily seen on the dashboard. It usually locates at the topmost or leftmost.
Reasons are the source of the problems. These parameters can affect the problems. For example, the number of sales of each product could affect company profit. You can group the parameters in many ways such as by country, product category, or the age of the customer. Users will get an overview of the problem.
The following questions may be a guideline to identify the reasons
- Which are critical parameters that could affect their decision?
- What do they want to measure?
- How could you calculate this parameter (problem’s parameter)?
Actions are like a kick-off point. As users know the reason for the problems, they should know how to solve them. The actions part should elaborate enough for the users to implement a plan from the data they know. It can show the detail for a marketing purpose such as which products give high profit but have low sales volume.
Visualization selection
You should have the skeleton of the dashboard in your mind now. Then, it is time to craft the visualization.
The visualization can be grouped into 4 groups based on the objectives:
1. Show change over time (Trend) such as a line chart, area chart
2. Compare data such as a donut chart, pie chart, bar chart
3. Find correlations such as a scatter plot, bubble plot
4. Highlight a value such as a card, gauge
It is crucial to select visualization to match the objective.
Visualization arrangement
You have many jigsaws of visualizations to include in the dashboard. The visualization placement is a mix of art and science. These 3 principles are a guideline for designing visualization locations.
- 5-second rule
The users should find key info/metrics from the dashboard within 5 seconds. The font size should be large enough, the color should not be too distracting, and the visualization alignment and size should be easy to follow. - Only 5–9 visualizations per page
Miller’s law states that the magical number 7 is the number of objects an average human can process. This 7 ± 2 capacity is a universal phenomenon that has been tested in several experiments and is widely adapted by UX/UI design. In dashboard design, it is great to start with this number to limit dashboard complexity. If there may be more than 9 visualizations on a page, you should consider splitting to a new page. - Place in the reading order
Normally, humans read from left to right and top to bottom (in English). It is the same for a dashboard. Users will read graphs from left to right and top to bottom. You can place the visualizations from problems to reasons to actions from left to right or from top to bottom, so the users can know the problems, see the overview reasons, and know the actions they need to take.
Conclusion
The user-focused dashboards must start with the users. Then, find the direction of the dashboard to the user objectives using the Problems — Reasons — Actions framework. Follow by selecting visualizations that match the purpose. Finally, place the visualizations on the dashboard when keeping the 5-minute rule, Miller’s law, and reading order in mind.
This article was written from my personal experience. If you have any commend or suggestions, feel free to share them in the comments.
This article was inspired by
Dashboard: Theory of Problem-Reason-Action
4 Principles of Dashboard Design
Please clap and follow me if you want to read more data analytics content.
