Dashboards: visualization and usability (part 3/4)

In this four-part series on dashboarding, in the first article we made clear to you when a dashboard is the solution for your information needs and what this dashboard must meet to ensure that you can extract the right insights from the data. Then, in the second part, we showed you how to turn KPIs into an information requirement and data model. In this third part, we will show you how to design visualizations and a dashboard so that it is user-friendly and answers central questions.

Choosing the best visualization for your dashboard

A well-designed dashboard draws your attention to parts where action is needed. Visualization plays a role in this not to be underestimated. People are much better at processing and appreciating images than written text. The prerequisite, however, is that the visualization be set up to best support the message. In our example, where a manager wants to achieve 20% sales growth over last year, the dashboard includes the number of new customers and average order value. Visualization becomes interesting when the result has a benchmark and when the use of color draws attention.

Start with the message

It is tempting to design visualizations in the tool, but our advice is not to do so! Start with the message. What is the most important thing I want to show? Who do I want to show this to? What question does this visualization answer? As we’ve advised before: start with a rough sketch of your idea on paper. As a result, you maximize the creative process. Only then take your idea into the tool you use to create the visualization. This method ensures that you keep focus on the message and that your attention is not consumed by all the fancy features a tool offers. So first sketch a general picture for each KPI you want to display and then choose an appropriate type of visualization. There are an extraordinary number of options available, if you need inspiration take a look here. This seems easy and an open door, but we come across visualizations every week that leave us wondering exactly what the message is.

Use of color in your visualizations

Once the foundation of your visualization is in place, you can expand it with shape and color. Keep in mind that color has a meaning in dashboards, so use it only for a purpose such as distinguishing between categories or clarifying a trend and keep color blindness in mind. Over 10% of the world’s population is colorblind, and choosing some colors that go well together is harder than you think. Choosing a color palette for building visualizations is best left to experts. We therefore make extensive use of a color generator that takes color blindness into account and chooses colors that match well, like the example below.

In many cases, you are limited in your choices because you rely on your company’s or client’s corporate identity colors. If these corporate identity colors prevent you from properly conveying a message, consult with the person ultimately responsible for the corporate identity. Often you can achieve a lot with the adjustment of saturation and hue.

Formatting your visualizations

In addition to color, the formatting of your visualizations plays an important role in getting the message across. It is especially necessary that few distractions be present. Therefore, use one font, the same font size between visualizations, and use accents (bold, italic) only when necessary, such as in a title. Also, don’t focus on adding additional features to a visualization, but rather ask what other parts you would remove. The fewer distractions the better, like the example below.

Source: Courtney Jordan, ‘Make your data speak for itself’

Guide your user

By now you know our position on dashboards, it must:

• answer central questions;
• Draw attention to parts where attention is needed;
• lead to action.

For dashboard construction, this means that the first screen shows the status of KPIs: Where are things going well, where is attention needed? Next, a user should be invited to explore root causes. Why is the average number of orders per customer lower than last year? You do this by using mouse-over tooltips with additional information or by incorporating navigation. In the latter, users click on a visualization to see a deeper look at the figures. With this, you provide the user with information so they can proceed to action.

Use a basic layout

As a rule, we work for clients with a wireframe, which is a basic layout/sketch of the dashboard where filters, navigation, explanations and legends have the same place. This also starts on paper and after coordination we take this design into the tool. There is no basic design that works best for every situation. Below you see an example, however, more important is that you work consistently. This makes it easier for end users to use and understand a dashboard. Place titles, buttons and filters in the same place when your dashboard consists of multiple tabs. And don’t forget to explain the “hidden” features. Such as interactions between visualizations or tooltips. With a symbol or short explanation in the subtitle, it is immediately clear that a visualization has these options.

Sample dashboard wireframe

If you keep adding new filters as you build your wireframe or your dashboard, then you might wonder if the client doesn’t need a dashboard but an analyst. A long battery of possible filters is cluttered for many users and leads to confusion. User A sees different results than user B because a slightly different filter selection was chosen. In our experience, the limit is 5 filters in a dashboard. Need more filters: then you either have a complex business model or the goals of the dashboard are not as sharp after all as you initially thought. In our first article in the series, we take an in-depth look at whether a dashboard is the best solution for your question.

About data visualization and dashboarding, we personally don’t run out of ideas easily. Want to know more? Then visit our website or register for our data visualization training or hands-on training in PowerBI.

Contact

Want to know more about this topic? If so, please contact us!