10 dashboarding best practices to get the most out of analytics
Dashboards are the control centers for your analytics, putting vital metrics at your fingertips and creating a visual representation of data mining. But building them can be difficult and sometimes confusing. Not all dashboards are created equal, and it’s important to understand the key components that will optimize your expertise for success, impact and peak performance. Design and simplicity play vital roles, as does a mindful focus on key performance indicators (KPIs) from the very first moment.
If you’re building a dashboard and looking for ways to optimize it, keep these 10 best practices in mind:
1. Consider a dashboard a blank canvas for your own work of art
An artist staring at a blank canvas and a tub of paint understands that these are merely tools for the end game: a work of art. When you design a dashboard, adopt a similar perspective by maintaining a mental target of your desired deliverables before you begin. What do you as a user want to show management or another audience via this dashboard, and what types of actions do you want to compel them to take when they look at it?
A dashboard can serve as the bridge between data and action, but to build it properly, you need to understand who will be looking at the dashboard and what its metrics will compel them to do. If you want to create a dashboard to look at the top spend and expenses in your organization, the goal that you should have in mind at all points of creation is showing management how much they spend in each specific timeframe. And let it literally be a canvas: I always recommend that you draft your dashboard’s entire layout on a piece of paper first before moving to the real thing, which will help you project what you see in your mind. For some people, this is a really helpful first step before building.
2. Focus on data to insight and decision
Now that you understand your dashboard is a work of art, let’s add another layer: your dashboard is also a critical tool helping your management and your audience to take decisive action for your company. When selecting metrics, keep this in mind and it will guide your design process and overall selection.
3. Give your dashboard a title
Just like PowerPoint slides, a dashboard also needs a clear message. When people open your dashboard, you want to make sure that the first thing that meets their eye is the name or title of your dashboard. This is their first impression of your entire analytics journey, and it should be clear and concise.
And don’t forget that first impressions can happen across a variety of devices. A lot of your audience will be looking at your dashboard on a mobile device rather than a desktop, so make sure you’ve optimized the viewing experience for both. Whether viewed on an iPhone, a tablet, a laptop or a desktop computer, anyone who opens your dashboard should know immediately what that the title is and what the information they’re viewing is all about.
4. Include data visualization context
Each data visualization on your dashboard needs a title, description, point of view, and date and time stamp to provide viewers with context and understanding. This information presents a guide and a model for further KPIs.
5. Ensure load speed is close to a second
Information is useless if it can’t be understood. In the case of a dashboard, less is more and simplicity is key. This will also help performance.
Ideally, a dashboard should open in 1-2 seconds. A good rule of thumb is that it opens in less than one minute. If your dashboard takes more than one minute to load, there might be a problem on the data modelling side. One easy way to boost query speed is to transfer the underlying database to the Autonomous Data Warehouse.
6. Keep maintenance in mind
While creating your dashboard, make sure that you know exactly how long it will take to maintain it for future updates. Data quality is just as important as the data visualization itself. If a dashboard does not have 100% accurate data, you can’t deliver facts and management can’t make fact based decisions. So, after you acquire the data and create your work of art, be sure you have a plan for updating it, including refreshing the data, and transforming it and displaying it, for the long haul.
7. Don’t forget about logos and branding
There was a time, several years ago, when standards for dashboard design were low. Those days are past, and it’s important to use all of the options in the toolbox. Remember that logos, branding, and company communication need to be clearly and consistently presented. If your dashboard is being communicated inside of your organization, it’s always good for people to be able to identify it right away. Use your company logo and keep it prominent, and also keep in mind the basic color palette and design philosophy to which your branding is aligned.
8. With KPIs, less is more
It’s very tempting to include 1,000 different pieces of information on the dashboard, with 10 or 20 different filters. No one wants to leave out critical information. But remember that less is more. Too many KPIs create a classic case of “too much of a good thing,” forcing the user into information overload and clogging up the dashboard so that data cannot be properly visualized. Overwhelming people with information can make them question your dashboard and the data, reducing your impact.
Instead, set five or six KPIs as your data visualization maximum. This doesn’t mean that you need to leave deeper information on the cutting room floor; it simply means that you should allow an option to drill down into these KPIs via sub KPIs.
9. Don’t forget about the footer
Dashboards need information context, which can be featured in the footer. The footer is the critical spot where you share the name of the dashboard’s author, and if appropriate, the data sources used.
The footer is also your tool spot for more context on viewer information. Is the dashboard restricted or confidential? Keep that information in the footer. It’s also the spot to put contact information like an email or phone number so that if any user has a question about the dashboard, they can reach out for clarification. These more than nice things to do, they are thoughtful touches that make your dashboard more usable, which your audience will appreciate.
10. Filter your filter use
Some people love filters. But filters need to be used sparingly, and with care. The more you offer filters to slice and dice data via categories like region, the more you open yourself up to the risk of the dashboard not being current or presenting incorrect data. Above all, make sure that your dashboard is clear and the data is understandable with all filters removed.
Putting It All Together
Here are the key takeaways:
- When you’re building a dashboard, take time before you begin to focus on your goals. Think to yourself, “When someone looks at my dashboard, what do I want them to see and what actions do I want them to take?”
- When you’re working, focus on delivering data that leads to insights and decisions.
- Keep simplicity and clarity in mind by providing a title, context, logos, branding, and a footer.
- Ensure that your load speed is close to a second and that your dashboard is easy to maintain.
- Be sure to choose five or six KPIs, and use filters sparingly.
Your dashboard doesn’t have to say everything and it doesn’t have to include everything. It’s a guide. Trust your judgment, remember less is more, and you’ll be on the right track before you even begin.
Author: Benjamin Arnulf