10 Principles for Building Effective Dashboards

Follow these guidelines to create a dashboard that resonates with your users and transforms your data into actionable information.

Therese Moriarty
Jan 21, 2019 · 8 min read
A Google Analytics dashboard | Photo by Lukas Blazek on Unsplash

A dashboard is the most visible part of a self-service analytics (SSA) suite; it’s the public-facing page where users can access and analyze their data. But many dashboards fail to captivate users and guide business decisions because they have poorly-designed displays. Even if a dashboard has a beautiful interface with attractive colors and data visualizations, it can still fall short of delivering any meaningful insights.

There are many dashboard templates floating around, however if you have unique tracking and reporting requirements, you’re better off developing your own format. If you learn a few basics about visual perception and managing data, you can build an effective dashboard that helps your users drive business growth.

Photo by Ben Sweet on Unsplash

Before you even build your dashboard, you have to engage with your prospective users during a planning phase. Gather details around which metrics they track, how much data they want to query, and how frequently they need to access new information. As with any product, people will only interact with a dashboard if it solves a problem for them — so understanding their needs and goals is a critical first step. This input from your users will inform the “blueprint” of your dashboard and promote a higher adoption rate.

Photo by Mika Baumeister on Unsplash

A dashboard is just one component of an SSA suite; it essentially provides a user-friendly method to retrieve information from a data source. So if a dashboard is slow or if it updates infrequently, the problem often lies with the data source.

Once you’ve explored your users’ needs, optimize the data source to pre-empt any performance issues. Make sure the dashboard platform can handle your data storage. Run a few queries to check the response time; if it takes a while to return results, you may have to re-organize the data. Optimizing the data source behind a dashboard is akin to laying the foundation for a house, a smarter configuration will lead to greater stability.

Photo by rawpixel on Unsplash

During your planning phase, you should ask your users about what type of screen they intend to view the dashboard on. Will it be desktop, mobile, tablet, or all of the above? If they need to access the dashboard across different devices, you will have to design it to be responsive. Alternatively, they may simply want to present the dashboard on a giant monitor in their office — in which case, you should forego interactive features like menu filters. Screens factor significantly into the dashboard design process and you should always account for them.

Abbreviated and full numbers of the total world population (Source: United Nations)

A dashboard should not be a detailed report, instead it should be a well-organized summary that allows users to quickly absorb information. Your performance monitoring needs to meet a specific objective and it should be as high-level as possible. Metrics like totals, averages, and rates always come in handy because they summarize whole data sets as single numbers.

And since people take more time to process long strings of digits, try to round and abbreviate your figures whenever you can. Unless they work in finance, your users probably don’t have to view precise numbers in order to get the information they need.

This eye tracking heat map shows a common dashboard scanning pattern (Source: Tableau)

If you keep the following UX criteria in mind, you can design a dashboard that effectively informs your users.

  • Most people start scanning pages from the upper left corner of the screen — this makes the area prime real estate for any metrics you want your users to focus on.
  • If people have to scroll to view more information, they tend to assume it’s less important than what they can immediately see. This is why you should design your dashboard to have a minimal page length. If you find yourself placing more metrics below the fold of a page, they may belong on another dashboard all together.
  • Miller’s law indicates that most people can only store around 7 items within their short-term memory. With respect to dashboards, this means you that you should avoid exposing your users to more than 9 menu filters. Anything more than that can overwhelm them, potentially leading to overchoice or analysis paralysis.
Here are 2 dashboards that visualize the same information. The spare use of color on the left is less likely to confuse users than the busy palette on the right. (Source: Information Dashboard Design: The Effective Visual Communication of Data by Stephen Few)

When it comes to dashboard design, keep everything consistent and simple. It’s OK to apply the same data visualization more than once, all that matters is that it works for the information you need to convey. The differences in color should also be purposeful, otherwise they will unnecessarily grab your users’ attention and prompt them to search for meaning that isn’t there.

Visual consistency will habituate your users, helping them consume information and spot changes more readily. Always resist the urge to tantalize them with an array of data visualizations and colors. If you do, you’ll probably disengage your users more than you will impress them.

Comparing serif and sans-serif fonts (serifs are circled in red)

Your users won’t be able to assimilate any information if they don’t know what they’re looking at, so it’s imperative to label everything on your dashboard. Give your metrics clear names and title every data visualization. Provide legends explaining the significance of any colors or symbols in your diagrams.

On top of being descriptive and coherent, every label has to be legible. This is why you should always use a sans-serif font in your dashboard. Unlike serif fonts, sans-serif fonts do not have any embellished edges around their characters. This makes them more readable on digital screens, especially ones with lower resolutions.

Goal lines and comparison values can put your metrics in perspective

When you add context to a dashboard, it can rapidly communicate if your activity is trending positively or negatively. For instance, a dashboard can display a quarter-to-date sales revenue of $64.2K — yet what does that really signify to a user? The number alone doesn’t indicate anything about how well their business is performing. How close are they to hitting their target? How does it compare to their performance during a previous period? Goal lines and comparison values are some elements that can offer valuable perspective and address these types of questions.

Photo by Agence Olloweb on Unsplash

“Garbage in, garbage out” (or GIGO as it’s commonly abbreviated) refers to the concept that an erroneous input will produce an erroneous output. It’s a computer science adage that’s highly relevant to dashboard development. Every user who views and interacts with a dashboard expects to see an accurate picture of what’s happening. But ultimately, a dashboard returns information from a data source…and if your data source is flawed, so are your dashboard’s results.

An error-prone dashboard can undermine user adoption rates and erode trust in your product. Therefore, you should routinely test your data source for accuracy. Not only during the planning phase, but for as long as people are using your dashboard. Since new bugs can emerge at any time, it’s crucial to check your activity regularly and nip any problems in the bud.

Photo by Vitto Sommella on Unsplash

Dashboard development is an iterative process. Much like testing your data source, you should continue engaging with your users after the planning phase. An ongoing dialogue will keep you informed about their evolving tracking needs and help you refine dashboard features. Before you know it, you may find yourself creating entirely new dashboards or connecting to different data sources!

While change is inevitable, there are a few things you can do to streamline your future development. Standard naming conventions and formulas might take a while to establish, however, they can become major timesavers down the road. At the very least, it’s always worth the extra effort to program code that’s clean and well-structured. Hasty coding is harder to debug and it’s more vulnerable to technical debt that will complicate your work later on.

Whether you’re building your dashboard from scratch or using a BI tool (like Tableau, Microsoft Power BI, Klipfolio, Google Data Studio, Grow, Looker, Yellowfin, or Sisense), you may run into some limitations that prevent you from applying all of these principles. Perhaps your selected dashboard platform does not support custom labeling or comparison values. As a result, your development may involve some trial and error; you may need time to figure out how to work within the boundaries of your dashboard platform and leverage every feature to its fullest extent.

All said, you will always have the freedom to curate your content as you see fit — so remember that minimalism is key. When you distill your dashboard down to its most essential information, your users can quickly update their awareness and work more productively. To quote the polymath intellectual Antoine de Saint-Exupéry, “Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.”


eyeful

We visualize data to help businesses make smarter decisions. Get in touch. 👇 https://eyeful.us

Therese Moriarty

Written by

I make pretty charts. 📈 https://eyeful.us

eyeful

eyeful

We visualize data to help businesses make smarter decisions. Get in touch. 👇 https://eyeful.us

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade