Making your organization more data-driven part 2
Generating excitement through design
This is part 2 of “Making your organization more data-driven”, my series on how to integrate data-driven decision making into your organization’s existing processes. You can find part 1 here.
Nobody wants to use unaesthetic tools. Anyone who remembers the early days of the internet will appreciate how far we’ve come in our approaches to digital design.
Your BI tools should be no different. Anyone can make a data table or a bar chart; it takes a strong eye for both data and design to assemble a beautiful dashboard that communicates a story. Taking the time to commit to user experience will help grab your users’ attention and keep their interest.
A lot has already been written on the subject of making beautiful dashboards, and I’ll link some of my favorites in the footer.
I don’t want to go into many specific examples here, but I do want to touch on some broader points to keep in mind as you carve out your BI portfolio & strategy.
Keep your dashboards focused.
When I think about the most egregiously useless dashboards I’ve ever seen, there is always one commonality: people try to fit too much onto one screen. Beyond the performance ramifications, creating cluttered dashboards is distracting and confusing for users.
If your tool is going to be a joy to use then you can’t make your users work for your/their insights. This means that their eyes should naturally move across the screen as they process your story. They shouldn’t have to scroll excessively. Marks needing comparison should be placed close together. Cramming too much disparate content into the same view works against these goals.
You would think this would be a well-known and accepted concept, but you’d be surprised at how many users or developers I meet that want to pack a thousand metrics into a single dashboard. I get it. At the end of the day, we’re problem solvers. Users have a myriad problem and we want to design a cure-all. This just doesn’t work in practice. You’re just going to overwhelm your user with intense complexity.
Focus on grouping metrics that go together and tell one cohesive story on the same page and then make other pages for other stories. Your individual dashboards can be simple! In fact, simple dashboards are often much more beautiful and less overwhelming for your users. You’re not here to prove how long it took you to work on your dashboard. Develop your ideas separately to present information clearly.
Furthermore, clustering multiple charts that share metric domains on their own dashboard will help you give your tools a specialty and a unified theme. As I discussed in part one, focusing on your audience is important. If you’re focusing on your users successfully, dashboards should be well grouped by function to solve a specific problem.
Adopt a design mentality.
Tackle your dashboards with a design-first mentality, especially for higher-level dashboards. After you’re sure the data is correct and the metrics are accurate, turn off your data analyst cybernetic eye and pull out your design lens.
Read some articles on web design. Go to a color palette website (this one is really cool) and try some out. I like to take away all the axes, excess lines and headers on my dashboards, then slowly add them back one by one. This allows you to be intentional with your design. Everything should exist for a reason, usually to add clarity. Your goal is to remove redundancy and noise so your story shines through.
Play with it, have fun with it, spend time on it! It’s probably rare for you as the analyst to have a chance to step back and take some artistic license.
I should also add that I prototype in black and white with mutable layouts, then spruce things up when we’re going to final reviews. This allows my users to discuss the flesh and bones of the tool while being undistracted by design choices (I warn you now: everybody has an opinion on design, and almost all of them suck). If you don’t get the content correct then the design doesn’t matter.
But it also doesn’t matter how great your analysis is if your dashboard isn’t a joy to look at. The balancing act of content and design is crucial.
Don’t go overboard on interaction.
Your goal is to answer a business problem, it’s not to answer every business problem.
I know that often your users will want to have every single possible filter to cross section their data in every way. You need to steer them away and try to answer the relevant questions through your visualizations or through different dashboards.
Very often, people don’t understand that enabling every cross section of the data possible in a single tool just makes it feel cumbersome. Even if they gain the ability to look at any slice they want, taking ten clicks to get there will be frustrating.
Keep the detail manageable. You will probably be the only person in the room during initial requirements gathering that understands the UX impact of the choices being made. You don’t want your users to have to navigate a maze of clicks every time they open your BI tool — you want to answer their question and facilitate their quick solving of a problem.
As long as they know where to go to get their answer, it’ll be ok to have everything split more into separate dashboards. It’s fairly common for me to create a dashboard that views specific slices of the data that people consume often, and then have a separate tool that allows users to download any cross section of the underlying data they want. Your goal is to provide a light bulb for a single story, not to create one tool to solve them all.
Hopefully these points help you get in the right frame of mind to make your dashboards more beautiful. I think this part of the process can be the most frustrating to get “correct” — design is a pretty subjective concept. That being said, there are people and companies that consistently get it right so there are definitely right and wrong principles at work.
Stay tuned for part three of four articles in this series!
I’m Eric, a freelance business intelligence engineer and co-founder at Magnifai. I’m interested in building analytic tools paired with beautiful design, especially for opportunities utilizing data for activism. I write about data analytics and productivity & motivation on Medium. Feel free to reach out or connect via LinkedIn.
