As an Tableau analyst, I feel like I answer this question quite often. And when I ask it, I get the perfect textbook answer:
A dashboard is a visual display of the most important information needed to achieve one or more objectives; consolidated and arranged on a single screen so the information can be monitored at a glance.
– Stephen Few
Perhaps because it comes from the textbook…And, I’ll spend time breaking it down because it really is a lot to swallow in one chunk. And, let’s be real, we memorize things like this, spout it off, and rarely pick it apart.
So what does this mean?
I’ll let you in on a little secret: humans are visual creatures. Even people born blind build a limited amount of spatial awareness. Charty types with fancy words call this “pre-attentive attributes” or “Gestalt Principles of Perception“. This too is in that textbook. For those who skipped buying books in school, this is what you need to know – anything “pre-attentive” or “perceived” means you don’t need to be literate in the field to get that piece.
So, when we look at charts, there is a limited, baseline understanding we all have. Some bars bigger than others – we get this. We all want the king-size candy bar as kids. Pie charts that look like Pacman – we know which piece we’d rather have if we’re talking Lemon Meringue. It’s when we don’t, that the food fight starts. We wonder how to visually communicate using charts when the obvious stops being glaring.
So, for this piece, there’s big business in chart-choosers. Or, you can take a statistics class. Or Six Sigma. Or analytics courses. Pick your poison, but if you’re not fluent in charts, this is the place to start.
This is one of the scarier parts of dashboard design. The Stephen Few definition calls this “the most important information.” If you ask your boss what the most important information is, you’ll be told everything. Don’t ask an executive for KPIs.
The Hitchhikers Guide already gave us the perfect answer to life, the universe and everything. We don’t need the answer. We need questions. Lots of them. If you’ve never worked as peon somewhere, you should. You’ll see all the problems workers face. Now imagine that experience in data. What would you want shown? How would you measure the times customers spit at you or threw things at you (or maybe this was just me?). Quantify that so it’s meaningful (occurrences per day? deviation from expectations? gross factor index?).
A lot of people want templates or reports that can do it all. Notice – report. Not dashboard. Dashboards are concise and focused for a reason.
Reports tell us about things. Dashboards tell us to do things, like change the oil in our car, fix the airbag, or get gas before we’re pushing the car or walking miles in the snow to the next gas station. The dashboards we make for business are not that different. We make them in the hopes that things will change, ideally for the better.
We want people to find things that support their work and be empowered to find the cause, or at least an area where to look.
We’re not bringing back ’80’s music (okay, maybe we are), but we are trying to digest our findings so others don’t have to do the same. Like bees and honey, we’re spitting this stuff back out so others enjoy it without all the work.
We’re simplifying and trying to make it easier for others to either come to our same findings, or discover something else equally interesting. The scary part about this one? We own it and shape it.
Remember the relevant piece and deciding what’s important? This is finding the “what else” without going to far down the rabbit trail. Knowing when to stop is just as hard as knowing what to get in the first place. Maybe harder…
This is where I push people towards the last piece.
Dashboards are just one view of the data. Often, there’s a few of these in a workbook. You may start with one piece for the first dashboard, explore a middle question for the second, and end on some detail for the 3rd. Or, you may have a few questions and choose to answer each one with a dashboard. There’s a million different ways to split these up and people smarter than me provide better guides. Bust out your Rubik’s cube and try getting all the colors to match. If that fails, pull off the stickers and try again.
Now, Mr. Few wrote his definition without interactivity in mind. As you look at the dashboard, consider where actions belong. While Tableau does make it easy for everything to filter everything with actions, I usually try to order mine in a sequence and avoid monster filter fights.
So what else are we missing?
We communicate for a number of reasons. So do our dashboards. Ever try to follow a presenter that makes too many points? (Don’t look this way.) Gives too many details? Provides too many options? (Hey, I saw you look!) Okay, now look at your dashboard and ask the same questions.
A dashboard that communicates effectively lets you know something. It informs, it passes down history, sometimes, it even shares culture. How does a dashboard share culture? What’s the focus? Is it good news, bad news, or a blend of both? What stories are consistent? What will get passed on?
Happy dashboard making. And maybe I’ll see you at the Tableau Conference. Thankfully, I’m co-presenting with the delightful and sensible Joshua Milligan.