Sexy Dashboards
I loooooove a great data dashboard. I’ve looked at particularly well-designed dashboards and said, without irony, “That’s sexy.”
It’s perhaps because of this interest that I’m usually underwhelmed when seeing dashboards. The majority of dashboards I’ve seen in organizations are poorly designed or used in a suboptimal way. The aspiration is good—figuring out how to synthesize the most important data and drive better decisions—but the intended outcome is often not achieved.
Following is what I’ve seen make these tools more effective.
Principle 1: Great dashboards answer specific, meaningful questions.
Too often, leaders think that creating a dashboard means simply putting all the information in one place. But that’s not a dashboard; that’s a report.
A great dashboard doesn’t just provide data. It answers specific questions.
Your monthly bank statement is an example of this distinction. The bank statement reports how much money is in each of your accounts.
However, the most meaningful questions about your finances might be: Do I have enough in the checking account I use for paying bills to cover next month’s bills? Or, Is my rainy day savings account enough to cover six month’s worth of expenses?
Because the bank statement provides information rather than answer the most important questions, it’s a report rather than a dashboard.
In organizations, a view that provides just the revenue and expense data is a report. A view that answers Are we on track to reach our revenue goal? or How much of our expense reduction targets have been realized to date? is likely a dashboard.
The first step, then, in creating a great dashboard is clarity in what questions you want it to answer for you. As you can see from the analogies above, the questions themselves dictate what data is most meaningful.
In terms of design, I’m actually a fan of putting the questions directly in the dashboard for clarity.
For example, in creating a health dashboard, I might put “change in weight over the past month” as one of the measures. This is probably better than just reporting the current weight.
But if I showed it to someone else without context, they might ask: Are you trying to lose weight or gain weight? How much? By when? For what reason?
Putting the meaningful question on the dashboard itself eliminates confusion.
This may seem trivial. It’s easy to think that everyone will know what you’re talking about. But the reminder is helpful for your boss, or your board, who don’t think about your work everyday. The reminder is even helpful for your team, as they can sometimes lose track of the overarching goals.
The final reason to use meaningful questions as the foundation of dashboards—and not the data—is that the most important questions may be best answered with qualitative judgments. For example: How effective is our organization? How robust is our new product pipeline?
When you start with data as the framing, the dashboard can end up being biased toward those things we can measure today, and leave out important elements of the strategy.
Even having blank spaces for “we can’t measure this yet” is a helpful thing to see!
Principle 2: Great dashboards distinguish between strategy and monitoring.
To understand the distinction, think about the difference between the fuel or battery gauge on your car’s dashboard and the navigation map.
The fuel gauge is definitely important, but all you really want to know is that it’s above a certain level. It goes up and down, but you’re not actively trying to make it be at a certain level.
The map, on the other hand, describes where you want to go, and it provides information relative to how well you are progressing toward that destination.
The best dashboards I’ve seen separate strategy metrics from monitoring metrics. This is part of clarifying the meaningful questions.
For these metrics, we’re asking “Are we making progress?”
For these other metrics, we’re just looking for, “Is there a problem that we don’t know about?”
Principle 3: Great dashboards should be management tools, not just reporting tools.
Put another way, great dashboards should be what leaders use day-to-day to run the organization. This is the ultimate test of relevance.
For example, imagine asking someone, “How are we doing on customer retention?” If they cannot immediately show you on their day-to-day dashboard, or don’t have the data memorized, here’s what is likely the case:
You’re not on the same page about how important customer retention is, or you have a different logic about how the business works.
The person is not actively managing customer retention, even though you believe that it’s strategically important.
There’s an inability to measure customer retention effectively.
In all cases, this is evidence of a lack of strategic alignment and should trigger a helpful two-way conversation about the work.
In practice, many dashboards are only created because the board or senior leader requests a synthesized view of what’s going on. In those cases, a helpful question when a dashboard is presented to you is: Which of this data do you look at day-to-day, and which is just generated for this meeting?
There’s another, super practical reason to ensure that the dashboard is actually a management tool: If the data in the dashboard is not generated in the normal flow of work, it becomes a pain to update. That makes it more likely that the dashboard will be lower quality and the effort to create it will be less sustainable.
Principle 4: Great dashboards should spark the right conversations.
Imagine getting a report card from school and never talking with your parents about it.
This was actually my experience. Because I was …well, a nerd who got all As, my dad would never ask any questions about my report card. It took about 30 seconds for him to review it, and we moved on. But that’s not a great practice!
To achieve their highest and best use, dashboards should exist to support an effective strategy conversation.
Step one in achieving this outcome is the principle from above—orienting the dashboard to answer meaningful questions. When the dashboard reflects the strategy, it can be a tool for strategic discussion.
The second step is to design the discussions around the dashboard to be most effective. Tactically, the agenda item shouldn’t be “review the dashboard.” Rather, it should be “review the business, using the dashboard to support that conversation.”
In the most effective routines, there is a clear expectation that the person presenting the dashboard—the person responsible for moving the data on it—comes to the meeting with:
A clear perspective on what is most important
The next-level analysis already complete (e.g., a deeper dive on those things that are off-track or surprising)
A proposal for what they’ll do next
Requests for the group (e.g., questions they need help answering, tactical support)
That’s what good looks like.
This is also why it’s important that a dashboard serves as a management tool and not just a reporting tool. If it reflects how they’re doing the job day-to-day, the leader is much more likely to feel ownership over the results and be thinking about how to move forward.
Principle 5: Great dashboards maximize the signal-to-noise ratio.
A good example of this is the Check Engine light on your car’s dashboard. It’s great because it’s just one symbol, and you know what to do when it’s on.
Now imagine that instead of one symbol, the Check Engine light was replaced by a readout of every single metric that’s used to calculate whether there’s a problem in the engine. It’d be overwhelming, confusing, and distracting.
This is poor signal-to-noise.
The takeaway: Dashboards should be powered by data, but they shouldn’t necessarily have all the data.
When the Check Engine light is on, that’s when you can go deeper and show why that’s happening. Otherwise, the details aren’t useful.
The book Universal Principles of Design provides some helpful tips on how to increase the signal-to-noise ratio of your dashboard:
”Maximizing signal means clearly communicating information with minimal degradation. Signal degradation occurs when information is presented inefficiently: unclear writing, inappropriate graphs, or ambiguous icons and labels.”
And: “Minimizing noise means removing unnecessary elements, and minimizing the expression of necessary elements. It is important to realize that every unnecessary data item, graphic, line, or symbol steals attention away from relevant elements.”
Here’s a great test of whether the signal-to-noise ratio is right: Can you see the most important takeaways at a glance?
If not, there’s room to improve.
Your dashboard might not reach “sexy” level, but if they meet each of these principles, I’d bet that they’ll be more useful!