I’ve been intimately involved in the creation and dissemination of hundreds, if not thousands, of dashboards over the past decade. Four years ago, I even wrote my most viewed piece ever (not posted here, I put it on Medium for some reason): Value from Dashboards…the real story. That post has been read thousands of times showing some demand to understand dashboards.
Dashboards relay information quickly and efficiently. The classic example is the dashboard in your car: it needs to display critical information in a glance that lasts less than a few seconds. If the information is confusing or takes longer to absorb, a life-threatening situation arises. If the data isn’t available at all, a life-threatening situation arises. If the information is incorrect or misinterpreted, a life-threatening situation arises. A dashboard that requires explanation is not much of a dashboard.
The other day I was presented with a conundrum that made me revisit my fascination with dashboards. How do you build and maintain a valuable dashboard when there is limited desire to submit/maintain data and limited desire to act on the data? The need for it is because there is real risk that needs to be called out in case action is actually desired after all. Should one even build a dashboard in this circumstance?
I’m not a fan of doing work with the goal of simply covering my ass. My work should be able to stand on its own without needing additional work to cover myself. Building something for the sake of building it isn’t interesting to me.
However, a case study from a few years ago provides a counterpoint to this argument. I was in a situation where there was a significant disconnect between groups about the current state of some accounts. The situation had existed for well over a year and continued to grow worse. To help bring things together, a few of us created a report showing the gap and established a process for dissemination. For months there was no movement, but then one day we discovered the report being used in guerrilla fashion by others behind the scenes. Just because demand isn’t obvious, doesn’t mean demand doesn’t exist.
Reports are not the same things as dashboards. Dashboards are summaries. Reports are the full data set, dashboards are the top 10 list. The use cases for each are very different. However, it is not uncommon for people to confuse the two and turn their dashboard into a report. To provide a simple rule of thumb: if your summary contains sentences, it is a report and not a dashboard.
Simplicity is the key for dashboards. They should be simple to update, simple to read, simple to distribute. The numbers in them should be actionable without wondering what they mean. Dashboards may cause controversy but they should not be controversial in themselves. Your executive summary is the dashboard but there are likely to be reports behind each summary element.
When your check engine light comes on in your car the reason often isn’t obvious, only the action is: get your car checked out. The dashboard is built to drive action: speed up, slow down, get your engine checked, fill tires.
Business dashboards should be the same. A red light should mean that something needs to happen. Green lights do not imply no action needed, only that the current actions are appropriate. Amber lights should imply caution required because something isn’t going well. Any text accompanying these lights should be self-explanatory. Names listed on the page should be completely versed in why the status is what it is.
A dashboard that doesn’t drive action isn’t a dashboard, no matter how many other characteristics it has of a dashboard. A page full of green lights means a lot of people are patting themselves on the back but not much else is happening. Very rarely is that actually the case. A misleading dashboard is one of the worst business tools in existence and should be questioned out of existence.
That’s the nature of dashboards.