Dashboard Design Lessons from the Pit Wall

Good dashboard design is mostly an exercise in restraint, and few environments teach restraint better than a motor racing pit wall. Engineers sit in front of screens fed by sensors on the car and decisions about fuel, tyres and strategy must be made in the time it takes a car to complete a lap. Nobody there wants a dashboard that looks impressive. They want one that makes the next decision obvious. Here are five habits worth borrowing for business screens.

Start with the decision, not the data

Teams do not display every channel they record. They display what supports a choice that is about to be made. When we are asked to build a dashboard, our first question is not which numbers exist but which questions the reader asks on a Monday morning. How many jobs are late? Which customers are waiting for a reply? Is cash coming in as expected? Each answer belongs on screen, ordered by how often it is needed. Anything that does not inform a decision can live one click deeper.

Show change and context, not just numbers

A lap time of 1 minute 32 seconds means nothing without the previous lap, the target and the conditions. The same is true of a sales figure. Show whether a figure is up or down against last week, against a target and against the usual range for the time of year. Small line charts, called sparklines, sit neatly beside a number and reveal the trend in a glance. Avoid decorative gauges and three-dimensional charts; they consume space and make comparison harder.

Use colour as a signal, sparingly

If everything is red, nothing is. Reserve strong colours for conditions that need action, and make normal states quiet. Remember that many readers cannot distinguish red from green, so add words, shapes or positions to carry the meaning. In a pit environment a flashing warning is understood immediately because it is rare. A dashboard that raises ten alerts a day teaches people to ignore all of them.

Decide how fresh the data must be

Real-time is expensive and not always necessary. Telemetry from a car streams constantly because decisions depend on seconds. A weekly management report does not. Ask how stale a number can be before it misleads, and design the pipeline accordingly. Show the time of the last update on screen, so nobody makes a decision from yesterday's figures without knowing it.

Build trust in the numbers

A dashboard is only used if people believe it. Define each metric in plain words, show where the data comes from and make it possible to click from a summary figure down to the rows behind it. When two screens disagree, find out why and fix it, rather than letting people drift back to private spreadsheets. Data quality at the source, such as consistent dates and spelling of customer names, matters more than any chart library.

Design for the real screen

Pit walls use large monitors; most business users look at a laptop, a tablet on a workshop bench or a phone between meetings. Test at each size, keep text legible and make touch targets large enough for fingers. Check accessibility too: sufficient contrast, labels for charts and a table alternative for anyone who cannot see the graphic.

Review and prune

Alerts should arrive with a next step

On a pit wall a warning without an instruction is an annoyance. The same applies to business screens. When a metric crosses a line, such as a stock level dropping under a threshold or a support queue growing beyond an hour, the alert should say what it means and what the usual response is, and it should link directly to the place where that response can be carried out. Keep a record of when alerts were raised and who acknowledged them, so the process improves over time instead of relying on one conscientious person.

Resist the urge to add one more chart

Every request for an extra figure feels reasonable on its own, and a year later the screen resembles a cockpit. Make a rule that adding something means removing something else, and agree who decides. Where people want more detail, give them a separate report rather than crowding the main view. The best dashboards we have built are the ones whose owners are slightly embarrassed by how little is on them.

Dashboards accumulate clutter. Every quarter, ask which tiles anyone has actually looked at, and remove the rest. If you want a fresh pair of eyes on a screen that has grown unwieldy, or you are replacing a set of spreadsheets with something live, we are glad to help you work out what deserves a place on it.