The dashboard nobody opens

A team we worked with had built the dashboard everyone in the room had asked for. Forty charts, every metric someone had named across a string of planning meetings, a clean layout that looked genuinely impressive projected on the wall. It shipped to nods and a round of thanks.

Then almost nobody opened it again. The demo turned out to be the high point of its life, not the start of a habit. Weeks later, decisions that the dashboard was supposed to inform were still being made off memory and gut feel, the same as before it existed.

The instinct when a dashboard goes unused is to add to it: more charts, a cleaner design, the one KPI a stakeholder mentioned after the launch that nobody had thought to include. That instinct treats the problem as a coverage gap, as though the right dashboard simply contains more than the wrong one did.

It doesn't. Adding a forty-first chart to a dashboard nobody opens produces a dashboard with forty-one charts nobody opens. The unit that was missing was never information; it was a reason to come back on an ordinary day when no one was watching.

A dashboard admired in a review and a dashboard used on a Tuesday morning are different objects wearing the same skin. One exists to demonstrate that the work behind it was thorough. The other exists to change what someone does next, today, without an audience. Building more of the first has never once produced the second, no matter how complete or well-designed it gets.

What actually moved it was a single question, put to each stakeholder separately: what would you do differently if this specific number changed? Most of the forty charts had no answer. People could describe what the number measured, but not what action followed from it moving up or down, which meant the chart was decoration wearing the shape of information.

A handful of charts had a real answer, and it came fast, specific, and identical across people who hadn't compared notes. Those were the ones already tied to a decision someone made on a recurring basis, whether or not the dashboard existed to support it. We kept the three that passed that test and retired the other thirty-seven.

It felt like a step backward at first: less coverage, fewer metrics, a dashboard visibly thinner than the one it replaced. It was the opposite. Coverage was never the thing making the original valuable, and thinning it down did not remove value, it removed the noise that had been hiding whether there was any. One number someone acts on beats forty they admire, and the dashboard that survived proves it: smaller, plainer, opened on a normal Tuesday instead of only in front of an audience, doing the one job a dashboard is actually for.