Nobody Looks at Your Dashboard. It Was Built Backwards.

Fourteen charts and six weeks of dust

Someone built it. Maybe you, on a weekend, with real enthusiasm. It's genuinely nice — revenue trend, pipeline by stage, project status, a couple of gauges. For about ten days everyone looked at it. Then a busy week happened, and nobody opened it, and nothing broke. That last part is the tell.

Now it sits there, a small monument to a good intention, and every so often you feel a flicker of guilt about not checking it. Don't. The guilt is misplaced — you're treating a design failure as a discipline failure. If the fix were discipline, it would have worked, because you are demonstrably one of the most disciplined people you know. You run a company. You didn't fail to open the dashboard because you're lazy.

Pull fails on the week you need it

The mechanism is direction. A dashboard is pull — the information sits still and waits for you to come to it. Which means the whole thing is gated on you remembering it exists and having the spare attention to go look. But attention is the scarcest resource in your company, and it's the exact resource the dashboard was built to conserve. You built a tool to save your attention that can only function by spending it first.

Then notice when it fails. Not randomly. It fails on your busiest week — the week the calls stack up, the week something is on fire — which is definitionally the week most likely to contain the thing you needed to see. A detection system that switches itself off exactly when risk is highest is worse than no detection system, because you believe you have coverage.

And it fails silently. A page nobody opens generates no error, sends no alert, files no complaint. It just quietly stops being true — a chart pointed at a field that got renamed keeps rendering a shape. So there's no moment where the business learns it's blind. It just is, indefinitely, behind a bookmark.

The pillar's version is short: telemetry is not a dashboard. A dashboard is a place you have to remember to go, which makes it one more thing depending on your attention — the exact problem you were trying to solve. When you are the system, you can't grow beyond yourself, and that applies to your eyes as much as your hands.

The second failure: no decisions attached

There's a compounding reason nobody opens it, separate from the direction, and it's why fourteen is the wrong number. It answers questions you never asked. Somebody built what the tool made easy to build — the charts the integration offered — rather than what a person was going to act on. So even when you do open it, nothing happens. You look, you nod, you close it. It's an aquarium.

Run the test on every chart: if this number moved, would we do something different? Fourteen charts, and honestly maybe two survive. The other twelve aren't neutral. They're camouflage. They dilute the two that matter, so on the day one of the real signals moves, it's a small line on a busy page nobody's reading.

That's what makes this expensive rather than merely wasteful. It isn't that the dashboard doesn't help. It's that it convinces you visibility is a solved problem. You've got a dashboard, so you don't build the thing that would have told you. The unbuilt telemetry is the actual cost, and the dashboard is what's blocking it.

Push the pulse, schedule the scoreboard

The rebuild splits the thing you built into two, because you were trying to make one artifact do two incompatible jobs. The live layer — the pulse — has one non-negotiable property: it comes to you. Push, never pull. A stalled deal, a silent client, an onboarding stuck on step two lands in Slack with the record attached, addressed to a named person, whether or not anyone remembered anything. Three or four signals, no more. A channel that fires forty times a day gets muted within a week, and then you're blind again with worse confidence.

The pattern layer — the scoreboard — doesn't need to be live and shouldn't be. It needs a slot and an owner. The numbers get computed the same way every time and read out loud at a standing review by a human whose job includes reading them. That's what makes a scoreboard work: not availability, attendance. Nobody ever needed the trend at 11pm on a Tuesday. They needed it said out loud on the first Monday with the team in the room.

Which means the artifact might survive — a Notion page, a simple view, whatever's cheap. It just stops being the mechanism and becomes the reference for a meeting that actually happens. The meeting is the system. The page is a prop.

This is the Telemetry pillar

OPERATE names this symptom exactly: the dashboard nobody opens. Someone built it. It has fourteen charts. It answers questions you never asked, so it has been closed for six weeks and nothing broke.

It's worth being generous to whoever built it, including you. The instinct was right — the business is opaque and it shouldn't be. What was missing is the direction and the restraint, and neither is obvious. Every dashboard tool on earth sells you pull, and every one makes fourteen charts easier than two. You built what the tool made easy, which is what everyone does.

The goal isn't to watch everything — it's to know the right things to watch. If you have a dashboard with dust on it, the useful question isn't how to make people open it. It's which three facts your business already knows and cannot say out loud to the person who needed to hear them. That's the shape of the pulse, and it's a fraction of the work of the thing you already built.