Kill the dashboard. Keep three numbers.

Somewhere in your company there is a dashboard with forty tiles. It updates in real time. It is screenshotted into a monthly deck. And it has not changed a single decision in six months.

Dashboards grow because adding a metric is free and removing one feels like an accusation. Nobody has to defend a tile’s existence, so the tiles accumulate like furniture in an attic. The result is not visibility. It is a place where accountability goes soft, because with forty numbers something is always up.

Measurement is for decisions

A metric earns its place by having a decision attached. Not a sentiment, a decision: if this number does this, we do that. Cost per qualified lead crosses a threshold, we pause the channel. Activation drops for two weeks, we stop the feature work and fix onboarding. No attached decision, no tile.

Run that rule over your current dashboard and be honest. Most teams keep three to five numbers. That is not a failure of ambition. That is what steering actually requires.

Choosing the three

The exact numbers vary by business, but they always answer the same three questions.

• Are we attracting the right people, at a cost we can live with?
• Do those people take the action that predicts revenue?
• Do they stay, expand, or come back?

One number per question. Everything else is diagnostics: useful when a top number moves and you need to find out why, invisible the rest of the time. Diagnostics belong in a drawer, not on the wall.

The weekly readout

Three numbers make a different kind of meeting possible. Fifteen minutes, every week, same format: what moved, why we think it moved, what we change before next week. No slides. If a number did nothing, you say so and move on. The discipline feels almost rude at first, which is how you know it is working.

Teams resist this because a big dashboard feels like control. But control is not watching many numbers. Control is knowing which few numbers would make you act, and actually acting when they move. Everything else is scenery.