Introduction

Every organization we've worked with can produce a dashboard. Fewer can tell you why the numbers on it are the ones that matter. Somewhere between the strategic plan and the reporting deadline, metrics tend to accumulate rather than get chosen — a KPI added because a donor asked for it, another because a competitor tracks it, a third because someone in a meeting once said "we should probably measure that."The result is a dashboard that's technically full and practically useless. It answers questions no one is asking, while the question that actually determines whether the mission is working — is this working, and how do we know — goes unmeasured.A well-chosen metric is not a bigger version of a bad one. It's a different kind of object entirely.

‍A metric earns its place on a dashboard by doing three things at once. It has to be tied to a decision — if a number changing wouldn't change what you do next, it's trivia, not a KPI. It has to be attributable — if ten variables move it and you control one, it will tell you almost nothing about your own performance. And it has to survive contact with a skeptical reader: a funder, a board, a partner who wasn't in the room when the target was set.Most metrics fail at least one of these tests. Few organizations notice, because a metric that fails quietly still produces a number, and a number on a page looks like progress whether or not it means anything.This is why logical framework development is not a paperwork exercise. Done properly, it's the process of stress-testing every indicator against the theory of change it's supposed to prove — forcing the question of what does this actually demonstrate before the indicator ever reaches a report.

The cost of the wrong metric

Choosing badly doesn't just waste a column on a spreadsheet. It actively misleads. A program that hits an easy target looks successful while the outcome that mattered goes untracked. A team optimizes for what's measured, not for what's meant — and if what's measured is the wrong thing, the organization gets very good at the wrong thing.We've seen this most clearly in monitoring and evaluation work for donor-funded programs, where the gap between an ambitious mission and a defensible indicator set is often the entire difference between a proposal that gets funded and one that doesn't. Funders don't fund good intentions. They fund organizations that can demonstrate, in specific and attributable terms, what changed because of the work.

The same discipline applies well beyond the nonprofit and donor-funded world. A startup tracking vanity engagement numbers instead of retention, a consultancy reporting hours billed instead of outcomes delivered — the failure mode is identical. The dashboard is full. The insight is empty.

Choosing well, in practice

Getting this right is less about finding a perfect indicator and more about running every candidate metric through a short set of honest questions before it earns a place on the dashboard:If this number moves, does it change a decision we'd actually make?Can we attribute the movement to something we did, or could ten other factors explain it just as easily?Would this indicator still look credible to someone who doesn't already want to believe us?Is there a simpler, more direct way to measure the same underlying change?Most metrics that survive this filter are less impressive-looking than the ones that don't. They're narrower, harder to hit, and less flattering in a slide deck. That's usually a sign they're doing their job.

Conclusion

The instinct to measure more is understandable — more data feels like more rigor. But rigor doesn't come from volume. It comes from the discipline of choosing a small number of indicators that are actually tied to the outcome you care about, and being willing to defend each one under scrutiny.That discipline is the difference between a logical framework and a spreadsheet, between a KPI dashboard and a wall of numbers, between an organization that can prove its impact and one that can only describe it. The weight of a well-chosen metric isn't in what it says. It's in what it makes true, and what it forces you to stop pretending.