“Dashboards create visibility”
A dashboard shows what was measured. Visibility needs a question and an action.
Published September 2026
Usually said
“Dashboards create visibility”
The shift
From: Show the data
To: Show the data a decision needs
More precisely
A dashboard makes a decision visible when each figure is tied to a named question and action, and it shows only the part of the situation that was measured.
In short
A dashboard puts shared numbers in front of people, which beats arguing from anecdote. It can only show what was measured, though, and it tends to show what was easy to measure. Once people are judged by a number, they work on the number. Visibility comes from tying each figure to a decision.
The argument
From the claim to a better one
The strongest formulation
Decisions made without data rely on anecdote and the loudest voice. A dashboard puts the same numbers in front of everyone, quickly, and makes problems harder to ignore.
Why it is attractive
It is a visible deliverable that looks like oversight. It turns a messy situation into something that fits on a screen.
Where it holds
It holds when each figure answers a question someone has already agreed to act on, and the people acting can see its uncertainty.
Where it breaks
It breaks when it shows what was easy to log rather than what matters. What is off the screen becomes invisible, and once a figure is watched people work to move it rather than the thing it stood for.
What evidence matters
Goodhart’s law, and Campbell (1979) on how a quantitative indicator used for decisions becomes corrupted and corrupts the process it monitors. The base rate matters for any alert. What matters is whether decisions changed, and whether they were better, after the dashboard existed.
So, more precisely
A dashboard makes a decision visible when each figure is tied to a named question and action, and it shows only the part of the situation that was measured.
Reading it
The precise version, word by word
A dashboard makes a decision visible3 when each figure is tied to a named question and action1, and it shows only the part of the situation that was measured2.
Tied to a named question and action
Each figure is there because someone has agreed what they will do if it moves. A figure with no action attached is decoration.
Only the part of the situation that was measured
What is off the screen still exists. A good dashboard says what it leaves out.
Makes a decision visible
The test of a dashboard is whether decisions changed, and got better. A clearer picture is the means.
What it changes in practice
- Start from the decisions people make, then choose the figures. Starting from whatever happens to be logged gets it backwards.
- Show uncertainty and base rates next to any alert.
- Watch for figures that improve while the thing they stand for does not.
- Retire figures nobody acts on.
A case, in general terms
A course team builds a dashboard of logins and video views. Within a term, staff are nudging students to log in, and logins rise. Pass rates do not move. A second version starts from one question: which students are at risk of failing the first assignment? It shows three figures, each tied to a specific outreach step. It is less impressive to look at and far more useful.
Where next
Follow this idea
Operationalise it
Learning analytics (opens in a new tab)Guide
Operationalise it
Data strategy and governance (opens in a new tab)Guide
Related argument
The analyst insideTool
Challenge it
The usefulness of an educational metric decreases when the metric becomes a substitute for the construct it was meant to approximate.Working proposition