In marketing, we usually know our metrics are imperfect.
ROAS is directional. Attribution is contested. Brand lift can be useful, but no one should mistake it for the voice of God. Even the cleanest dashboard is still a partial view of reality, filtered through assumptions, definitions, time windows, channel constraints, and whatever the analytics team could reasonably stitch together before the next executive readout.
Yet inside organizations, a strange thing happens when the subject shifts from customers to employees.
Suddenly, measurement becomes much more sacred.
Engagement scores, performance ratings, personality tests, leadership surveys, “high-potential” assessments, and manager evaluations are often treated as if they reveal something definitive. A number appears in a spreadsheet, a quadrant, or a dashboard, and the organization begins behaving as though the measurement is the person.
That is where measurement becomes management theater.
The survey already knows what it wants to hear
A few years ago, I saw a leadership survey designed to evaluate direct managers and inform executive understanding of leadership quality. On the surface, it looked like a standard internal assessment: employees would rate their managers, results would be aggregated, and leadership would receive a readout.
But the design was doing more work than the questions admitted.
The scale skewed positive. There was no neutral option. The wording made unfavorable evaluations harder to give and easier to interpret as outliers. Before anyone answered the survey, the instrument had already tilted the room toward a more flattering picture of management.
That may have been accidental. Plenty of internal surveys are built by people with good intentions and limited measurement training.
But in organizations, design choices are rarely neutral in their effects. A scale that makes criticism harder does not simply measure sentiment. It shapes sentiment into a more convenient form.
And once the results are presented upward, the theater begins.
The organization can say it listened. Executives can say they have data. Managers can say the numbers validate their leadership. The process appears empirical, even if the underlying instrument was designed in a way that protected the conclusion.
Bad measurement does not become good because it is quantified
This is the central mistake: organizations often confuse quantification with rigor.
A weak survey with a five-point scale is still weak. A personality inventory with poor scientific grounding does not become more valid because it has a polished report. A performance score does not become objective because it appears next to a decimal point.
Numbers can clarify reality.
They can also decorate bias.
That is especially true in talent management, where assessments are often used to make consequential judgments about hiring, promotion, leadership potential, culture fit, and performance. The language may sound scientific — reliability, validity, benchmarks, profiles, competencies — but the actual practice may be much thinner.
Was the assessment validated for this job?
Does the scale capture meaningful differences?
Are we measuring performance, popularity, compliance, confidence, or proximity to power?
Would the result be stable if measured again?
Is the tool identifying leadership potential, or merely rewarding people who already resemble the organization’s preferred image of a leader?
These are not academic questions. They determine who gets promoted, who gets coached, who gets labeled difficult, and who gets quietly moved out of the pipeline.
Measurement often protects the existing order
This is why management measurement deserves more skepticism than it usually receives.
Assessments do not only describe organizations. They can help maintain them.
A leadership survey can become a legitimacy machine. A performance rating can justify decisions already made informally. A personality test can turn cultural preference into diagnostic language. A “high-potential” label can make sponsorship look like merit.
The organization gets to say, “This is what the data shows.”
But sometimes the data is only showing the values, assumptions, and power arrangements baked into the measurement system.
That does not mean all measurement is corrupt. It means measurement is never innocent. Someone decides what to count, how to count it, how to label the result, and what consequences should follow.
Those choices matter.
The problem with pseudo-scientific confidence
The most obvious examples are personality tools that organizations continue using long after serious doubts have been raised about their scientific value.
They persist because they are easy to administer, easy to discuss, and emotionally satisfying. They give people a vocabulary for themselves and their colleagues. They create the feeling of insight.
But a feeling of insight is not the same as evidence.
This is where management culture often reveals an uncomfortable preference: it may want the appearance of scientific legitimacy more than the discipline of actual scientific thinking.
Real measurement is messy. It requires humility. It forces leaders to ask whether the tool is reliable, whether the construct is valid, whether the criterion is meaningful, whether the interpretation is warranted, and whether the results should actually be used for the decision at hand.
Management theater skips most of that.
It moves directly from instrument to insight to action.
What better measurement would look like
Better measurement starts with a more modest premise:
A metric is not the truth. It is a structured way of asking a question.
That means leaders should be much more careful about the questions their tools are really asking.
A good employee survey should not be designed to reassure leadership. It should be designed to surface reality clearly enough that leadership can act.
A good performance system should not reduce a person to a single annual label. It should help distinguish outcomes from conditions, effort from opportunity, individual contribution from team context, and short-term output from long-term value.
A good leadership assessment should not merely identify who looks executive. It should examine whether someone creates trust, improves decisions, develops others, handles dissent, and strengthens the system around them.
And above all, good measurement should preserve judgment rather than replace it.
The real test
The real test of a measurement system is not whether it produces a number.
It is whether the number helps the organization see more clearly.
If a metric makes leaders more curious, more honest, and more precise, it is probably useful.
If it makes them more certain, more defensive, or more comfortable with what they already believed, it may be theater.
Organizations do not become more scientific by putting weak assumptions into dashboards.
They become more scientific by caring whether the dashboard reflects reality.
That is harder work. It is also the work that separates measurement from performance.
