EN
Needle gauge connected by tubes to a wood-and-metal base.

The illusion of precision in talent dashboards

A dashboard can display 73.4 and still fail to explain what was observed, how the evidence was combined, or how much the conclusion could change. Visual exactness does not guarantee conceptual precision.

PRYSMAP4 min read

Decimals, rankings, and small variations carry an air of authority. They imply that the phenomenon has been captured in enough detail to compare people, rank priorities, or anticipate risks. Sometimes that is true. At other times, the design presents a degree of resolution that the measurement process never achieved.

The illusion does not necessarily arise from false data. It may appear when sound data are interpreted beyond what they can support.

The figure inherits every preceding assumption

Before the result reached the dashboard, someone defined the skill, wrote the levels, selected sources, decided how to combine them, and determined how to handle missing or contradictory responses. The final result compresses those decisions.

If strategic thinking means something different across areas, an average with two decimal places does not repair the inconsistency. If one leader observed a person for a year and another for only six weeks, placing both results on the same scale does not make their access to evidence equivalent. If a self-assessment and an interview answer different questions, averaging them may hide more than it integrates.

The dashboard does not create these problems. Its visual cleanliness can make them less visible.

Screen resolution and evidence resolution

In metrology, uncertainty is part of the result. The NIST guide on measurement uncertainty establishes principles for evaluating and expressing the uncertainty associated with a measurement. A later NIST guide describes it as the uncertainty that remains about the true value after measurement.

Talent assessment is not physical metrology and should not pretend to be. The principle is nevertheless transferable: reporting a value without making relevant uncertainty visible invites overinterpretation.

A 3.42 may come from a scale with only four broad anchors, three observations, and meaningful disagreement among sources. The decimal exists because a formula can produce it, not because the evidence can defensibly distinguish between 3.42 and 3.38.

Displaying fewer digits does not solve the problem by itself. The important question is which differences are meaningful. If two results would activate exactly the same interpretation and action, their visual distance may be operational noise.

Aggregation erases the path

Every dashboard needs to condense. The risk appears when condensation removes the dimensions that would change the decision.

An overall indicator can hide:

A metal caliper measures the thickness of a gray felt piece on a wooden table.
  • a critical capability below expectation, offset by less relevant ones;
  • sources that disagree because they observed different contexts;
  • evidence concentrated in simple situations;
  • a stable score built on an unstable definition;
  • missing data treated as a neutral value.

An average shows what the combined result is. It does not answer why, with what confidence, or under which conditions. When the decision concerns development, mobility, or succession, those questions usually matter more than the decimal.

Design also decides what appears important

Ranking people from highest to lowest suggests that the main difference lies between individuals. Showing a rising line suggests progress even if the instrument, population, or opportunities to observe have changed. Using green, yellow, and red turns debatable thresholds into categories that appear natural.

These choices are not neutral. They direct attention and can produce consequences: prioritizing someone, rejecting a candidate, assigning training, or declaring a risk.

Depending on its intended use, a responsible dashboard should make it possible to trace:

  • the definition and expectation assessed;
  • the sources and observation period;
  • relevant dispersion or disagreement;
  • calculation rules and treatment of missing data;
  • limits of interpretation;
  • the minimum change that actually alters a decision.

Not everything needs to appear at once. Synthesis can be preserved if these layers remain traceable and the primary view avoids presenting irrelevant differences as findings.

Useful precision is a property of the decision

Definition What the indicator represents exactly.

Evidence Which sources and time period it covers, and how it treats missing data.

Decision Which difference would actually change an action.

A dashboard does not need to display the system’s full complexity. It needs to show enough information for the decision it is intended to support and allow closer examination when the conclusion carries consequences.

The question is not how many decimal places we can calculate, but what the smallest difference is that we can interpret and defend with the available evidence. Sometimes it will be a quantifiable difference. Sometimes only broad categories can be distinguished. In other cases, the correct answer will be insufficient evidence.

Acknowledging that limit does not make the dashboard less sophisticated. It makes it more honest about the relationship between what it displays and what we actually know.

  5. The cover of Sorting Things Out on a wooden table beside a small paper box.

    Sorting Things Out: every taxonomy describes and decides

    The book Sorting Things Out examines how categories and standards organize the world while distributing visibility, work, and consequences. Reading it through the lens of a skills architecture extends that inquiry to the decisions a taxonomy makes possible.