EN
The cover of How to Measure Anything on a wooden table beside a rolled measuring tape.

What How to Measure Anything can teach us about talent

The most productive contribution of How to Measure Anything is not the promise of a number for everything. It is the demand that we specify which uncertainty we want to reduce before we begin collecting data.

PRYSMAP6 min read

Douglas W. Hubbard wrote How to Measure Anything in response to a common organizational claim: this is important but intangible, so it cannot be measured. His answer is provocative. If something affects a decision, there is almost always some observation capable of reducing, even modestly, our uncertainty about it.

The Wiley description summarizes the book’s purpose as helping decision-makers become better informed and take less risk. Read through the lens of talent, the book does not offer a model for assessing people. It offers something that comes first: a discipline for framing measurement problems more effectively.

Measurement does not mean finding a final answer

Hubbard defines measurement as a set of observations that reduces uncertainty, with the result expressed quantitatively. The shift may seem small, but it changes what we expect from measurement. Measurement does not require eliminating every doubt or capturing a concept in full. It requires us to know more at the end than we did at the beginning.

In talent, this idea helps us move away from outsized questions such as what is this person’s true potential? That formulation combines time frame, role, context, motivation, learning, and opportunity. A single number is unlikely to resolve it.

A measurable question would be narrower: how uncertain are we about this person’s ability to make autonomous decisions in a broader role over the next six months? The formulation still requires definitions and evidence, but it already connects measurement to a decision and a period.

This is one of the book’s strengths: before looking for data, it forces us to examine the question.

Clarification is part of measurement

Hubbard devotes considerable attention to clarifying the problem. He asks exactly what the concept means, which decision depends on it, which outcomes matter, and which observations would change our position.

Applied to capabilities, saying strategic thinking is not enough. We would need to identify its manifestations relevant to the role: recognizing dependencies, working across different time horizons, comparing scenarios, anticipating secondary effects, or explaining the assumptions behind a decision. Not all of those manifestations need to be present or carry equal weight. Their relevance depends on the work.

Measurement begins when the concept stops functioning as a label and is translated into observable conditions. That translation is not neutral: it defines what will count as evidence and what will remain outside.

Decomposition reduces the sense of intangibility

Another recurring idea in the book is decomposition. Problems that appear impossible to measure often contain smaller variables that can be observed or estimated.

An organization asking whether it is ready to execute a strategy can separate the concern into critical roles, required capabilities, current coverage, dependency on a few people, development time, and options for mobility or hiring. No variable answers the original question by itself. Together, they help identify the uncertainty that truly affects the decision.

Metal ruler with a blue-handled magnifying glass resting on a frosted glass panel.

Decomposition also prevents us from measuring what is easy simply because it is available. Counting completed courses is straightforward. Estimating whether people apply a capability with greater autonomy is more demanding, but it may be much closer to the decision that matters.

Not every uncertainty deserves the same effort

Here we encounter one of the book’s most useful and least intuitive ideas: the value of information. If an additional observation has no reasonable chance of changing a decision, measuring it may cost more than it contributes.

An explanation published by INFORMS notes that calculating the expected value of information helps focus resources on the additional data that would be most useful. The principle changes the conversation about talent dashboards. The question stops being how many indicators we could add and becomes which uncertainty could change a relevant decision.

This also sets a limit. Not every difference requires greater precision. If two possible outcomes lead to the same development action, collecting more evidence may have no immediate value. If, by contrast, a succession decision depends on an uncertain gap that would be difficult to reverse, better observation may be crucial.

Experts must calibrate their uncertainty too

Decision Which choice needs better information.

Uncertainty What is not yet known with sufficient confidence.

Hubbard does not dismiss human judgment. He seeks to make it explicit and calibrated. The book proposes training probabilistic estimates so that assessors recognize their own overconfidence and express ranges instead of artificial certainties.

Applying this idea to talent assessment requires care. Assigning a percentage does not turn an opinion into evidence. It does remind assessors to distinguish between what was observed, what was inferred, and how confident they are in the conclusion.

A statement such as they are not ready claims a degree of finality that the evidence may not support. Saying there is consistent evidence of performance below the expectation in two contexts, but we still need to observe the person under greater pressure offers a more useful conclusion. Not because it sounds cautious, but because it identifies what is known and which observation is missing.

What the book does not resolve

The strength of How to Measure Anything can also invite overreach. Reducing uncertainty quantitatively does not decide what we should value, which consequences are fair, or who has the authority to define an expectation. Nor does it correct a poorly defined construct or an assessment in which people have unequal access to opportunities to demonstrate capability.

People change their behavior in response to metrics, contexts change, and a capability cannot always be observed directly. What we observe are situated decisions, outcomes, procedures, and explanations. Applying the book to talent requires adding validity, fairness, interpretability, and governance.

My reading is that Hubbard is most valuable before the dashboard, not inside it. He forces us to ask which decision we are trying to improve, which uncertainty limits it, and what the smallest useful observation might teach us. Only then does it make sense to discuss scales, instruments, or visualizations.

That order does not make talent measurement simple. It makes it less ceremonial.

  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.