How to Measure Anything

About the book

How to Measure Anything addresses a common workplace standoff: a decision hinges on something nobody knows how to quantify, so the discussion stalls or defaults to gut feeling. Hubbard, a consultant in applied information economics, frames measurement as reducing uncertainty rather than achieving perfect precision. The book sets out the reasoning behind that reframing and the methods for putting it into practice.

Its central tools are calibrated probability estimates from informed people, small samples that still carry information, and the recognition that a measurement only needs to be precise enough to change a decision. Hubbard connects these ideas to valuation and to intangible assets, two areas where organizations often claim numbers are impossible. The treatment is practical and argumentative, aimed at readers who want a defensible way to put bounds on uncertainty.

How to read it

Approach the book as a methods argument with worked examples rather than a narrative. The early chapters establish why the intuition that some things cannot be measured is usually mistaken, and later material turns to calibration exercises, sampling, and ways to model uncertain variables. Readers who want to apply the ideas should plan to try the calibration and sampling exercises rather than read passively.

The prose is plain and example-driven, but the statistical content rewards attention. Some sections assume comfort with basic probability and expected value; others can be skimmed on a first pass and returned to when a specific problem calls for them. Keep the decision context in mind, since the methods are framed around choosing whether more information is worth gathering.

Who it suits

Managers, analysts, and consultants who face decisions involving hard-to-quantify factors will find a structured alternative to guessing. It suits readers who want practical estimating techniques with justification for their assumptions, and those responsible for valuing intangibles or assessing risk. Readers with a statistical background may move quickly through the fundamentals and focus on the applied examples.

Reading considerations

The book's confidence in calibrated estimation depends on honest participation and reasonable data, and its examples reflect a consulting context that may not match every organization. Later editions and related works expand the methods, so check which version you have before relying on specific techniques. Test the central claim against a decision where measurement has previously failed.