library / lib3722b8841142bbf2
How to Measure Anything: Finding the Value of 'Intangibles in Business'
Douglas W. Hubbard
In a sentence
A practical guide arguing that anything a manager cares about—however 'intangible'—can be measured by reframing measurement as the economically justified reduction of uncertainty to inform decisions.
Douglas Hubbard dismantles the costly myth that important business quantities like quality, risk, security, employee morale, or public image are immeasurable. Drawing on his Applied Information Economics (AIE) method and inspired by 'measurement mentors' Eratosthenes, Enrico Fermi, and nine-year-old Emily Rosa, he shows that measurement means reducing uncertainty—not achieving exact certainty—and that even a few clever observations can dramatically improve big, risky decisions. The book equips readers with calibrated estimation, Monte Carlo risk modeling, value-of-information calculations, sampling shortcuts (the Rule of Five, the Urn of Mystery), Bayesian updating, and methods to turn human experts into reliable instruments (Lens models, Rasch models). The central revelation—the 'Measurement Inversion'—is that organizations routinely measure what is easy and ignore what truly matters, and that computing the economic value of information tells you exactly what to measure, how much, and when to stop. Through real cases (Veterans Affairs IT security, EPA drinking water, Marine Corps fuel forecasting, ACORD standards), the book proves that resourceful, decision-focused measurement pays for itself many times over.
The four lenses
- Science
- Statistics
- Systems
- Strategy
Tags
The model
A causal path model in which clarifying a decision and the conditions for measurement drive psychological states (reduced uncertainty, calibrated judgment) and behaviors (selecting and applying measurements) that improve decision quality and economic outcomes.
Decision and Object Claritydesign lever
The degree to which the decision to be supported and the thing being measured are unambiguously defined in terms of observable consequences and decision impact.
Value of Informationdesign lever
The computed economic worth of reducing uncertainty about a variable, determined by the chance of being wrong and the cost of being wrong relative to a decision threshold.
Calibrated Judgmentpsychological state
The trained ability of estimators to assign probabilities and confidence intervals that match real outcome frequencies, correcting systematic over- or underconfidence in subjective estimates.
Uncertainty Reductionpsychological state
The quantitatively expressed decrease in the range or probability spread about an unknown quantity resulting from observation, which is the operational definition of measurement itself.
Measurement Method Applicationbehavioral pattern
The behavioral selection and execution of appropriate empirical instruments—decomposition, sampling, experiments, Bayesian updating, expert models—matched to the information value and uncertainty.
Cognitive and Measurement Biascontextual condition
Systematic distortions—overconfidence, anchoring, halo/horns, bandwagon, expectancy, selection, observer biases—that degrade unaided human judgment and observations.
Decision Qualityoutcome metric
The degree to which decisions are well-informed, with quantified risk and return, leading to fewer costly errors and better allocation of resources.
Economic Return on Decisionsoutcome metric
The monetary value realized from making better decisions under reduced uncertainty, including cost savings, avoided losses, and improved returns.
How they connect
- decision clarity → predicts information value
- information value → influences measurement method selection
- calibration skill → predicts uncertainty reduction
- measurement method selection → predicts uncertainty reduction
- uncertainty reduction → predicts decision quality
- decision quality → predicts economic outcome
- cognitive bias − moderates calibration skill
- cognitive bias − moderates decision quality
- decision clarity → mediates measurement method selection
The story
The reader A manager, analyst, or decision maker who must make big, risky decisions and wants better information about things others have dismissed as immeasurable.
External problem
Critical 'intangibles'—quality, risk, security, customer satisfaction, productivity, brand value—appear impossible to measure, so decisions are made under unnecessary uncertainty.
Internal problem
They feel stuck, intimidated by statistics, and anxious that any measurement will be too imperfect, too expensive, or simply impossible.
Philosophical problem
It is wrong to allow self-imposed ignorance to drive resource misallocation when reducing uncertainty is both possible and economically valuable.
The plan
- Define the decision the measurement supports and clarify exactly what the intangible means in observable terms.
- Quantify your current uncertainty with calibrated 90% confidence intervals and probabilities.
- Model the decision and its risk (e.g., with a Monte Carlo simulation).
- Compute the value of additional information to decide what and how much to measure.
- Apply economical measurement methods—sampling, experiments, Bayesian updating, expert calibration.
- Make the risk/return decision once uncertainty is reduced to a justified level, then iterate.
Success
- Better-informed decisions with quantified risk and return.
- Resources allocated to what truly matters, saving money and avoiding costly errors.
- Confidence that even 'intangible' factors can be measured economically and iteratively.
- A repeatable, scientific approach to decisions across the organization.
At stake
- Continued reliance on unaided intuition and overconfident estimates.
- Measuring the wrong things while ignoring high-value uncertainties.
- Misallocated resources, rejected good ideas, accepted bad ideas, and wasted money.
- Decisions made under avoidable ignorance that put profits, safety, or welfare at risk.
Chapter by chapter
ch07Quantifying the Value of Information
ch12The Human Mind as a Measurement Instrument
ch14Applied Information Economics
Questions this book answers
- What does it actually mean to 'measure' something, and why does that definition make almost everything measurable?
- How do you quantify your current uncertainty about an unknown quantity?
- How much is it economically worth to reduce uncertainty before making a decision?
- Which variables in a decision should actually be measured, and how much effort is justified?
- What practical methods (sampling, experiments, Bayesian updating, expert calibration) reduce uncertainty cheaply?
Glossary
- Decision and Object Clarity
- The extent to which the decision being supported and the quantity to be measured are clearly and observably defined, including alternatives, thresholds, and consequences.
- Value of Information
- The economic worth of reducing uncertainty about a variable for a given decision, based on the probability and cost of being wrong relative to a threshold.
- Calibrated Judgment
- The trained ability to assign subjective probabilities and confidence intervals that match actual outcome frequencies.
- Uncertainty Reduction
- A quantitatively expressed decrease in uncertainty about an unknown quantity resulting from observation; the operational definition of measurement.
- Measurement Method Application
- The selection and execution of empirical instruments appropriate to the information value and the nature of the variable.
- Cognitive and Measurement Bias
- Systematic distortions in human judgment and observation that degrade estimate accuracy and decision quality.
- Decision Quality
- The degree to which decisions are well-informed with quantified risk and return, yielding fewer costly errors.
- Economic Return on Decisions
- The monetary value realized from improved decisions made under reduced uncertainty.
Related in the library