BOOK REVIEW · MEASUREMENT
How to Measure Anything
In short: Hubbard’s thesis is liberating: if something matters, it is observable; if observable, it is measurable. The book gives you the tools — calibration, decomposition, value of information — to measure things marketers usually wave off as intangible.
What it covers.
Hubbard reframes measurement as uncertainty reduction, then provides practical methods: calibrated estimation, decomposition, small samples, and the economic value of information.
- Measurement as uncertainty reduction
- Calibrated probability estimates
- The value of information
- Decomposition of hard quantities
- Small-sample inference
- Measuring “intangibles”
Who it’s for.
Anyone who has said “you can’t measure that” about brand, awareness, or sentiment. A mindset shift for measurement leads and skeptics alike. Finance and ops readers love it too; the methods are domain-agnostic by design.
Evergreen
- Measurement = reducing uncertainty
- Value-of-information thinking
- Calibration training
- Decomposition method
Read with a 2026 eye
- Examples skew enterprise/risk
- Not marketing-specific
Key ideas worth stealing.
“If it matters, it is measurable” dismantles the excuse that brand or awareness can’t be quantified.
Value-of-information analysis tells you which measurements are even worth taking — a rare and valuable discipline.
How it reads.
Methodical and a little dry, but genuinely mind-expanding. The calibration exercises stick with you.
The RGM verdict.
Not a marketing book, but one of the most useful measurement books a marketer can read — it kills the “it’s intangible” excuse. Pair with RGM’s MMM guide for the marketing application. Once you internalize that uncertainty reduction is the goal, “we can’t measure that” stops being an acceptable answer in any meeting.
The marketing application is yours to make; the thinking transfers directly to MMM and brand measurement. Hubbard’s clarification that even a few observations can dramatically cut uncertainty is genuinely freeing for marketers used to demanding perfect data before acting.
Asked & answered.
What is the main idea of How to Measure Anything?
That anything which matters can be measured, because if it matters it has observable effects — and measurement is really about reducing uncertainty, not achieving precision.
Is How to Measure Anything about marketing?
No, it is general, but its methods apply directly to the “intangibles” marketers struggle with — brand, awareness, sentiment.
Who should read it?
Measurement leads, analysts, and anyone who needs to quantify something others call unmeasurable.