How to Measure Anything: Finding the Value of "Intangibles" in Business by Douglas W. Hubbard

How to Measure Anything: Finding the Value of "Intangibles" in Business by Douglas W. Hubbard
My rating: 4 of 5 stars

A nice book. A must read for sceptics like me who think there are many immeasurables in business. The key message the author gives is, instead of taking a path or avoid taking path because one does not find the right measurement, an attempt should be made to find out what can be measured to reduce the risk if the path is taken or not taken. This will help make a more sensible decision than just saying there are immeasurables. In short some information is better than no information.

It is recommended that one of the tools be leveraged to carry out the measurement with whatever data is available.
1. Monte Carlo
2. Markov Chains
3. Bayesian Probability and Bayesian Inversion for Ranges
4. Rasch Model
5. Lens Model
6. Simple sampling. The key is that the samples must be really random.
7. Brunswik's method
8. Daves Z Scale
7. Objective Model if Historical Data is available

They myth that is dispelled in the book is that when you have a lot of uncertainity, you don't need too much data to reduce uncertainity significantly. Event a very small amount of relevant data will go a long way to reduce the uncertainity.

Some of the issues that must be avoided are
1. Bandwagon effect
2. Halo effect
3. Choice blindness
4. Don't over measure. The theory of Diminishing Marginal Returns starts applying and only adds to the cost reducing the risk significantly.

At a high level the steps outlined are
1. Define the decision and variables that matter to it
2. Model the current state of uncertainity about those variables
3. Compute the value of additional measurements
4. Measure the high-value uncertainities in a way that is economically justified
5. Make risk/return decision after the economically justified amount of uncertainity is reduced.

A must read for all decision makers, which is all of us.

View all my reviews