Over the past few weeks I’ve been reading Decisive by Chip Heath and Dan Heath.  For those of you who haven’t read the book, it’s essentially a collection of vignettes about different people who’ve had to make hard decisions accompanied by suggested strategies for making those decisions.  For example: one story was about a man who had a rare disease and had to decide whether or not to have a risky operation, one was about VP at a non-profit who had to decide which vendor to use for a specific technology, one was about Andy Grove at Intel and his decision to kill the company’s line of memory chips and go full-force into processors, one was about a recent college graduate deciding on whether to go to nursing school, etc, etc.

The book has – at times – been entertaining to read, but throughout there has been one big problem that’s really been nagging me.

To properly explain, first I need to tell a story.

In a past professional role, my primary job was analyzing the effectiveness of optimization algorithm changes.  In essence, I helped my team decide if we should make a specific algorithm change, or not.  During my time in that role I quickly learned an important lesson about data and making decisions.  I found that if we were able to set up a perfectly controlled experiment with one single independent variable (something we would change) and one single dependent variable (a success metric) – as long as we executed the test properly – the decision was easy to make.  We would control the modification of the independent variable and observe the change on the dependent variable; if we saw the desired results, we’d make the change.  However, if we did an experiment with two independent variables and two dependent variables, the problem wasn’t simply two times harder, it was actually ten times harder.  If we had more than two variables on either side of the equation, the problem became nearly impossible.  With more variables, problems get exponentially harder to analyze.

In Decisive, Dan and Chip Health explain real life problems as if they are controlled experiments.  Maybe in hindsight it’s possible to boil a problem down to its core independent and dependent variables, but life is really complicated.  There are almost always many more variables to take into account than can be analyzed scientifically – and often you don’t even know which variables matter until after you’ve already made the decision.

Don’t get me wrong, the tactics they present for solving problems are totally legitimate and helpful frameworks to remember, but the problems they present are over-simplified and manipulated so that they fit perfectly into their various decision making frameworks.  It’s not the solutions they present that are phony; it’s the problems.

If you’re into the Heath brothers and want an entertaining read, this might be worth checking out.  Personally I think you’re better off reading Daniel Kahneman’s work, which is the source material for many of the Heath brother’s decision-making strategies anyway.

Decisive: How to Make Better Choices in Life and Work
  • I do love those interactions between independent variables. Makes life fun!

    Did you ever read The Drunkard’s Walk by Leonard Mlodinow? I find the concept of randomness to be quite unintuitive, but rather powerful.

    And: I’m not sure the world will ever get tired of post hoc ergo propter hoc and its many variations.