On Humility and Making Better Decisions

Dr. Daniel Kahneman features on the latest Farnam Street podcast and it’s a surprising episode. Kahneman wrote Thinking Fast and Slow. I admire Kahneman a great deal. Not for his Nobel or for his work, which are both impressive, but for his humility.

Some of the key tenets of Kahneman’s work in his famous book were disproved. And he owned up to it, both in print and on the podcast. That’s the hallmark of someone with great integrity, and it’s a sign to trust someone more.

So I listened to his podcast with great interest. Kahneman discussed the ways to make better decisions. He recommends delaying your intuition and breaking down a decision process into smaller pieces.

Suppose you are interviewing executives for a VP role. Ahead of time, the hiring team creates a rubric detailing a few key dimensions, maybe five or six. Cultural fit, intelligence, domain expertise, ability to recruit and management style, hypothetically. The hiring committee interviews the candidate and then reassembles to debrief.

Most of the time, we run unstructured meetings. Kahneman recommends structuring the agenda by dimension. Everyone discusses the first parameter and scores her on cultural fit. Once that’s done, the group moves onto intelligence. Everyone scores the candidate, and onto the next one. At the end, the group reviews the score matrix and decides.

In this way, everyone delays their intuition and thinks more deeply about the decision. This is a structured interview, which is three times better at predicting outcomes than an unstructured interview.

Another way of breaking down decision is the Fermi Process, named for Enrico Fermi, another Nobel laureate, father of the atomic bomb. Fermi was a terrific estimator. He famously estimated the power of the atomic bomb to within an order of magnitude by dropping leaves of paper during the detonation of a test bomb.

The Fermi Process or Fermization, as Tetlock called it in a great book called Superforecasters, is about breaking down a problem into smaller parts. Then estimate each of the subcomponents and then calculate the answer. How many piano tuners exist in Chicago? Can you estimate it? The answer is at the bottom of this post.

This idea of breaking down predictions and forecasts into smaller bits has merit and has been proven across several different fields. It’s worth considering as a tool the next time you need to decide on a complex topic. As Kahneman says, the best decisions are made by algorithms and the closer we can get to structured decisions, the more we can supress bias.

The podcast has several other interesting jaunts on different topics in work psychology. But the thing that I admire most about the author is his integrity. It’s rare to see a man of such prominence admit fallibility. It’s a trait to admire and emulate.

*Chicago has about 290 piano tuners. See this explanation