Review: Noise (audiobook)

When I saw a new title co-authored by Daniel Kahneman, I knew it’s going to be a good, insightful piece. Like the “Thinking, Fast and Slow” was. Yes, it comes from a renowned brand, and yes, expectations are set high. Nobel-high. So, here is the “Noise” by Kahneman, Sibony and Sunstein. I picked the audiobook version for ease. Did it deliver?

By all means, this title is a must-have in 2021.

Oh, really? [surprised]

It’s Jam-packed with insights

To be clear;

For me, insightful – as it should – means it is both thought-provoking AND backed with strong research data. This comes as granted. No concept is left floating here. Everything is grounded in either academic or corporate research.

Now, what I found most compelling was the thorough analysis, following the exposure of volatility in countless decision making processes. Just sheer awareness of this puts you one step ahead in understanding certain group dynamics, like the group thinking. Then comes the differentiation between noise, and bias. Finally, how these two relate to each other, and what that implies for organisations and individuals alike.

The example of Google downsizing from 25 to just 4 interviews on average, really got me thinking.

Tackling the hiring bias

Truth be told, I knew little about the scale and impact of biases in hiring decisions. Interestingly, we’ve been recruiting heavily ourselves, and I recognised − unfortunately − some not-so-good patterns. Yes, this is crazy relevant these days. Too bad that I’m not in the position to address all of those on the spot.

Turns out, that we all need to be much more thoughtful about biases in hiring processes. In fact, we should stick to a well-researched process in the first place, defined as series of structured or semi-structured interviews for greater decision quality and minimal bias. But there are many more types of biases in play – from halo effect to prejudice.

And there’s noise in all this too. The example of Google downsizing from 25 to just 4 interviews on average, really got me thinking. You can read more on this here. That’s an extreme amount of input information (x25!) on one hand, but it’s impaired by plenty of noise though. On the other hand, running a single interview round is equal to a blind guess, with chances of a good decision in the range of 50%. Statistically, according to Google’s findings, 4 interviews lead to an optimal judgement and therefore, hiring quality.

De-noise your organisation

But there’s much more to this. No one would accept only 61% agreement rate in medical diagnosis between highly qualified professionals. But still, this is the case! Even seemingly obvious judgments like x-ray evaluations result in higher than expected delta of judgements.

Or take seasoned business executives affected by noise levels in the range of 34-70%. With a generally accepted rate falling between 5-10% of such decision making noise. This. Is. Crazy.

So, Should you read it too?

There are 2 take-aways here:

1. Go read this book. Or at least listen to it, like I did.
2. If you make decisions “of matter” in your organisation, run a noise audit exercise. Seriously.

/K.