Risk Insider: Beaumont Vance

No One Wants To Believe the Data

By: | December 18, 2015

Beaumont Vance is an executive risk manager. He has managed strategic risk for Fortune 100 firms for the past 15 years. His multidisciplinary approach weaves together such disparate fields as quant modeling, statistics, behavioral economics, biology and game theory into practical solutions and insights.

I recently had the awful of experience of getting data.

I know, I know, we are supposed to see data as a good thing, especially when it provides insights.

But this data refuted more than 30 years of experience and proved me wrong about many things. It was a ghastly experience.

The shock was delivered by a new smart shirt that I bought from a company called Athos. It is one of those workout gadgets that monitors your body and transmits the data to your phone. This particular shirt is packed with sensors that record how much each muscle is firing and displays it in real time on the phone.

I have lifted weights on and off for about 34 years. I take this seriously and have read many books on the subject.  I consider myself an expert. The shirt thinks otherwise.

big data photoI found out that many exercises that I thought worked out muscle A, actually work out muscles M and Z.  It is as if you found out that exercising your right leg actually worked the muscles on the left side of your neck.

But the results were incontrovertible. Sensors don’t lie. So what did I conclude at first when I got this data? That the sensors were lying, of course!

It took me about three weeks of emailing the shirt company to finally believe what I was seeing. And that is when I realized why Big Data efforts are doomed.

If the data contradicts our cherished beliefs, we will ignore it.

For example, imagine you work at a firm that has a strong service culture. The firm believes that customer service is the main key to its success. Then one day you run the data and find out that, in reality, the customer service efforts have a negative effect on the business.

There is no substitute for doing what actually, demonstrably, measurably works versus what one just “believes” to work.

What is management most likely to do?

I sincerely doubt that management will collectively say, “Well, we had that all wrong! Slash the service budget, retrain the employees and forget the whole customer service thing.”

Nope, they are much more likely to say, “That can’t be right. Let’s spend more on customer service and less on the unreliable analytics stuff.”

As Kahneman, Tversky, Ariely and other behavioral economists demonstrated, our cognitive filters effectively alter reality for us.

Thanks to my new shirt, I experienced this firsthand.

Unfortunately, if I am like other people, my experience does not bode well for Big Data nirvana.

However, for those who accept the radical change implied by analysis, there is a whole new world opening up. After I came to accept that I had to do everything totally differently in the gym, my results improved drastically.

There is no substitute for doing what actually, demonstrably, measurably works, versus what one just “believes” to work.

You can argue with reality all you want; you just won’t ever win.

More from Risk & Insurance