The Undoing of Recruiting
I picked up Michael Lewis’ 2017 New York Times Bestseller, The Undoing Project because I loved Liar’s Poker, The Blind Side, and Moneyball. In Moneyball (2003), Lewis told of how Billy Beane (Brad Pitt in the movie) reinvented the cash-strapped Oakland Athletics by using data analytics to outsmart the richer teams. The Undoing Project is a kind of sequel to Moneyball. Lewis felt that in Moneyball he had only addressed one side of the equation: the data science. The other side: how do people make decisions?
Lewis starts with Daryl Morey, General Manager of the NBA Houston Rockets. Morey was a nerd who had never played basketball in his life. He had, however, developed a complex set of algorithms that were able to predict, with astonishing success, which draft prospects would become the best NBA players. His predictions were consistently better than the seasoned scouts and executives. The sports world was shocked that a major team had appointed a data geek to run their franchise.
As a recruiter, I was hooked. If the Houston Rockets could use data analytics to predict success, then recruiters should be able to do it too. But the bigger question was, why do the seasoned execs and scouts get it so wrong so often? We see this often in our business – companies passing on our seemingly bulls-eye candidates, then hiring someone who (based on their own requirements) we never would have even considered. I was stunned by the answer: Daryl had analyzed the execs mistakes and found that they were not random blunders, but that there was a consistent and predictable pattern of human error.
Morey was not the first with this theory. Lewis tells the story Danny Kahneman - another young data nerd who was recruited in the 1940’s by his employer, the Israeli Army - to predict which young men would make the best soldiers. He gave thousands of tests to tens of thousands of recruits, with unprecedented prediction accuracy. Danny then met another brilliant psychologist, Amos Tversky, and together they wrote hundreds of articles for obscure psychology journals read by only a few other geeks. Danny and Amos, two scholars, psychologists, and Israeli soldiers, eventually won the Nobel Prize for their theory of the mind. The implications for marketing, healthcare, and economics were game-changing, and even gave birth to behavioral science.
But from this headhunter’s eyes, Lewis’ book is all about recruiting – or more precisely – the decision of whom to hire. It’s arguably the most important decision a company makes. In our last newsletter (http://bit.ly/2wjgkUi), we wrote of a new world where applicants are chosen by algorithms and interviewed by bots. There’s no doubt that better data helps us make better decisions. But like Lewis after Moneyball, we’ve only addressed one side of the equation. The other side is: how are hiring decisions made?
Kahneman and Tversky attacked this problem with a set of mental test-tricks; paradoxes that would fool even the master statisticians. Then they would break down the decision-making mistake to isolate and identify the errors, which they called heuristics - mental rules-of-thumb we each use when faced with a decision – and put them into categories like familiarity, confirmation, and anchoring. Heuristics are mental biases that sometimes cause us to ignore data and make bad decisions. The premise wasn’t that the human mind is flawed (we all know that), but that it’s flawed in a systematic, consistent, and predictable way. They even tested thousands of doctors and proved that data analytics can predict the odds of success for cancer treatment options better than the doctors could themselves.
Danny and Amos had a knack for asking what hadn’t been asked, measuring what hadn’t been measured, and then dissecting the results.
Lewis called his book The Undoing Project because the cast of characters ‘undid’ a lot of the accepted beliefs about how we think and make decisions.
The more I read, the more I thought about how many aspects of recruiting need to be ‘undone’. For example, your HR department probably gives you statistics on how your company’s hiring is going (giving you stats like time-to-fill, interview-to-hire ratios, quality-of-hires, and so on). But do they give you stats on the bad decisions you’ve made by not hiring someone? Interviews ten years later with the superstars who you passed on? Or do you get stats on the people who didn’t respond to your ad? That’s what Danny and Amos would have done – and did - incomprehensible volumes of interviews and research for over three decades.
The Undoing of Recruiting is simply a suggestion to think about undoing a lot of what we think about recruiting. What heuristics and mental biases do each of us bring to the decision process? If hiring talent is indeed the most important decision companies make, how can we help them improve the odds of success, and reduce the risk of costly errors?
To that end, we’ve rolled out two new recruiting products this year. SourceSheet™, which supplies clients with real-time raw candidate source data, and ScoreSheet™, a tool to forecast candidate success for a particular role. We’re trying to address both sides of the equation. We may never be able to predict, with total certainty, the probabilities of success and failure. But there’s no doubt that we can help our clients improve the odds, and over the long haul, that means helping them build a better team.