The most dangerous part of the hiring process is what you think you know, but which might not be true. I just finished reading Michael Lewis’s Moneyball, a best seller from a few years back that EVERY business manager should read. The book is officially about baseball, but if you can’t relate it to your own business, it may be time for you to retire.
Baseball has more statistics than almost any part of your business, except your sales department. When I was growing up, my best friend never read a book, but he studied the Dodger statistics every day. While most companies were managing on no more numbers than you’d find in a quarterly financial statement, baseball fans were devouring detailed daily statistics.
Surprisingly, though, baseball management was ignoring most statistics. As Moneyball explains, baseball managers and scouts had traditional ways to looking at players, and pretty much ignored the hard data available.
Billy Beane, General Manager of the Oakland A’s, hired a quantitative analyst who pored over the research of many amateur baseball analysts. Then he went digging in the data himself. The result was player arbitrage. Take the expensive players that other teams wanted, replace them with less expensive players who had the right stats. Billy Beane won more games than average, with fewer dollars of salary than average. Anywhere else you’d call it efficiency.
So do you just “know” how to spot a good customer service representative, just like baseball’s old scouts just “knew” how to spot a standout minor leaguer? Any chance that you’re judgment would stand up to statistical scrutiny?
Beane (actually his quant guy) also learned that some of the traditional statistics don’t do the best job of describing values. They ended up ignoring fielders’ errors. Instead of batting averages and RBIs, they focused on less common measures, such as on-base percentage and slugging percentage (a weighted average of bases per hit).
The business lesson here: what you think you know may not be true. The metrics you are using to justify what you are doing need to be subjected to tests and validation.
This is a great book to read after, or even before, Competing on Analytics, which I reviewed earlier. If it motivates you to rethink your hiring strategy, take a look at this post: The 4 Steps to Using Metrics in Hiring Decisions .
I suggest the following standard:
Go over in close detail what the resume claims vs. what the person actually knows. This is more difficult than it sounds. It means that if the fellow says he was in Pittsburgh in 1997, you find out what was going on in that city at that time and ask him questions about it. If he says he worked in a particle physics lab, you learn what you can about hadrons and quarks, then quiz him on it.
This approach has an important result. One is less likely to hire a liar. Liars are the most destructive new hires one can have.
You clearly do understand the problems involved in metrics. Many managers do not and do not have the patience and independent thinking skills necessary to evaluate trial quantities.
Credit scores are a classic: all things being equal, they will be worse for younger people. Does one want to screen out people on that basis? They may be lower for women who have gone through a divorce. Does one want to screen for that? They will be lower for people who have had medical problems. Employers might want to screen those people out, but it's a mistake. An employee who has had cancer and been cured is often an exceptionally loyal and dedicated employee. Lower credit scores can hit veterans, who have been hauled off to Iraq. On and on, one can come up with reasons why this could be a very, very bad metric.
But screening against liars helps all legitimate businesses.
Posted by: Charles Utwater | June 19, 2007 at 10:31 AM