The Obstacle
The Boston Red Sox have changed the way they draft and deploy personnel. Instead of relying on gut instincts, the team now wants managers to make decisions based on statistical analysis. The transition hasn’t always been smooth. In Game 7 of the American League Championship Series against the New York Yankees last fall, Sox manager Grady Little decided to leave ace hurler Pedro Martinez in to pitch the eighth inning with Boston ahead 5-2. Martinez was obviously tired, having thrown 115 pitches, and the stats over the season showed his effectiveness fell by more than half after 90 pitches. The rest is now baseball lore: The Yankees rallied and won and, a week later, Little lost his job.
So how do organizations get managers in charge of human resources to make decisions from numbers instead of instinct?
The Response
Give managers something they can really use: One of the keys to getting managers to use statistical analysis for human-resources decisions is finding practical metrics that can give a team new ways to achieve success. Over the past few years, Bill James, now an advisor with the Sox, and other baseball statisticians took a hard look at the existing baseball metrics to see if there were other—perhaps better—ways to evaluate players. They came up with some interesting results. For instance, instead of looking at just a hitter’s batting average, James combined hitting with walks and the total bases achieved during an at-bat to come up with a more accurate reflection of a player’s ability to get on base and contribute a run—thereby giving his team a better chance to win.
Show them proof: Even with new and improved metrics, long-held attitudes are difficult to change. Yet many of the most stubborn managers will alter their opinions about personnel when presented with documented successes. That was the case at the Georgia Office of Child Support Enforcement. When senior human-resources advisor Charles Brooks, after poring over reams of data, discovered that some enforcement agents were simply better than others at collecting money from deadbeat dads, other managers resisted using his metrics. But Brooks developed a profile of the characteristics and skills the agency’s top performers possessed and used that information to hire new agents with similar profiles. The average amount collected per agent jumped by 42%, and turnover was reduced by 50%. It was hard for anyone in the organization to argue with that kind of success.
Show them more proof: Just like anything else, the more you prove something works, the more acceptable it becomes. Oakland A’s general manager Billy Beane uses stats to sign free agents and to trade for additional talent. But he found other ways to use the metrics. In baseball, scouts often work hard to find promising young high school players, and teams have dedicated a significant percentage of their first-round draft choices to the scouts’ finds. But an analysis of draft pick success showed Beane that pitchers signed out of high school are less likely, by a twofold margin, to make it to the big leagues than college pitchers, and high school position players (first baseman, catcher, etc.) are four times less likely. Beane saves his draft picks for college-age players—without objection from his scouting staff.
Cite the advice of experts: Fans and the media screamed for Oakland to keep slugger Jason Giambi when he became a free agent in 2002, but Beane reasoned that it didn’t make sense to pay Giambi $10.5 million per year if it meant the team would only be left with enough money to surround Giambi with a supporting cast of minor league talent. Beane told his critics that businesses have to use cold, hard analysis when making decisions or risk undermining the financial health of the organization. He got his point across by invoking the name of his favorite pundit, value investor Warren Buffett, a strong advocate of taking emotion out of business and managing by the numbers. “Everything he says I can put to use in baseball,” Beane insists.
