The Major League Baseball season is upon us and, in the past three weeks, I have attended no less than three presentations where Moneyball, the feature film based on the book by Michael Lewis, has been referenced in regard to the success of big data. In the narrative presented, baseball’s low-budget Oakland A’s use big data to beat their rivals at baseball. And, while it is a fabulous story of using statistical analysis and arbitrage to gain a competitive advantage, using it as big data isn’t entirely accurate and somewhat misses the point...
Moneyball isn’t a story of the Oakland A’s installing giant servers and using complex analytics in an initiative to get *more* data than their competitors and beat them with a barrage of data (although teams would do this later). The A’s had access to generally the same data sets as every other team at the time – they simply used the data more effectively to help them focus on what really mattered. Specifically, baseball had traditionally focused on measuring success of a single player through measures like batting average, RBI’s or slugging percentage. The logical path was that a team of successful players would clearly make a successful team.
The revolution in baseball that the Oakland A’s launched was less about big data and much more about measuring the right things. Advanced statistics started to measure ultimate outcomes of individual performance in Wins (quantified as Win Shares or Wins Above Replacement), which in baseball is what ultimately matters. When they started measuring in terms of Wins, it turned out that a lot of the conventional wisdom in baseball was wrong. Defense matters a lot more than had been previous attributed. Walks matter a lot and Stolen Bases do not. It’s changed the way teams are constructed and the way the game is played. Its success has even sparked a second evolution that IS about big data – collecting data on the speed and location of every pitch’s speed, location and movement for advanced analysis.
When I look at metrics in HR, I see a lot of parallels to the Moneyball revolution in baseball. We’ve gotten pretty good at our standard metrics of HR and measuring the performance of hiring. Most recruitment teams can speak intelligently about their time-to-fill, aging requisitions, source data and perhaps even cost-per-hire. SHRM has embarked on an admirable journey to standardize these types of metrics. The problem is pretty similar to baseball’s breakthrough with the Oakland A’s. The metrics HR currently used are very good at trending the success of HR, but not very good at relating to organizational success. The organizations at the forefront will measure to the bottom line of the organization, not to trend lines within their function.
Such a shift is very difficult and requires serious work in deep diving on HR’s impact on the bottom line (or other success metrics for non-profit organizations). A good example is the difference in the cost-per-hire and cost-per-vacancy metrics. The cost-per-hire calculation is easy and comfortable, typically consisting of advertising spending, recruiter salaries and other direct costs. It’s easy to compute and easy to trend. In a cost-per-vacancy calculation, the business need is computed, generally using an individual’s salary (or a multiplier of that salary) under the foundational economic principle that a person is more valuable to the organization than what you pay them. The problem is that cost-per-hire greatly undervalues the cost of the vacancy to the organization and therefore undervalues the speed at which positions are filled. Concretely, an organization might see as a win that they spent less money on advertising and reduced cost-per-hire, but the organization perhaps lost because it took longer to fill the roles, which increased spend on contingent staffing, lowered morale, or generated overtime expenses.
As I see the HR world get excited about big data, I have to admit to getting excited about the possibilities as well. We will have more data points than ever before which will lead to new insights and new possibilities for our profession. But, when I think about the Moneyball example, I remember that it’s not how much you measure that matters, but rather ensuring that you’re measuring the things that matter. The Oakland A’s gained an advantage by calculating value in the context of what really mattered – wins. I’m hopeful that HR can do the same.
Post contributed by Adam Godson. You can follow Adam on Twitter at @adamgodson or connect with him on LinkedIn.
