Moneyball-grade analytics in HR: Be careful what you wish for
It has been almost a decade since Michael Lewis’ Moneyball: The Art of Winning an Unfair Game was published, illustrating the power of analytics in developing a successful major league baseball team despite severely limited financial resources. HR has since taken keen interest, as analytics such as Moneyball demonstrated holds great promise to transform organizations, especially as data about everything is becoming plentiful and cheap to collect and maintain. Even as HR struggles to become a strategic partner, analytic techniques like those described in Moneyball represent a significant source of competitive advantage.
What would HCM really look like if business organizations were able to gain insights from their data that analytics promises? Moneyball gives us both a good idea, and an admonition to be careful what we wish for.
We will likely discover that the labor market is appallingly inefficient, and HR is a big part of the problem. We will find that talent is terribly mispriced and mismeasured. We will find our “tried and true” performance metrics contain little information about what individual contributions put points on the scoreboard, and are no more valid than the venerable though errant figures that have been reported in box scores for over a century.
Not what you were expecting? Read on.
Pay is disconnected from productivity
Talent management assumes that talent is scarce, particularly at the managerial level and in specialized skills. As with many of today’s big payroll teams, whether in baseball or business, compensation policies for such talent reflect these assumptions.
However, the first and most obvious lesson from Moneyball is that pay and productivity are not highly positively correlated. Like the New York Yankees or Texas Rangers in Moneyball, having a lot of big money stars in the heart of the lineup will not drive increased win productivity. In addition star performance is not likely to be replicable or portable, and so paying big money for star performance turned in at another firm can be highly risky.
Analytics is helping us discover that truly valuable human capital is indeed firm-specific, and worth the price of investing in its development. Yet just like baseball scouts, baseball managers (field managers responsible for rosters, player substitution and daily performance), and baseball general managers (executive level managers responsible for talent acquisition, trades, compensation, and player development), HR departments are frequently poor assessors of human capital, and do not evaluate it properly or at all during selection. In a Moneyball world, analytics will illuminate what aspects of human capital we should be measuring, and determine what we should be paying for.
Talent and skills are mispriced
Economic theory practically guarantees that talent will be mispriced due to the information asymmetry between companies and persons on the labor market. Moneyball shows us that a truly transformed analytics function will reveal by how much. Regularly accessed salary benchmarking databases will likely be found to be full of meticulously kept, but flawed data, revealing little information about which job categories drive business performance. There will likely be many surprises about which roles are overvalued, and which roles create great value but accrue very little of it to the employee.
Interestingly, some highly visible executives and performers in the organization whose compensation packages are contractually determined might actually be more fairly priced than assumed, as their performance is subject to greater internal scrutiny, their performance metrics more holistic, and compensation subject to greater market discipline.
In a Moneyball world, the assumption of talent scarcity will also be challenged. An analytically transformed HR function will likely find that some types of talent may not be all that scarce, and that artificially induced scarcity such as forced ranking, “up-or out” systems of performance management, or delayering really benefited no one, despite their decades-old practice.
Standard and valid metrics will change the labor market, and HCM practices
As with professional baseball, analytics will give rise to new or recombined performance metrics that will eliminate inefficiency within the organization, an area where HR has become particularly adept since the economic recession of 2007-2008. Yet the other side of gaining a whole new set of metrics is that inefficiencies that had been previously exploited will offer few such opportunities in the future, as the true drivers of business performance become more widely known and measured. Analytics will kill as many opportunities as it discovers.
For example, it is unclear right now who benefits from the labor market information asymmetry, but it is clear that someone is paying too much. As the value of individuals and the pay they receive becomes more a matter of fact than speculation, some job categories may demand a greater portion if the gap appears too wide in favor of the company.
For the last twenty years strategic human resource management has sought the missing link between HR and tangible business results, and has found dozens of measurable business inputs that are correlated with hard business and financial outcomes. However we still do not know why these input behaviors drive business value, although we are fairly sure they are not direct inputs of financial performance. Analytics will be sought out less for determining what HCM investments drive performance, and more for determining the behaviors HCM influences, which in turn drive business performance. That is much more difficult to do, but will become the hallmark of a truly analytically-transformed HR function.
I loved the movie "Moneyball" and this blog opened my eyes to what would happen if those concepts were applied in HR. Intriguing!
To me that's the biggest lesson of Moneyball - HCM analytics will change the whole competitive environment for all those who use it.
Hi Ray,
I still haven't seen the movie yet, but I read the book several years ago and found it very interesting. When you consider the amount of folk lore and invalid 'rule of thumb' measurements that accumulate around sport, which provides almost immediate feedback, then it is not surprising that the 'People Management' field has similar issues.
I believe that the breakdown between added value to the employer (pure talent won't necessarily correlate, as anyone who has seen a techie promoted to management will attest) and fair compensation for that added value began when the Personnel Department became the Human Resources Department. The implication in Resources is that employees at certain levels and niches are interchangeable. We all know that is not so; For example, the body of research that suggests that an expert developer is 10 times more productive than the average developer is as solid as any research that’s been done in software engineering, but we very rarely see this reflected in salaries or conditions.
While my head tells me that you'll get more productivity (output divided by cost) by using money-ball style analytics, my gut and (eyes) tell me that treating people as people, not resources, will get you even more.
hth
Hello Martin:
Thank you for you very thoughtful comments. You raise several important points yet I am afraid I will be able to address only one of them right now, referring to people as a resource.
Indeed, a good deal of folly in management thinking in the past has proceeded from the notion that every aspect of a business organization in engineerable, including human performance. Stanford business professor Jeffrey Pfeffer warned HR of this in his 1997 paper, Pitfalls on the road to measurement: The dangerous liaison of human resources with the ideas of accounting and finance. Some of today's more enlightened management gurus are constantly reminding us that algorithms are no substitute for good judgment.
My colleagues and I are acutely aware of this. When we use the term resource among ourselves, we mean it in a neutral way, as there is no implication that people ought to be treated like machines. Resource to us denotes simply something that is measurable. Of course not all human performance activity is measurable, and HR has suffered over the last few decades from poor data. But conscientious data science and good manager judgment together go a long way in capturing value-driving behaviors by means of HR metrics. Our aim in analytics is to help sharpen judgment so that we can get better at treating people as people, rather than creating algorithms to substitute for good judgment.
What Michael Lewis does is also helping interstand how to uncover unexpected relationships between metrics, that is to say relationships that are invisible to the bare eye.
By the same tokens, I believe the next challenge in HR data will be to indentify relations between HR metrics and business metrics that are invisible to the bare eye of a manager. THis will give them leverage. If I know that allowing employees to bring their kids to work one day per week can significantly improve the way employees are dealing with customers, thus increasing customer satisfction, this is an easy level I can pull. But I don't know that until I am told based on correlation analysis.
Savvy customers and customers with a history of quantitative analysis already do this, but it is important to see how it can be generalized.
Hi Jean-Bernard,
You make an excellent point that I think is often overlook: "uncover unexpected relationships between metrics". It is these unexpected relationships that can provide a competitve advantage to an organization. Being able to correlate why particular metrics or KPIs are occuring can help to resolve/utilize/amplify action (depending whether the correlation is positive or negative to the organization). In this age I don't think it's enough that analytics are purely for reporting on what has happened; I think organizations require predictive analytics and systems that can identify correlations between analytics - this is what will help drive business efficiency and operational excellence as the world becomes even more competitive.
I know SAP are working towards this area and I'm really keen to see what they come up with to help businesses move towards these types of benefits.
Best regards,
Luke
Hi Ray,
This is a great blog and highlights some key points about the usage case for analytics. I really hope businesses begin to move towards using correlated analytics and predictive analytics because that will provide significant benefits towards operational excellence.
I think it can be tough for organizations to find correlations, depending on what sort of personnel they have available to them. Anything SAP can provide to bring the business closer to this will be a big step forward in my opinion.
Best regards,
Luke
Hello Luke:
Agree! We just concluded a research study with the Human Capital Institute on the state of HCM analytics, and found that technology alone is not enough for an organization to transform itself through analytics. While this might seem self-evident, we should note that when SAP brings top-shelf solutions to customers, we should also bring insight, not only to help users become smarter users, but to back up our solutions with the knowledge that went into developing them. Herein lies an opportunity to help our customers get to high levels of analytical proficiency much quicker, and to use our products the way they were intended to be used.