HR metrics are usually top of mind for talent leaders. At least they have been since the ’70s with the publication of Jac Fitz-enz’s “The Measurement Imperative” in the Personnel Journal. Metrics have come a long way since then, with numerous metrics identified and benchmarked and a burgeoning industry of HR metrics conferences and consultants ready to collect and offer decision-making data. But have HR metrics fulfilled their promise, and are we positioned to capitalize on the emergent next generation of human capital analytics?
For some organizations, the answer is yes. IBM leverages HR metrics and analytics internally and also runs a large global human capital analytics and business intelligence consulting practice.
Randy MacDonald, senior vice president of HR for IBM, has been a fervent advocate of HR metrics, beginning with a focus on characterizing labor cost in the early 2000s that guided IBM’s expansion of its global footprint before it was fashionable to place huge bets on India or any of the other BRICs. HR analytics guide people decisions at IBM through use of tools such as comprehensive scorecards that help to keep the company at the forefront of HR leadership and innovation.
Google is also in the vanguard of the analytics movement. Analytics are at the heart of its business model, so the application to HR, or people operations as Google calls its HR department, is a natural progression. At an Institute for Corporate Productivity (i4cp) meeting on evidence-based human resources in June, Google’s manager of people analytics, Brian Ong, identified an analytics value chain (Figure 1), which organizations progress through in their quest to get the greatest return on their investment in HR analytics.
At any given time, organizations rest at varied points along the length of the value chain. Absent standards such as generally accepted accounting principles in finance or regulations requiring the use of specific metrics such as those mandated for proxy statements, there always will be organizations at the opinion end of the chain, many of which aspire to evolve out of it. Other organizations will progress toward the action end of the chain, reflecting the growing trend toward competing on talent analytics.
The Rise of Predictive Analytics
Over time, however, organizations’ motivation for moving out of the opinion end of the chain has changed. Fitz-enz, who observed the evolution firsthand, tells the story as follows in his 2010 book, The New HR Analytics. Through the 1970s, HR focused on cost and quantity metrics to demonstrate that the function was managing expense, largely as a defensive maneuver. In the 1980s the motivating force was a desire to show HR was adding value to the business. This led to an obsession with demonstrating a return on investment on people initiatives through the 1990s. This was the point at which the paradigm shifted from running an HR department to managing the organization’s human capital. The new millennium has seen an increasing application of statistics to explore correlations and causation, which has blossomed into the current age of human capital analytics (Figure 2).
Ideally, organizations want the ability to demonstrably impact business results through smart talent decisions. That human capital decisions impact business outcomes is obvious. What’s not so obvious or easy is to demonstrate the existence of a causal relationship in which talent actions significantly affect business results such as revenue growth, profitability, market share and customer satisfaction. The ability to predict what will happen to profitability during the next five years if the workforce profile is altered to rely on offshore resources, or what will happen to market share as a result of investing a million dollars in executive leadership development will cast talent leaders in a new light in the C-suite — not to mention at HR conferences.
What Drives Business Results?
Are scorecards and benchmarks — the bedrock of the age of metrics — still valuable in this new age of predictive human capital analytics? There are differing points of view. On one hand, organizations evolve in terms of capability and culture as they progress up the analytic value chain — learning how to ask the right questions, collect the right data, set up the reporting infrastructure, glean insights from the reports and benchmarks, and take appropriate action. Some may assert that becoming an analytical company is a journey and there are no shortcuts — each step builds on the previous step. And each is necessary to move the organization forward culturally where data and metrics are appreciated and expected to support points of view.
On the other hand, with the ubiquity of data, metrics, benchmarks and tools available thanks to the age of metrics and the increased availability of analytically oriented talent leaders, many organizations can go for the gold: tying human capital metrics to business results in a reliable and predictive way. Other than getting their data infrastructure into shape, they don’t need to build scorecards or worry about benchmarking against the market. Talent managers likely will admit scorecards lose their charm or become obsolete quickly, and benchmarks can be argued away on based definitions and peer group composition. Rather than boiling the ocean with all HR data and metrics, some quick wins using an analytical, evidence-based approach help to sway senior leadership toward further analyses.
The focus is on identifying the human resources processes such as rewards and recognition or learning management as well as human resources capabilities such as leadership quality and workforce adaptability that drive the growth, productivity and innovation, that lead to total shareholder return, return on equity and other measures that comprise i4cp’s market performance index: revenue growth, profitability, market share and customer satisfaction.
I4cp’s 2010 research on talent management metrics demonstrated that higher-performing organizations are more likely to take a strategic approach to human capital metrics than lower-performing organizations. One characteristic of higher performers in terms of talent management prowess is that they tend to focus on certain human capital capabilities more than others, especially these four:
• Leadership success.
• Robustness of talent pipeline.
• Overall employee engagement.
• Management satisfaction with the talent management process.
However, even with the requisite data and strong notions of how HR processes and capabilities impact business results, it is difficult to demonstrate a causal relationship. At least three criteria need to be met. First, the human capital activity and the business outcome need to be correlated. Second, the business outcome needs to occur after the human capital activity. Third, other variables that might be driving or contributing to the relationship must be controlled. These are difficult conditions to meet in the best of circumstances, let alone when dealing with many talent and business outcome variables in a complex environment. Research results that base predictive assertions on correlation need to be looked at carefully.
The Pitfalls of Predictive Analytics
Then what about predictive analytics has captured HR professionals’ imagination in recent years? With so many organizations having graduated beyond the metrics segment of the analytics value chain, this is surely the next step — to build the predictive models that will launch HR into the strategy orbit currently owned by finance and marketing. Predictive models have been around all the time — any statistical model, such as a correlation or regression, is predictive since statistical models deliver inferences by definition.
The problem is relevant statistical models, such as logistic regression or data mining approaches, are complex. They are hard to explain, and the results need to be qualified with so many statements the audience soon loses interest and confidence. Telling the story the data and analyses reveal is a rare skill often located outside of the HR or talent management function. Further, the outputs from predictive models are denominated in probabilities, odds and likelihoods, numbers the average person isn’t comfortable interpreting. Last but not least, because results are probabilistic in nature, there’s no guarantee a prediction will come true. When it doesn’t, no one will remember the prediction was qualified; woe to the talent manager unfortunate enough to make two faulty predictions in a row.
It can be difficult to connect talent management activities with business outcomes in a compelling and credible way. There are data and system issues to overcome, HR process and capabilities to be characterized, accounting and business knowledge to be familiar with, advanced statistical models to master and a knack for storytelling to be honed. None of these are insurmountable challenges.
An organization needs champions to push for this type of HR capability and the means to develop the capability internally. However, “there’s gold in them thar hills.” Organizations can extract competitive advantage from their talent. Companies such as IBM and Google do it every day. There is a thirst for HR to demonstrate its impact on the business, to leverage science and technology and the talent available to do it. Is your organization game?
Amit Mohindra is director of primary research for the Institute for Corporate Productivity. He can be reached at firstname.lastname@example.org.