A Harvard Business Review Analytics study of 230 executives suggested a stunning rate of anticipated progress: Today, 15% said they use “predictive analytics based on HR data and data from other sources within or outside the organization,” yet 48% of them predicted they would be doing so in two years. However, today’s HR analytics are mainly used for cost avoidance, not to drive broader strategic success.
An IBM global survey of over 1700 global CEO’s found that 71% identified human capital as a key source of competitive advantage, yet a global study by TATA showed only 5% of big data investments were in human resources. “Sharing information across silos” was the #1 challenge to big data effectiveness.
Can HR analytics progress as fast as HR leaders hope? The key may lie in getting beyond HR analytics advancements to HR analytics leadership. If HR leaders and their constituents are seduced by the power of HR analytics, they may reinvent the wheel or miss the bigger picture by fixating on predicting traditional things like costs and turnover. The cure is not more powerful analytics, but taking a different perspective on HR analytics.
An Evidence-Based Approach to Avoid Reinventing the Wheel
Companies try to hire the smartest candidates they can find. Some are legendary for interview questions like “Why are manhole covers round?” and “How many golf balls can you fit into an airplane?” In a New York Times interview, Laszlo Bock, then Google’s senior vice president of people operations, revealed that an analysis of Google’s evidence showed such questions don’t predict anything, and “serve primarily to make the interviewer feel smart.”
Google’s HR leadership replaced these questions with “structured behavioral interviews,” that ask all candidates a consistent set of job-related questions, such as “tell me about a time when you solved a really difficult analytical problem.” This was a significant breakthrough, but decades of research existed showing that structured interviews predict better.
[Recommended Read: 4 Best Practices for Choosing the Right HR Analytics Technology]
Google used HR analytics to examine what really set apart their best managers from others, finding that a good manager did eight things such as: “Is a good coach,” “Empowers the team and does not micromanage,” “Helps with career development,” and “Has a clear vision and strategy for the team.” These factors are consistent with prominent findings from scientific studies of leadership, teams, and employee motivation.
HR leaders are often unaware of evidence-based findings, particularly those in scientific journals, in their quest to analyze their own data. Certainly, not all the answers are buried in the studies published in journals, but the journals can better communicate such findings to leaders. And these journals are like scientific repositories of useful ideas, similar to heirloom seed banks that can help avoid agricultural over-reliance on just a few food plants.
HR analytics might more efficiently hone in on promising ideas by drawing on existing evidence-based findings and frameworks.
Avoid Being Seduced by Powerful Analytics Applied to a Narrow Question
The power of HR analytics can be seductive, creating a danger of myopia. A too-narrow analysis can miss the pivotal decisions. Employee turnover and talent acquisition provide two cases in point.
HR analytics improves employee turnover predictions. Google pioneered big data to predict employee turnover, and “get inside people’s heads even before they know they might leave.” Credit Suisse calculated who is likely to quit, and what new career offerings would keep them. Yet, ever-better employee turnover predictions can tempt leaders to assume that turnover is bad and must be prevented. Conversely, optimally investing in people can mean humanely encouraging employees to leave, such as in cases of poor fit and/or when better candidates exist.
Optimal decisions require embedding HR analytics in a framework akin to inventory optimization, helping leaders to not simply lower turnover, but rather optimize employee turnover to enhance the long-term workforce value. Inventory optimization, not simply turnover prediction, reflects the pivotal decision to optimize workforce value.
[Recommended Read: It’s “Like” Harnessing Analogies to Accelerate HR]
HR analytics garners headlines using big data and social media to predict applicant success. One expert said, “if you apply for a job today, you can be sure your prospective employer is going to be checking out your personal brand across all the social networks you are part of to see if you are a good candidate to hire.” Yet, a fixation on the increasing hiring predictions can actually reduce cost-effective talent acquisition.
For example, the 2014 Wharton People Analytics Conference featured a student competition in which teams analyzed data on thousands of Teach for America applicants. The Wharton MBA’s (many from outside HR) did not improve predictive accuracy. Instead, they applied “optimal test sequencing” to identify that interview questions were sequenced wrong.
Each of three interviews rounds was predictive, but the second round screened out very few applicants. Re-ordering the questions in each round meant early questions could screen out more applicants, avoiding costly multiple interviews. HR analytics should consider not only whether questions predict, but what is the optimal sequence. The principle of “optimal test sequencing and frequency” is a staple of MBA classes, since at least the early 1990’s, but in Operations Management not HR.
Making HR Analytics a Priority Through Collaborative Leadership
HR analytics leadership should be a priority, but it’s not about simply applying the latest innovations or waiting for HR to deliver new analyses.
True leadership means understanding and enhancing pivotal decisions. That requires resisting the seductive power of analytics to produce traditional HR outcomes. Instead, leaders need to use business frameworks and evidence-based research.