Analytics: A Resource for Human Resources
If you want to use analytical data to assess employment candidates for low-level hourly positions, or to judge their performance once they're hired, there's usually plenty available. Hourly workers can be judged by objective benchmarks like sales, attendance records and point-of-sale efficiency. These can be correlated with things like commuting time, education levels and employment history to guide the hiring process.
But what about the c-suite?
The use of analytics to guide human-resource decisions about high-ranking executives has a somewhat checkered history, but in general, it's less prevalent for them than for retail floor workers. This is partly because more evaluative data exist for the latter jobs, and partly because those are seen as easier to evaluate in the first place.
"It's easier to measure performance for call centers or salespeople than for managers because the jobs of the former are more straightforward," says Peter Cappelli, a professor at the University of Pennsylvania's Wharton School of Management. "Good performance is easier to measure, so then [it's] easier to get the data."
John Hausknecht, a professor at Cornell's school of industrial and labor relations, agrees that it's easier to do data analysis of lower-level workers simply because more data are available. A big retailer might have hundreds of thousands of floor-level workers, overseen by only a few thousand management-level execs.
Companies are "just not going to have the data and the statistical power to do much on the higher levels," Hausknecht says.
Before a company can use "human resources analytics" at any level, it has to decide what the term even means. Broadly speaking, it refers to the use of quantifiable data of any kind to guide or augment major personnel decisions–hiring, promotion, salaries/bonuses, etc.
"If you take and expand the definition of what is HR analytics, it means something to the effect of any kind of data analysis and reporting that helps to enhance the improvement of individuals and/or organizations," says Laurie Bassi, CEO of McBassi & Co., an executive recruitment firm. This could encompass something as simple as number-based performance reviews, or as complex as predictive modeling.
At least one major CPG retailer is ramping up its efforts to use HR analytics. Last year, CVS Health advertised an opening for a "human resources analytics – service delivery senior manager." The job responsibilities were described as "development and governance of HR data standards and reporting, [and] HR reporting delivery." (CVS did not respond to a request for comment.)
The use of analytics in human resources has a long, and not always distinguished, history in American business. Its first widespread application was after World War II, when it formed what might be seen as part of the "Organization Man" mindset. A shortage of labor, combined with a trend in the social sciences toward quantification, led large corporations to administer a battery of psychological tests–IQ, personality and others–to job candidates. This died off over the decades, mostly because the tests proved to be not very effective. The capacities they measured, whether intelligence or personality traits, just didn't correlate well with things like job performance and company loyalty.
Correlation is the key to what might be called the next generation of human resource analytics. Aided by increasingly sophisticated software, companies are able to determine how well a given metric correlates, if at all, with an individual's subsequent performance.
Some of these correlations revealed that some metrics, including ones often accepted as self-evidently relevant, weren't so at all. The performance of computer programmers, for instance, has been shown to have little if anything to do with the prestige of the colleges they attended–or even if they attended any college.
Evolv, a San Francisco-based company that does third-party screening for low-level job applicants, found that metrics like continuity in employment history, general intelligence or even criminal records did not correlate with job performance or turnover.
MEASURING THE METRICS
If choosing the right metrics for low-level employees is tricky, it's doubly so in the c-suite, due in part to the relative lack of data for correlative purposes. Most companies start with what might be considered obvious ones like education, experience, skills and job mobility. Even some of those, however, can be interpreted in different ways. How much should experience in another field count? Are frequent job changes a sign of flexibility, or disloyalty?
This uncertainty reflects the inexact nature of analytics as a predictive tool, especially when there's a lack of correlative data.
"It's a question of how well can we model and predict individual performance success, and of course, if the truth be told, the answer is, not all that well."
McBassi & Co.
"It's a question of how well can we model and predict individual performance success, and of course, if the truth be told, the answer is, not all that well," Bassi says. "There is still a lot of what you might think of as noise in the data. The best model, the best testing, the best assessment, still doesn't always get it right. You can put your executive recruits through a whole battery of tests and some of them will still fail to be a good match for your organization."
Josh Bersin, principal of Bersin by Deloitte and a frequent writer and speaker on HR and talent management, is even blunter: "In HR, nothing's ever right. Everything is open for debate. I actually think we're going through a process right now of redefining what the metrics are."
To define the metrics of human resources analytics, it helps to define the goals–to determine what HR analytics is intended to accomplish. This is another way of asking: What are the qualities the company most prizes in its executives?
For instance, many companies want loyalty from their top-level executives–they don't want them to leave.
"They're generally high-potential people that are ambitious, and they want to move around from job to job, so the issue is, how do we keep them?" Bersin says. "And so what people use analytics for is to see why people are leaving, who's recruiting them, and what are the missing elements?"
Such analytics have been used in the information-technology sector. They correlate metrics like job history, title and education, along with the previous headhunting tendencies of area competitors, to determine the likelihood of an employee getting snatched.
"There are vendors that sell access to that data, so now the company can go to that person and have a conversation and say, 'Hey, just want to check and see how things are going with you. Are we giving you everything you need here? What can we do to make your job better?' before they disappear," Bersin says.
Another way to gauge an executive's performance is through feedback from peers and subordinates. This has been done on an informal basis since there have been executives, but technology now provides ways to analyze it systematically. This even includes social media like Twitter.
"We're seeing more measurement of sentiment analysis from a social perspective–what are employees tweeting about?" says Ron Hanscome, research director for human capital management technologies at Gartner, an executive recruitment firm. "What's their collaboration within social platforms within organizations, and how is all of that measured and factored into the analysis? And then this gets tied back to the performance of executives."
Such social-media analysis can serve to pinpoint problems or concerns within a company. "If I have a set of executives in my grocery chain, and it's highlighting that this part of the workforce, this set of stores, is less engaged, and the big issue is that senior management is not taking things in the right direction, then maybe I have to dive into what are the skills and competencies of my leaders around being able to communicate," Hanscome says.
More generally, executives can be assessed through systematic feedback from those around–and below.
"There are some companies in the consumer space that have done some pretty cutting-edge work around measuring sales manager performance, and the performance of sales teams, based on the ratings of the regional sales managers," says Dan Kaplan, a management partner and head of global HR practice for consulting firm CTPartners. "There are some companies where HR is starting to measure the characteristics of consumers and employees to see if there is a natural bridge between the type of people the company attracts and hires, and the type of customers that buy the product."
Hanscome says that such data–evaluations from peers, subordinates, major customers and others–can be collected and assessed systematically through surveys, for what he calls "360-degree feedback."
"All of that is pulled together into an assessment of where that leader is doing well, and where they might need to be trained or mentored, or where they need to develop in order to be more successful in the organization. And so that's nothing new," he says. "What is newer in that space is the ability to pull in data from the responses of subordinates and the like from these engagement surveys. They highlight issues or areas of strength, and that becomes part of this broader assessment of leaders in the organization."
HELP WITH HIRING
These kinds of external assessments can be useful for evaluating the performance of executives already in place. But how can human resource analytics help with c-suite hiring?
Some companies use various forms of simulations to gauge candidates' reactions. These can range from simple verbal responses to hypothetical scenarios, to more elaborate computer-scored answers to questionnaires.
"There are assessments that can be developed that a person can take to demonstrate just how their brain is wired, how do they adapt to new situations?" Hanscome says. "There are more immersive simulation-type environments that have been done to try to highlight how flexible and how adaptable they are."
Kaplan, on the other hand, believes that in c-suite hiring, experience is a more reliable guide than simulations.
"When they were under fire and needed to make the right decision, did they use the right judgment? The real, experiential questioning tends to provide more valuable answers."
A candidate "can project a great answer on how they would handle something in the future, but a better predictor of that is to understand how they handled a similar situation in the past," he says. "When they were under fire and needed to make the right decision, did they use the right judgment? Did they execute the right outcome; were they able to build what was needed? The real, experiential questioning tends to provide more valuable answers."
Retailers should be especially amenable to using data for HR purposes, at whatever point in the career process, Bersin says. After all, retailing is one of the most data-driven business sectors; it's a question of carrying that mindset over to HR.
"Retailers are generally filled with data people, but they're on the product side, the consumer-analysis side, not the people side," he says. "I think if you just think about the exact same sort of problem and apply it to people–instead of product, name and brand, you apply it to the person–you can do similar analysis."