HR Meets Big Data: How Analytics Are Helping Companies Identify and Develop Top Talent

Resumés are notoriously inaccurate and misleading. Studies show that interviews are not as effective as managers think they are. Plus, getting applicants to fill out questionnaires, whether online or off, produces poor results. It’s even hard to persuade tech savvy people in their 20s and 30s to sit down and fill out a questionnaire by hand. “Hiring is very challenging and difficult,” explains CEO Halfteck. “Job seeking for people is also very challenging and difficult. I was determined to find a better way.”

His search led him to game-playing, which may seem trite at first glance but is actually quite profound. He discovered that if job applicants can escape the formality of an interview process, they open an authentic window into how they really behave and think. So Halfteck launched a company called Knack, headquartered in Palo Alto, California, which provides simple online and mobile games similar to Angry Birds that companies can use to screen job applicants. In less than 10 minutes of game playing, during which the player must react to and interpret facial expressions of customers in a restaurant, Knack gathers enough information to make key judgments about her potential. The company applies its own algorithms, written by data scientists and behavioral neuroscientists, and runs them on’s cloud-computing server farms.

“It’s kind of astonishing how poor the analytics have been in human resources given how much money goes into it. Most  companies haven’t bothered to try to figure out who is the best employee or why they quit.”

“We are able to get very quick insights into personality characteristics, decision-making capabilities and the overall soft skills that a person has,” Halfteck says. “We can measure those attributes and use the data to predict whether that person will be a successful employee.”

Big Data has taken on somewhat of an ominous overtone following the disclosures of massive spying by the National Security Agency, but that is not stopping CEOs from continuing to explore new ways to exploit the advantages of massive computing and data crunching know-how. Already in use for mass-marketing endeavors, those tools are now being applied in one of the dark corners of the corporate world—human resources. “It’s kind of astonishing how poor the analytics have been in human resources given how much money goes into it,” says Peter Cappelli, a management professor and director of the Center for Human Resources at the University of Pennsylvania’s Wharton School of Business. “Most  companies haven’t bothered to try to figure out who is the best employee or why they quit.”

But CEOs are starting to realize that they can get smarter about the entire lifecycle of human resources—recruiting, hiring, onboarding, training, measuring performance, retaining and predicting departures—and they can do all those things in relatively cost-effective ways. Companies from the Fortune 100 down to regional restaurants, hotel chains and hospitals are embracing the new analytic tools.

“This is going to forever change how senior executives manage and interact with their work forces,” says Max Simkoff, founder and CEO of Evolv, a privately held startup in Silicon Valley that provides tools to help companies improve their HR capabilities. Evolv is one of the many small, non-traditional players to emerge at the intersection of information technology, human resources and psychology. “When you look at who is buying our solution, it’s typically someone in the C-suite,” Simkoff adds. “They don’t relegate it to the HR department.”

To be sure, potential hazards can arise from any major corporate initiative that borders on becoming a fad. The most obvious is that companies come to depend on their analytic tools at the expense of basic common sense—the human factor. “A lot of companies have said, ‘We’re just going to use our software to screen the candidates. It’s cheaper,’” says Cappelli. “But that has largely been a disaster.” It’s also a highly fragmented field with different vendors offering different tools for the various pieces of the employee’s lifecycle. Even eHarmony, the online matchmaking service, has sought to apply its romance algorithms to unraveling the mysteries of the employee-employer relationship.

The Holy Grail for any CEO is to combine data from the recruiting process into onboarding and assessing how an employee performs over time and whether he is at risk of leaving. However, that information is difficult to achieve, even by investing in large software packages from Oracle, SAP or IBM. “All the big guys have suites of options you can put under the same umbrella, but the issue of integrating them remains difficult,” says Cappelli. “I don’t know that anyone is really good at that.”

Digging Into Dilbert
Whereas Knack’s specialty may be in assessing potential new hires, Evolv concentrates on understanding the health and efficiency of an existing workforce. Its tools, Evolv CEO Max Simkoff developed by industrial organization psychologists and data scientists, look at how long it takes a worker to come up the learning curve to do his or her job properly, at productivity once an employee is performing well and at attrition—or why someone quit.

It can study the impact of a supervisor’s conduct and whether the pay level was a factor in a decision to leave. It combines that info with external data from different industries and geographies to offer CEOs insights into how to identify, develop and encourage their best workers. Its customers skew toward larger companies, but it says customers with as few as 1,000 employees can benefit from its predictive analytics. In most cases, “we’ve improved productivity by 10 percent, lowered attrition by as much as 35 percent and lifted customer satisfaction scores by 4.9 percent,” says founder Simkoff. “What that means for senior executives is they can optimize and improve their existing workforce to deliver better performance.”

As with any technology solution, Simkoff acknowledges that algorithms should never replace human judgment. “Tools like ours are about supporting executives to make better human decisions,” he says.

In practice, the way Evolv’s technology works can be glimpsed at The Results Companies, which is based in Fort Lauderdale, Florida and has roughly $120 million in sales. Results helps other companies outsource their business processes and operate call centers or any sales or service function. Forty percent of its employees are in the U.S., 40 percent in the Philippines and 20 percent in Mexico. “If training and experience and having customer interaction is important to your business, analytical hiring and measurement is suited for any size organization,” says Alec Brecker, Results’ CEO.

The company’s software rates prospective hires with a color-coded system in which red means “don’t hire,” yellow means “hire with caution,” and green means “hire.” By concentrating on hiring green candidates, it has increased the tenure of its own employees by more than 20 percent, Brecker says, and it has helped customers increase the tenure of their employees by 5 to 8 percent.

One of the most intriguing possibilities is introducing customer-satisfaction data into the mix so that a CEO can know what type of employee is best at making customers happy. “That data is constantly being fed back and reviewed to ask, ‘How do we tweak the hiring profile of people?’ Brecker says. The company’s tools also can increase the total amount of sales per employee and reduce the number of customer complaints that a health insurance provider, for example, faces every day. If safety is a priority in a manufacturing company, The Results Companies can screen applicants in terms of their attitudes toward being careful.

The bottom line? CEOs now have tools that can give them unprecedented insight into how to hire the right employees and maximize their performance.

Sidebar: HR Tips From Moneyball

" William J. Holstein : William J. Holstein is a journalist, consultant and speaker. He is the author of, "The Next American Economy: Blueprint For A Sustainable Recovery." For more of his work, visit"