article » Hiring in the Age of Big Data

Hiring in the Age of Big Data

October 24, 2013
3 min read

Wasabi Waiter looks a lot like hundreds of other simple online games. Players acting as sushi servers track the moods of their customers, deliver them dishes that correspond to those emotions, and clear plates while tending to incoming patrons. Unlike most games, though, Wasabi Waiter analyzes every millisecond of player behavior, measuring conscientiousness, emotion recognition, and other attributes that academic studies show correlate with job performance. The game, designed by startup, then scores each player’s likelihood of becoming an outstanding employee.

Knack is one of a handful of startups adapting big data metrics to hiring. The companies are pitching online games and questionnaires to corporate recruiters frustrated by the disconnect between a good interview and an ideal employee. Based on records of how star workers responded to the same tests, these services predict whether a candidate will be suited for a particular job. Clients use the tool to help winnow piles of applications. “People are our biggest resource, and right now a lot of them are mismatched,” says Erik Brynjolfsson, an adviser to Knack and director of the Center for Digital Business at the Massachusetts Institute of Technology Sloan School of Management.

Evolv uses questionnaires to evaluate candidates for hourly work at Xerox (XRX), marketing company Harte-Hanks (HHS), and other businesses. Questions range from how close to the office a person lives to how many social networks he uses. Since 2007 the roughly 100-employee company, based in San Francisco, has collected terabytes of data tracking the results of its surveys and candidates’ real-life employment history—how long they stayed with the company and various measures of job performance such as customer satisfaction surveys. Evolv’s software translates new applicants’ results into a traffic light system for hiring managers. (Green means an applicant has “high potential.”)

BLOG: Big Data Not Living Up To Its Promise? Change the Way You Work

Workers at Harte-Hanks’s call center who were selected by Evolv missed about 29 percent fewer hours of work in their first six months and handled calls 15 percent faster than those hired before the company began using the software, according to a study published by Evolv. For “our largest telco customers, a single percentage-point increase in any customer experience metric they track, they can correlate to additional percentage points in subscriber base,” says Chief Executive Officer Max Simkoff.

David Ostberg, Evolv’s vice president for workforce science, says the data the company has collected debunk common recruiting assumptions. For instance, Evolv’s analysts have found that a history of job-hopping or unemployment is a poor predictor of how long a person will stay on a job. “As human beings, we’re actually pretty bad at evaluating other human beings,” he says. Simkoff says Evolv focuses on evaluating hourly workers, because “the performance data is not there yet” for employees engaged in higher-level tasks.

Two-year-old New York-based startup ConnectCubed says it has the data to tailor its online questionnaires and games to evaluate candidates for any position. Using personality surveys along with simple spatial-reasoning, trivia, or memory games, ConnectCubed tests a company’s longtime stars to develop ideal behavioral profiles for each job. “When new people apply, you can say, ‘Wow, this guy has all the makings of our top salesmen,’ ” says CEO Michael Tanenbaum. “These are things that are impossible to measure from a résumé, especially with educational backgrounds that are often more determined by socioeconomic status than your innate ability.”

Like Knack, ConnectCubed is too new to have tracked the long-term performance of the employees its software has recommended or the bottom-line value for businesses. “My concern is, with only a 9.5-minute sample of behavior, is that really enough?” says Frederick Morgeson, a professor of management specializing in personnel psychology at Michigan State University, referring to Wasabi Waiter. “Are we sampling enough of those behaviors to be confident that we’re capturing what the person might do in the totality of their complex behavior?”

Algorithms will never replace the interview altogether, but more information about each candidate helps make the hiring process less of a guessing game, says Erik Juhl, head of recruiting at video-ad startup Vungle, which uses Wasabi Waiter. “You have this enormous pool of people that’s being missed because of the way the entire industry goes after the same kinds of people, asking, ‘Did you go to Stanford? Did you work at this company?’ ” says Juhl, formerly a recruiter at Google (GOOG) and LinkedIn (LNKD). Eventually, he says, he wants Knack to be as big a factor in his hiring process as the opinion of his most senior colleagues.

For the full-length version of the article, please click here.