Companies consider plenty of data when making hiring decisions. Unfortunately, it’s usually the wrong data, said Daniel Enthoven, vice president of marketing for the San Francisco-based consulting firm Evolv, which specializes in helping companies tie productivity data to hiring decisions.
“Data can be used to make incredible predictions,” Enthoven said in a May 1, 2012, presentation at the HRO Today Forum near Washington, D.C. But, he added, there are “tons of data that is used very poorly.”
For example, 60 percent of employers use credit checks to screen applicants, he said, referring to a Society for Human Resource Management research report, but “we don’t have research to show a statistical correlation between what’s in somebody’s credit report and their job performance or their likelihood to commit fraud.”
In addition, many companies tend to screen out applicants who have changed jobs frequently, but “we looked at the data—there’s no predictive value at all,” Enthoven said.
Evolv has analyzed applicant data and employment outcomes from more than 21,000 call center agents in five major contact centers. According to Evolv, there was “virtually no difference in employment outcomes based on how many jobs a person had, or how many short-term jobs they had previously.”
In addition, “there was virtually no distinction between the ‘perpetually unemployed’—applicants who had no jobs for five years—and applicants who had many jobs in that period.
What about conducting background checks, analyzing handwriting samples and scouring Facebook pages? “No,” Enthoven said. Data gleaned from such activities haven’t been proven to reliably predict hiring outcomes but are nevertheless used commonly because “people want to feel they have a good process. The hard part is throwing out the weak [predictors]. It’s a security blanket,” he said.
Screen Job Candidates for What Matters
What should companies use to screen potential hires? According to Evolv research, the applicant’s personality, aptitudes, work style, technical skills and fit for the position are predictive of performance and attrition.
Companies usually measure recruiting success by time-to-hire, cost-to-hire and quality of hire, Enthoven said. “Two have obvious numbers, so companies optimize on those two. The third gets short shrift. The primary metric should be quality of hire. You need a program to hold the recruiting function accountable.”
To predict what will make an applicant a good hire, first understand what makes for a high degree of fit for that particular job, Enthoven said. Know the skills, aptitudes and work styles of the best employees in the position.
Sometimes companies overlook obvious data they could use to find the best person for a job, Enthoven added. A call center company eventually found “one data point” that indicated which applicants would stay on the job longer.
It was a typing test.
“All candidates were taking the typing test, but the company didn’t realize that for one position, typing was a key predictor,” Enthoven explained. “The company always had the data but never used it.
“Follow the data,” he urged. “I think there’s going to be a next wave of productivity that will be driven by HR using big data to drive quality.”
Stephenie Overman is a freelance writer based in Arlington, Va.
SHRM Online Staffing Management Discipline