Dive Brief:

• A study released earlier this month shows that computerized recruiting, aka recruiting with a testing process, is very effective – but hiring manager humans “really screw things up,” the Washington Post reported.
• The study involved a dataset of 300,000 hires at 15 companies that test for low-skilled positions, like call centers. Researchers measured how people were assessed initially and whether a manager overruled a low test score to bring them on and how that affected work performance later on.
• Not only did testing “improve job tenure by 15%,” according to the Post, but when humans intervened, the results were spectacularly worse.

Dive Insight:

To add salt to the wound: Although the workers picked by the algorithm didn’t turn out to be more productive than those for whom a hiring manager stepped in, they weren’t less productive, either – meaning recruiters weren’t effectively trading off between effectiveness and longevity on the job, according to the Post.

Julie Moreland, senior vice president for strategy and people sciences at Peoplematter, told the Post that “about one-third of hiring managers don’t place enough weight on the assessments.” Moreland said that may partly be due to high turnover in management positions leading to inexperienced bosses who don’t know how to use them effectively.

Unfortunately, while algorithms certainly aren’t free of internal biases, having inexperienced humans step in to perform hiring often leads to even greater prejudice entering the system. The best scenario? That HR humans become experts on a company’s hiring needs that can be repurposed into an algorithm, saving time and money.

“What true professional HR providers realize is they’ve taken something and made it more efficient," Moreland said, "and therefore they can spend more of their time on strategy rather than interviewing.”