When it comes to getting hired, people always say, “It’s who you know.” More than that, it’s what you have in common. A shared alma mater or mutual friend is often the deal-sealer in job interviews when the recruiter is forced to choose between two equally qualified candidates. But such chummy behavior can sometimes mask hidden prejudices: the fact that two people went to the same school can often reflect their socio-economic similarity.
A recent New York Times article described the emphasis on cultural similarity as latent discrimination. “A cultural fit is an individual whose work-related values and style of work support the business strategy,” according to Lauren Rivera, a researcher at Northwestern’s Kellogg School of Management. “When you get into a lot of the demographic characteristics, you’re not only moving away from that definition but you’re also getting into discrimination.”
Now, with tech companies like Facebook receiving criticism for their lack of diversity (only 31 percent of Facebook’s employees are women, and at 57 percent, most are white), growing tech companies with roles to fill are wondering if algorithms might be able to find the most qualified candidates in a hiring process that seems, at first glance, blind.
How Do They Work?
There are a few different companies offering different algorithms at this time. Gild, for one, scours the web for publicly available job data posted on sites like LinkedIn or GitHub and looks to match skills to the company’s job description, searching beyond what’s listed on a resume. That means that if your hobby is relevant to the job description, but not listed on a resume, you might still appear in search results (if, for example, you’re a barista who writes brilliant code at night). Gild also takes career patterns into account — searching for candidates who are likely ready to jump ship soon, for example, based on how long they’ve been with their current company.
Doxa caters to women in particular who are looking to work at tech companies that will actually appreciate them. In addition to matching skills with company needs, their algorithm uses anonymous employee surveys that describe the nebulous, off-the-page qualities of companies (how often employees work nights and weekends, time spent in meetings, whether there is perceived gender bias) to match candidates with compatible company cultures.
Another service, GapJumpers, allows companies to write questions that applicants answer anonymously. The answers are then ranked based on how well they match the company’s, who is then given a list with the identities of the candidates revealed so they can start scheduling interviews.
How Well Do They Work?
A Harvard Business Review analysis looked at 17 hiring studies that used algorithms and found that they outperformed human recruiting by about 25 percent. The data was based on post-hiring stats, such as the supervisor’s ranking of the hiree after having been at the job for a while, the number of promotions, and how well the employee did during training.
The study noted that without an algorithm, recruiters are 85 to 97 percent likely to rely on intuition when assessing candidates.
“We don’t advocate that you bow out of the decision process altogether,” the researchers concluded. “We do recommend that you use a purely algorithmic system, based on a large number of data points, to narrow the field before calling on human judgment to pick from just a few finalists—say, three. Even better: Have several managers independently weigh in on the final decision, and average their judgments.”