LinkedIn has upended corporate recruiting in the past decade, allowing talent scouts to scour a vast database of 200 million people’s career profiles. That was just the start. Now LinkedIn has created algorithms that might do the sorting even more nimbly. The result: a digital cheat-sheet for recruiters, called: “People You May Want to Hire.”
Fruits of the new algorithms won’t be visible to ordinary users of the workplace social network, who get a basic service for free. Instead, the latest offering is aimed exclusively at LinkedIn’s best customers: the 16,400 enterprises that buy LinkedIn’s top-of-the-line profile-searching system, known as Recruiter. In a San Francisco briefing today, LinkedIn executives explained that they are about to roll out an updated version of Recruiter, with “People You May Want to Hire” as a prime example of what’s new.
In an interview, Parker Barrile, the LinkedIn product manager overseeing Recruiter, explained some of the not-so-obvious factors that will guide the new algorithms. By sorting through its own data, LinkedIn knows that computer engineers migrate back and forth between New York and San Francisco a fair amount. So California companies with a tightly drawn set of skills requirements might be shown a few New York candidates. “But we wouldn’t show them someone from Reno,” he said — because there’s very little job mobility between that Nevada market and Silicon Valley.
In finance, Barrile added, it may be appropriate to offer some London candidates to a New York bank trying to fill investment banking positions, because of similar mobility patterns. It wouldn’t make sense to show trans-Atlantic candidates for hedge-fund jobs, though, because migration patterns are much more limited.
LinkedIn’s candidate-spotting systems are designed to be quick learners, Barrile said. Initially, LinkedIn will analyze recruiters’ current selections of job candidates — and then find more prospects with similar careers, skills and education. If recruiters click on some suggestions and start an e-mail dialogue with candidates, those will be regarded as especially good matches. If suggestions are ignored, those will be scored as poorer matches. Over time, LinkedIn will realign its matching patterns in an effort to deliver more of what’s valued and less of what isn’t.
The new algorithms can’t run without a base of human-generated candidate preferences. But in the years to come, LinkedIn’s algorithms could play an increasingly important role in shaping companies’ overall prospect lists. Barrile pointed out that LinkedIn already offers similar algorithms to help ordinary users build out their contact lists, under the label “People You Might Know,” as well as career leads under the title: “Jobs You Might Be Interested In.”
Before such products existed, Barrile said, 100% of users’ job prospecting was based on their own searches. Now, he said, more than half of job prospecting is driven by what the algorithms recommend.
This article first appeared in Forbes