When researchers examined the retention rates of 300,000 low-skill service workers across 15 companies, they found that employees stayed in their jobs 15% longer when a computer algorithm helped determine their “employability.” Based on the results of an online test that questioned job applicants about their technical and cognitive skills, personalities, and fitness for the job, an algorithm then rated the results according to a traffic-signal scale: Green (high potential); yellow (medium potential or red (low potential).
Not surprisingly, would-be employees given the green light stayed in their jobs the longest, followed by yellow, and then the universal color code for “stop.” Human managers who dared to defy their digital counterparts in hiring decisions tended to hire applicants who turned out to have a shorter tenure. The secret to the algorithmic success of digital hiring may be that it works best in filling lower-skilled positions, where tenure and productivity are key components.