Can Big Data Make Your Next Staffing Decision?
New algorithms can help identify great potential hires. But executives will need to better understand their organizations' needs to make use of it.
What was your decision process for the last person you hired at your organization? Were the skill sets listed on the resume the most important thing, or was it how she aced the interview? Was the hire a product of rigorous thinking about broader strategic goals, or did you pretty much just like the cut of his jib?
Most likely it was a bit from column A and a bit from column B—balancing the hard facts of what tasks you need performed versus the character of the person who’ll perform them. But an argument is circulating that column A can get you pretty much everything you need.
In “They’re Watching You at Work,” the Atlantic‘s Don Peck reports on how some companies are using rigorous data analytics to better hire, manage, and (if need be) fire employees. Xerox, for instance, has been using an online evaluation for customer-service jobs that includes scenario and cognitive-skill testing; after a test-run of the program, its attrition rate fell 20 percent. Royal Dutch Shell has partnered with a start-up called Knack to help identify some of the company’s best “idea generators.” Interestingly—or creepily, depending on how you sit with these things—Knack has metrics for characteristics that on their face seem to be more squishy and intuitive, such as “mind wandering,” social intelligence, and conscientiousness.
The cut of your jib? There’s an algorithm for that.
Peck is optimistic about all this, albeit cautiously. “[T]his is undeniably unsettling,” he writes. “Should job candidates be ranked by what their Web habits say about them? Should the ‘data signature’ of natural leaders play a role in promotion?” I have my doubts about this as well, which is apparently a more provocative position than I would’ve thought. A couple of weeks ago, when I suggested that leaders need to engage in some human interaction as well as run the numbers when making strategic decisions, commenters pushed back. “Just one more excuse for association leaders to procrastinate on analyzing their data,” one wrote.
But I think we can agree on a couple of things here. First: When it comes to hiring and management, leaders too often rely on gut-feeling approaches that neglect what an employee has accomplished and what roles need filling at an organization. Second: The metrics we tend to use, when we use them, are insultingly reductive. Peck notes that our economy tends to define one’s workforce potential almost exclusively on our educational background. And once we’re on the job, we’re subject to performance evaluations that stuff employees into a forced-ranking methodology or a soul-crushing annual questionnaire—“this weird form you fill out every year that has nothing to do with everyday life,” as one organizational psychologist described it to the New York Times.
Any skepticism about a big-data approach to HR will speak to the sense that the system is already deeply dysfunctional and impersonal, and that the brave new world Peck reports on is just old Tayloristic wine in new algorithmic bottles. The Atlantic‘s cover treatment for the story deliberately stokes this concern, depicting three boss-types, one of whom is holding a stopwatch, hovering over an employee.
No, associations shouldn’t procrastinate on analyzing their data, for HR or anything else. But the programming term “garbage in, garbage out” is instructive here. Without close attention to the host of skills—hard and soft, tactical and strategic—you need throughout an organization, and thought to the ways those skills will be measured, a data-driven hiring plan can only make things worse. So assuming that, per Peck, that pretty much any skill is measurable, what are the skills you most want to measure?