Our country's heated presidential race has raised important questions about our next leader's temperament, from the kind of person best suited to mediate our increasingly hostile racial and socioeconomic debates to who should have their finger on the nuclear button as terror groups proliferate around the globe. As voters, we know how important these questions are. You might also find them frustratingly immeasurable.
And if you're like me, you might have also encountered these questions in your own personal life, especially as they pertain to your health care. Indeed, I know that like me, you would want to know the temperament of the surgeon who is about to remove a tumor laying atop your artery, and that of the nurse who is struggling to do a routine blood draw from your screaming toddler.
Resumes, however, are imperfect predictors of temperament or, for that matter, success on the job. There simply isn't enough information in the decision-making process to know if a candidate has the attributes to succeed in a particular position. Resumes, on their own, correlate highly with socioeconomic status, but not that well with talent. Interviews are equally troubling - arguably more useful for recruiting than for predicting whether a candidate is a good fit.
But what if we had more than just resume and interview data? In fact, if we had lots of information about job applicants, including data on current successful employees, we could use technology and data science to more accurately analyze and predict real on-the-job success. It's not magic or rocket science, but big data analytics that make it possible.
Revisiting that nurse trying to draw blood from your screaming child, let's say she did a great job. In fact, let's say there is a subset of nurses at that office who can reliably draw blood from anxious, fidgeting kids. Those nurses are improving patient satisfaction, saving time, promoting efficiency, and enhancing the quality of care. Any hirer wants more nurses just like them. Yet, a resume could never communicate that; all it does is demonstrate some field experience. And while an interview or reference might go further in determining certain personality traits such as compassion, perseverance, confidence, humor, that too does not provide a measurement of such a characteristic.
If instead of looking at one nurse at a time, we look at large groups of nurses who are successful at specific tasks, we can begin to see patterns in the data. We can see that nurses who routinely improve patient satisfaction in a particular department of a particular hospital share specific characteristics. Then as hirers, we can focus our search on applicants with these characteristics. This is big data analytics in action.
Of course, most hirers think they can do this without big data analytics. They just know what a great nurse is. Over the years, they've developed a set of notions of what "success" is, and hire to that arbitrary standard, wed to assumptions they've been using for years. We've all seen this behavior as hirers and applicants: only certain schools matter, only certain experience matters, "job hoppers" need not apply, the list goes on.
This kind of bias is arguably less troubling than the kind we see in national politics: Americans who refuse to vote for Secretary Clinton because of her gender or for President Obama because of his race. Federal rules attempt to protect us from these types of categorical prejudices in everyday hiring. But we are not protected from these more benign beliefs, and they are bias all the same. Perhaps we believe that a great pediatric nurse has to have a friendly tone of voice, has to have experience with children, or has to be able to name seven Disney princesses. And perhaps these assumptions are leading us to hire the wrong people.
When we begin to link a job applicant's concrete data to specific organizational outcomes, we are doing several important things. First, we are democratizing hiring. We are challenging conventional thinking about education, experience, and socioeconomic status - asking whether these things actually matter for on-the-job success. We are transforming our inherent biases into data-driven search, and opening our applicant pools beyond traditional lines.
Second, we are allowing our organizations to evolve. We are rejecting our historical notions of success and replacing them with current strategic goals. We are allowing our labor force - the driver of success in almost every institution - to be best matched for those goals.
Third, we are building a productive, satisfied workforce. Individuals who are appropriately matched to jobs are most likely to be successful, and therefore mostly likely to be fulfilled and stable.
Unfortunately, with only two job applicants, big data analytics has little to offer the presidential race (or should I say YUGE data analytics) but it has the potential to revolutionize hiring for so many other jobs.
Michael Rosenbaum is founder and CEO of Arena, a national company what utilizes algorithms and analytics to enhance recruitment and improve retention at more than 450 health care systems and institutions across the country www.arena.io