Our Chapter 10, Human Resources, Job Design and Work Measurement, covers almost every OM aspect of dealing with employees–except how to hire them. As I think about all the instructors and staff I have hired over my 40 year academic career, I realize many personnel decisions were based on resume length and intuition. So The Wall Street Journal’s article (Sept.20, 2012) on how computer modeling is upending the way workers are hired caught my attention.
For more and more companies, the hiring boss is an algorithm. Jobs that were once filled on the basis of work history and interviews are now determined by data analysis. Under pressure to cut costs and boost productivity, employers are trying to predict specific outcomes, such as whether a prospective hire will quit too soon, file disability claims or steal.
The new hiring tools are part of a broader effort to gather and analyze employee data. Globally, spending on so-called talent-management software rose to $3.8 billion in 2011. Though hiring is a crucial business function, conventional methods are usually short on rigor. Depending on who decides, what gets candidates hired can vary wildly—from academic achievement to work experience to appearance. Managers hunches generally have little value in predicting how someone will perform on the job. The statistical approach to hiring can improve results by reducing the influence of a manager’s biases.
When looking for workers to staff its call centers, for example, Xerox used to pay lots of attention to applicants who had done the job before. Then, a computer program said that what does matter in a good call-center worker—one who won’t quit before the company recoups its $5,000 investment in training—is personality. After a short trial that cut attrition by a fifth, Xerox now leaves all hiring for its 48,700 call-center jobs to software that asks applicants to choose between statements like: “I ask more questions than most people do” and “People tend to trust what I say.”
Discussion questions:
1. What are the advantages of this software-driven approach?
2. How can algorithms help determine how much to pay workers?