One of my favorite terms in HR is "base rate of success." Its also rarely discussed, but absolutely critical.
Simply put, the "base rate of success" tells us the proportion of employees who would be successful on the job if we hired randomly instead of applying some predictor.
As Gatewood and Feild (2010) write in their book "Human Resource Selection:"
"If 100 percent of the applicants would be successful if hired, there is no added information provided by the predictor; one would simply randomly pick from the pool of superstar candidates. Conversely, if every applicant in the selection pool would fail if hired, the predictor cannot help identify even one applicant who would be successful; none are successful."
We invest considerable time in our selection tools. We apply certain assessment tests believing they provide added information that would help differentiate amongst qualified candidates. We spend countless hours interviewing the "best" applicants hoping it will give they final impetus on who to choose.
How much better are these predictors than simply throwing darts or picking out of a hat? Have you looked at your measures to see if they are providing you that edge? How much would your company pay improved your ability to select by 3%? 5%? 10%?