I was talking with a very smart guy last week, who told me that weather prediction has gotten 50% better in the last 10 years. This involved a lot of modeling and AI in combination with hundreds of data scientists looking at every variable that affects the weather within 2-3 degrees in your zip code at any time of the day within 30-60 minutes. This is pretty cool and pretty exciting stuff, yet predicting on-the-job performance is still done with pseudo-science, bias, gut feelings and thumbs.
What the heck is happening? I’ve been thinking about hiring when I first become a hiring manager 45 years ago and 1,500 hires later (as a recruiter). Mostly I’ve been thinking about why hiring processes are still as outdated as they were when I became a recruiter (in 1978). Here’s what I find puzzling:
- Why do recruiters need to review 150 resumes to make one decent hire, but only 3-4 referred candidates to make one great hire?
- Why do we hire people we know based on their past performance and potential, but we hire strangers based on their past experience?
- Why do hiring managers want to hire candidates who can hit the ground running, but the best people who can hit the ground running want to run on different tracks?
- People who see the job as a career move don’t need as big a salary increase compared to people who are accepting ill-defined lateral transfers. So, why do we negotiate the salary, the location and job title as the condition for having a conversation with someone?
- While we all want to hire more diverse people, why do we expect them to have the same skills and experiences and look and sound like everyone we’ve already hired
- Why is it we still can’t measure quality of hire?
- Why do people in the top half have less or different experience than those in the bottom half? That’s how they got into the top half.
- Why do we design the candidate experience largely for candidates we’re not going to hire, rather than for the candidates we do want to hire?
- Why do recruiters and tech vendors get excited about doing the wrong things more efficiently?
- Why are we still fighting the war for talent?
Since I’ve had a chance to track the performance of hundreds of people who were and weren’t hired for many years after the hire, one problem is that there are just too many cooks and too many recipes being used on how to hire people. Based on this tracking, I’ve summarized the best and worst predictors of on-the-job performance in this table.
While the list is near perfect, getting everyone to use an evidence-based process to collect these best predictors and avoid the worst ones turns out to be the biggest challenge. (I’ll be discussing this concept at my next webcast.)
Predicting the weather requires the same insight and technical ability as rocket science. Predicting quality of hire requires a year or two of statistics, a good understanding of human nature, recognizing the difference between steps and systems, and an agreement by everyone who is involved in making the yes/no decision that evidence trumps bias, emotions, thumbs and gut feelings.