Am I that predictable? How AI can predict when you’re about to quit with 95% accuracy
For many, quitting a job is one of the hardest decisions they’ll make. For others, it is like changing a pair of socks. What we all have in common though, no matter how many self-help books we imbibe, is that we are predictable.
IBM’s Prediction Program
IBM, using the famous AI computer Watson, has created a ‘predictive attrition program’ which can predict when employees are about to leave the company. Its accuracy rate: 95%. It can also suggest further actions by HR staff to stop the person from leaving.
The program is being championed by IBM CEO Ginni Rometty, who stated in March 2019 that the program had, so far, saved $300 million USD in attrition costs. These were stemming from costs in relation to things like training and loss of income. This is not to mention the institutional knowledge that employees take with them, but also the morale damage to other employees of losing a colleague.
IBM now offers the solution to external clients. While they have not provided the secret recipe as to how the patented program works, they have revealed some of the more obvious data sets on which it operates. These are:
Time since last promotion compared to peers
Time since last change in employer
Compensation compared to peers
For any company or HR professional, it may be time to start measuring these factors to see where the pain points and imbalances, and then seeing what good can be done with the data. For example, with overtime and commute time, a flexible approach could be considered. As for promotions, it could be a matter of either promoting the person, or even mentoring the employee into a position where they can be promoted. Businesses should pay attention to the insights that can be pulled from the data: it can be used to provide a better workplace for employees.
Felix Blumer is a Consultant at Rutherford, the Legal and Compliance executive recruitment specialists.
Contact us for a confidential discussion about your search at firstname.lastname@example.org or see our vacancies.