Predictive Analytics Applied to Human Resources
Updated: Jul 9, 2019
Does your company suffer from low employee retention? Below is a great example of how HP overcame that same obstacle.
Hewlett-Packard is an American IT company headquartered in California, with satellite offices worldwide. The company has come a long way since being founded in a one car garage, as it is now one of the leading manufacturers of personal computers and computation services for small and medium sized businesses throughout the world.
In 2011, a couple of HP managers noticed that attrition within the sales team grew to be as high as 20% of all employees within a certain period of time. They called for a human resource study and the results changed the way the company evaluates employees forever.
The managers that took on this project were named Alex, Halder and Dey. They met at an IT conference and decided to make a difference. Together they studied employee behavior and career patterns over the course of two years relative to a sample of 300 salespeople spread across all countries. Some of the variables studied were days requested off, sick days, promotion, salary, job ratings and job rotations.
After examining this information, Alex, Halder and Dey were able to predict employee flight risk with 90% accuracy. They began to implement tests and began calculating and managing those employee’s compensation.
The three created a system that would change (and improve) the quality and ability of their human resource departments forever. This multi-level approach only took a few months to completely integrate. They had dubbed this newfound metric the individual ‘Flight Risk Score’.
HP was now reaping the benefits of this project because they realized how much added value is created by employee retention. The most prominent example was that much of their overhead was expended due to the need of getting each new hire trained and ready to sell such intricate products. They also factored in the level of productivity and associated costs.
This new system of action was implemented in daily human resource operations and continued to gain traction as HP acquired many smaller organizations.
This model was able to identify $300 million in potential savings for HP. 40% of the employees with a high ‘flight risk score’ consisted of 75% of all employees who quit HP.
Predictive modeling is not just about data. It’s about gathering and analyzing accurate historical information, machine learning and many (independent) variables interacting with one another to make the best possible decision.
If you’d like to utilize the potential of predictive modeling, contact us today.
Let us make your data Tangible.