12/14/2011 11:30 am - 12:30 pm
Finding Predictive Patterns of Employee Turnover
(Lawrence Livermore National Laboratories)Soo Tan
(Lawrence Livermore National Laboratories)
Employee turnover poses significant risks at the Lawrence Livermore National Laboratory (LLNL). During this session, we will present how we use predictive modeling techniques to find high risk patterns of employee turnover. The output of the technique we will propose is very visual and intuitive, and allows us to answer important questions.
• Are high risk employees in critical roles?
• Do high risk employees have skills that are difficult to replace?
With many variables effecting employee turnover, it is important to identify which variables are important, and what are the drivers of attrition. The session will cover a powerful predictive modeling technique that can sift through large amounts of data and variables to find high risk patterns of turnover. The output of this technique is visual and intuitive, which make it easier to understand and communicate to your organization. This session will also include an introduction to data mining and predictive modeling. How predictive modeling is commonly applied outside of HR, but how you can take that same concept and apply that to your HR data. We will also show you how data mining and predictive modeling has been applied in different scenarios at the Lawrence Livermore National Laboratory, and our approach to managing our data using automation techniques in Excel.
In the end, predictive modeling provides HR with a laser precision focus on employee retention by isolating potential problem areas of turnover. This enables the effective development of retention plans and strategies designed to prevent our best and brightest from leaving.