Presentation Topic:
Predictive Attrition Using Survival Analysis

More about presentation:
Attrition metrics can often under or oversell the sentiment of your employees. They are impacted by external labor market factors and are not consistently predictive: 10% turnover for the last 2 years does not predict 10% turnover in the upcoming year for example. Survival analysis gives an alternative and complementary view to attrition. Borrowed from epidemiology, this analysis looks at the likelihood of churn as an employee reaches tenure milestones with the company. It shows when your teams are at-risk and leads to a much more meaningful intervention or informed succession strategy. It's also adaptive: with new data being fed, this ML based approach can adjust expectations based on timely departures and can be modeled for different labor-market scenarios. We have used this approach at TIBCO to increase our annual turnover forecast accuracy from ~60% to ~80% and it allows us to see hotbeds of attrition before the trend emerges. In this brief presentation, I'll introduce the concept of survival analysis, demonstrate it with a sample dashboard, discuss its applications among trends such as the Great Resignation and then be available for questions.

About speakers:
Alex holds a Ph.D. and has over ten years of academic and industry experience in applied data science and machine learning in chemistry, biotechnology, and physics.

Presentation Topic:
AI HR Tech at large enterprise