CASE STUDY
Corporate Bank
Predictive Model for Employee Resignation
Key facts
Total number of employees: 47,000
Employees analyzed by Yva: 300 for the pilot, 8,000 for the production stage
Period of analysis: November 2018 - February 2019
Number of emails: 40,880,000
Data lake size: 2 Terabytes
Server number of stations/processor cores: 4 stations/4 cores each
Time spent to process the data: 59 days
The challenge
The bank has been nominated one of the best employers in the geography and is investing heavily on employee development. In fact, minimizing key employees churn is one of the most important tasks on the current agenda of its leadership team. The bank conducts regular 360 surveys and pulse surveys, but it doubts its real value in defining HR roadmap as well as to take timely actions.
The pilot
In the framework of the pilot project, the bank called Yva.ai to develop a predictive model of employee resignation and check its validity on a training sample of 200 dismissed employees. Following the historical test, the bank also launched a subsequent test for 300 employees (mixed data on employees and dismissed employees).
The solution
Development of a predictive model for employee resignation based on AI-driven analysis of survey data and a digital footprint for 8,000 employees in the central office.