Predicting Employee Churn with Data Mining

Use Data Mining to Identify Employees at Risk of Churn

For organizations that have a lot employees that are in high turnover positions predicting employee churn with data mining can help to reduce and retain top talent.  By collecting data on employees and then building a predictive model using employees that have left the organization.  You can create models that will give you new insights into the characteristics of employees that at risk of leaving.  By having this score you can then match up the performance of the employee to determine options to keep your talent and prevent them from leaving.  For example a call center, sales team, or temporary agencies could all benefit from building models to determine why employees are leaving.

Build the Model

For this example model, I used a sample dataset that contains various information for each employee.  This data contains both customers that are currently employed as well as customers that have churned.  The example data is shown below:

Data Mining Employee Churn

Review the Results

After building a few different models, I settled on a decision tree that let’s me visually see how the characteristics that lead to an employee staying or leaving. My model profiled the data and determined the following fields were the 5 most important factors that determined churn:

  1. Environment Satisfaction
  2. Job Satisfaction
  3. OverTime Pay
  4. Relationship Satisfaction
  5. Stock Options

Now with the model built I can apply the model to the active employee base and get insights into employees at risk of leaving and what is the path to determine their score.

Data Mining Employee Churn Prediction

Interested in learning more…contact me today – dwilson@cdoadvisors.com

Comments are closed, but trackbacks and pingbacks are open.