Recently we were approached by a long-time client who stumbled upon a new challenge which many might not have even considered - they wanted to predict the probability of their employees leaving, in particular, senior employees that are difficult and expensive to replace. Their reasoning was that since this was a senior position, they wanted to have a pipeline of people to fill these roles, but weren’t sure how to structure this pipeline. The biggest challenge was that they needed to know how many people would leave before year’s end in order to properly size their promotion/hiring pipeline.
What was most surprising about this request wasn’t actually part of the initial question, but the skeptical responses from employees charged with solving this problem.
Several of them, including an analytics supervisor, admitted that they had never considered churn outside of looking at customers.
In reality, churn can be used for any circumstance where there are a number of people or items which need to be replaced from time to time, for example office equipment such as laptops. It might seem strange to think of ‘laptop churn’ but from a data scientist’s standpoint, the two problems are nearly identical, they just have different attributes that lead up to a ‘churn’ event. I’ve digressed though, ultimately most businesses have no need to model their laptop churn as it’s a rare event with relatively low associated costs. On the other hand, replacing an employee, as replacing a loyal customer, can cost a great deal of time and money to replace.
While hiring pipelines were the reason presented, there are several other reasons these kinds of predictions can be quite useful.
- Hiring and training a new employee is almost always more expensive than retaining a current one, this is especially true in an industry with skilled positions that require intimate knowledge of complex systems to perform their jobs. Knowing that one of your key employees is a churn risk could start a conversation about what the company can do to keep them.
- A historical analysis of employee churn risk can uncover shifting trends in employee behavior that can help you better understand your business and plan for the future
- Your predictions and analysis might uncover some systemic problems with certain departments or roles which are much higher churn risks. Knowing this can allow you to examine why that might be, and potentially mitigate factors that cause this increased risk.
Employee facing analytics tend to be relatively simple to create, as employee data tends to be very descriptive and complete. Information such as age, time in grade, compensation, seasonal reviews, disciplinary actions, and promotions/job changes are a part of most any employee database. Having access to complete, well-defined data makes for better predictions and takes less time to process.
When it comes to the success of your company, an understanding of internal churn is just as important as external churn. Few businesses trust predicting customer churn based on gut feelings, and yet surprisingly this is how most identify internal churn risk. If your business relies on skilled labor, knowing the probability of your employees leaving is vital to keeping your business running, should that be left to gut feelings?
For more insight into the predicting customer churn, contact a data and analytics specialist at 813.265.3239 or email@example.com for more insight.