Much has been written about customer churn - predicting who, when, and why customers will stop buying, and how (or whether) to intervene. Employee churn is similar - we want to predict who, when, and why employees will terminate, in order to avoid attrition.
In many ways, it is smarter to focus inward on employees. For one thing, it is far easier for a company to change the
The most important difference between employee vs. customer churn is that a business chooses to hire someone. So, predictive analytics can play a formative role in the employee mix. There is more at stake for both employer and employee. This person will literally be the face of your company, and collectively, employees produce everything your company does.
Customers provide profit right away, so customer churn analytics tries to keep the train rolling a bit longer. We will see that employee churn analytics is more like trying to get the train to run long enough to provide any value at all.
This article will walk through a proven, holistic framework for measuring the cost and attrition of employee churn. It enables the calculation of employee lifetime value and enables the use of advanced predictive analytics to solve the churn problem.
The Employee Value Curve
Employee Cost and Benefit
In this stylized example we show the cost and benefit of an employee in dollars. It is simple enough to document the hard costs that make up the initial stages of employee life-cycle: recruitment, training, and on-the-job
Hard, documented costs, such as the cost of training, are preferable over less defensible “soft” costs like morale. A multi-year view should include wage increases; beyond a few years it will be necessary to include inflation and the time value of money.
Employee benefit, the amount that an individual contributes to revenue, can be reasonably estimated for sales or productive roles. For roles with a less direct connection to corporate gross revenue, value can be apportioned by geographic or departmental lines.
Employees are generally not productive their first day - often it can take months to get up to speed. Different jobs will have different curves, but we show a typical S-shaped learning curve, followed by a gradual increase in benefit. In this hypothetical job, an employee takes a year and a half to ramp up to full productivity.
Figure 1: Cost and Benefit for One Employee
Figure 1 shows a stylized cost/benefit plot for one employee across five years of tenure. At time zero, costs are high - an expensive recruitment process, administration, training, supplies are all above the normal flow.
We see that employees in this role start to provide daily value after 9 months, when the benefit curve crosses costs. This is called the “Daily Breakeven.”
The cost and benefit curves cross into a “quantitative scissors” - the left “red zone” is an inescapable reality of doing business. To decrease the overall costs due to employee churn, something has to budge on these curves.
Cumulative Employee Value
Figure 2: Cumulative Value for One Employee
We sum these costs in Figure 2, showing a cumulative value in the familiar “hockey stick” shape. The plot shows the total net cost or benefit accrued by an employee if they get to a specific tenure.
This curve, the employee value at time t, is a valuable business object. Once the “debt” from training and startup is paid off, the line crosses zero. This is called the “Cumulative Breakeven.” The red region shows that terminations before this point are a cost to the business.
In this stylized example, the employee starts providing daily value after 9 months, and does not pay back the startup “debt” until after 24 months. An employee who leaves at 9 months represents a net cost of over $25,000 in lost startup costs. By comparison, in our engagements we often see impressive attrition after just 3–6 months.
Normally, we rely on aggregated figures to calculate an average curve and breakeven for each role. Advanced analytics can use transactional sales or production figures to calculate individual curves and breakeven for each employee.
There is a sub-field of
Attrition: the Employee Survival Curve
With a cost curve in hand, we now turn to evaluate attrition. We hope to illustrate a far more intuitive and useful visualization than the popular business metric, annual attrition.
The Attrition Histogram
The histogram is a useful first glance of how attrition plays out in an organization. It is easy to produce from simple HR records, and the graphic already tells a deeper story than simple averages or turnover rates.
Figure 3: Histogram of Attrition
Figure 3 shows a basic histogram of tenure. The horizontal “X” axis maps out the number of years tenure in a specific role. The vertical “Y” axis shows the count of how many employees had that amount of tenure. We can see a stack of early departures in the first 9 months, then another rise later.
The Hazard Curve
But this is deceptive. Histograms are messy, and don’t show cumulative effects.
Figure 4: Hazard Curves - Daily Probability of Termination
In this simple example we have two hypothetical regions: Chicago is blue, with longer tenure. New York is brown, with high turnover in the first year.
This is just an example - the real world has more nuance and ambiguity. This same analysis can apply with one, two, or many groups of employees. Likewise, the model works for continuous variables like employee engagement. Good machine learning algorithms can identify useful clusters.
The Survival Curve
Employee attrition falls into the same class of “survival” problem as machine failure rates or medical research. This domain has brought us solid statistical innovations, including visualizations known as the “hazard curve” and the “survival curve.” The “bumps” in Figure 4 are in fact “hazard curves” - the hazard of someone terminating at time t.
Figure 5: Survival Curves - Probability of Reaching Tenure
The survival plots in Figure 5 use the same data as the hazard plots, but put it all together into one holistic sum. Across the horizontal ‘X’ axis we see tenure, from zero to many years. On the vertical ‘Y’ axis we see the probability of an employee surviving to that tenure. In fact some people don’t show up to their first day of work - but on day 1, survival is near 100%.
The normally-reported annual attrition is here - at the one-year mark, we see 86% survival for the Chicago curve, which subtracts from the top for 14% annual attrition. Likewise, the New York curve shows 58% survival at one year, which subtracts to 42% annual attrition.
But, one year is somewhat arbitrary. Remember that our breakeven point is 2 years - arguably this is a more useful threshold. At this 24-month breakeven point, we see 65% remaining from Chicago, but only 21% from New York.
Please note that there are special techniques, notably the Kaplan-Meier Estimator, that must be performed to properly account for current, unterminated employees. These methods are built into modern statistical software, and can even be done in
Employee Value + Survival = Two Fundamental Curves
We believe that survival curves, along with the value curves developed above, are two vital tools that should be produced by every HR or Finance department for every significant role. They capture the essence of the role’s impact on the company, and are the basis of powerful calculations.
Lifetime Value of an Employee
One important application of these curves is to obtain the lifetime value of an employee. Many have worked out the lifetime value of a customer to 5 decimal points, but few have ever considered the lifetime value of an employee.
We can look up the cumulative value at an arbitrary date, say 5 years. The chart shows that an employee at that tenure is worth $78,851. However, not all employees last 5 years - only 18% from Chicago and 0.2% from New York.
Figure 6: Comparison of Employee Lifetime Value
Borrowing a technique from finance, we
In this sample, a Chicago employee’s lifetime value is $31,487, while the New Yorker is a loss with -$9,343. A far cry from the potential $78,851. These groups have the same cost curve, but different survival curves - which makes a huge difference.
Scenario Planning
Employment, and business in general, is not a laboratory environment. We don’t get do-overs for failed scenarios, and our ability to “try things out” is limited. Customer analytics is slightly more amenable to A/B testing, just because the relationship is thinner, and there are many customers.
With this model of lifetime value, we can simulate the impact of programs. What happens if we:
- Increase wages by x%, and assume it will shift survival curves up by y%?
- Reduce wages by x%, and assume it will shift survival curves down by y%?
- Spend $x more on training, and assume it will accelerate learning by y%?
- Spend $x more on coaching, and assume it will help New Yorkers stay y% longer?
- Hire x% more employees in Chicago?
Predictive Analytics Baseline
Sensitivity analysis of these scenarios will quickly reveal that better employee selection is the key to higher lifetime value. All of the other measures (except reducing salary, which is often impractical) have lesser effects.
In predictive hiring, we look for signals that reliably show a candidate is more likely to stay longer in the role. The result of a good model is better employees and a higher average lifetime value.
Conclusion
These metrics are especially valuable for high-volume, high turnover roles that need attention. They are the measuring stick for improvements to hiring selection, engagement efforts, and performance improvements.
Is it time for your organization to begin using a predictive analytics approach to your high volume, high turnover roles?
by Pasha Roberts, Chief Scientist & Co-founder, Talent Analytics, Corp.
Pasha Roberts is Chief Scientist and Co-founder at Talent Analytics Corp., a company that uses data science to model and optimize employee performance in areas such as call center staff, sales organizations and analytics professionals.