Three Turnover Methods: Details Matter

Turnover is a term that every manager in business knows and every company of moderate size and up measures. It is a measure that is represented as a rate/percentage of workers that leave a company within a defined time period. Turnover is an easy concept to grasp but can be elusive to measure.

Three of the most important factors in measuring turnover can be gleaned from the definition. First, we need to determine “what” we are measuring as the baseline or denominator in the equation. Are we measuring employees or employees and contractors? Do we include interns?

Second, are we measuring voluntary or involuntary turnover? Finally, what is the time period for which we are looking to measure turnover? Understanding the business need for performing this calculation and anticipating what further cuts you may be asked to make is fundamental in determining the pool of workers that represent the denominator.

Once we’ve established the pool, the turnover type and the time period, we can address the different ways it can be represented. There isn’t one method that is more correct than another statistically, so we need to ensure that we understand the story we’re telling and that it’s done consistently. Once we’ve determined our method of measuring turnover, it’s critical to maintain the same method so that trending is correctly represented. Given that many HCM and HR Analytics systems have turnover metrics built in, we may also choose to align with their definition to ensure advanced analytics can be tied back to other HR visuals that the team utilizes. 

Method 1: Using the Average Worker Population for the Start and End of the Time Period

This is the method utilized by many HR systems and is the most straightforward. 

In this scenario, we need only three figures. We will take the defined worker population at the start of the time period, the population at the end of the time period and the count of the population that left during the time period. In this scenario, we divide the exiting population by the average of the population on the beginning and end dates. Mathematically we’d represent that as:

Average Population = ([Population at Start] + [Population at End])/2

Turnover Rate = [Population Exiting] / [Average Population]

This is a quick and easy way to calculate turnover but does present some challenges. In many industries, there are cyclical changes or business restructurings that can materially change the total population at any given point. Given that we are taking ONLY those two numbers (the population at the start and end of the time period), this could provide less insight into the turnover of the workforce than we would hope. 

Conversely, knowing there are cyclical changes could ensure that we always take the start and end dates from the same point in the cycle. Where the first scenario skews the numbers, the second helps to guarantee a consistent view of the workforce. For example, if a landscaping company had 100 workers at the start of their season in April but laid 25% of the workforce off for the winter months, we would see a much different turnover percentage based on when we chose the start and end date. If we measured when the company had hired in the seasonal staff and ended at the end of the season, we would see this in our turnover calculation:

Average Population = (100 + 75) /2 = 87.5

Turnover rate = 25 / 87.5 or 28.6%

However, if we did that same calculation in June and assumed 5 people had left for various reasons, we would get a 5.1% turnover rate. 

Understanding the latter value could be incredibly important to a business. In this case, the owner may choose to hire over his work capacity by a few percent knowing that he will have future attrition or only take bookings to 95% of his workforce’s capacity.

There is one more challenge with this method which is quite significant. If you were to present the monthly turnover values and yearly turnover values, you would discover that adding 12 monthly values doesn’t equal the same number as the annual calculated value using the formula above. One client of mine, an HR analytics team, found they spent more time explaining why adding the monthly values did not equal the annual value than they spent actually discussing turnover. This was especially true when presenting to finance leaders who are used to summing up monthly values on financial statements. After three years of this challenge, the team changed their turnover formula to match Method 3 below.

Method 2: Using the Average Worker Population Throughout the Time Period

To help mitigate cyclical or unexpected changes in the workforce, many companies choose to use the average of a predetermined population (usually monthly) during the period being measured. In Method 1, if you were measuring turnover between January 1 and July 1, you would take the population on January 1 plus the population on July 1 and divide that by 2 in order to get the average population (denominator value).

However, if you had a surge of temporary help on May 1 for the summer, basing the average population on only the number of workers on January 1 and July 1 wouldn’t be representative of the population during this six-month period. For this reason, in Method 2, we add the worker populations for each month covered (January, February, March, April, May, and June). In this scenario, the average population value better represents the population over that time frame.

Formulaically we’d solve this with:

Average Population = ([Population January] + [Population February] + [Population March] + [Population April] + [Population May] + [Population June]) /6

Turnover Rate = [Population Exiting] / [Average Population]

Our landscaping company above knows that some people quickly find that the hard work required on the job isn’t for them, so the company loses more people at the beginning of the season. By the end of the season (November 1), workers rarely leave. Knowing this, the landscaping company chooses to use Method 2 to calculate its turnover rate. The company lays off 25% of workers on November 1. 

This method is also far from perfect. If our business owner effectively replaces the exiting employees in the same month as they leave, then the denominator would remain at 100, thereby minimizing how much of the turnover is represented. To keep things simple, if the company lost 27 people, but was able to replace them before the close of each month, then the average population in the denominator would be 100.

If the company had three openings each month and did not replace them in the same month, then the average worker population in the denominator would be 97.

Another issue with this method is that a month that is out of character with other months being measured (e.g. flu hits the workforce substantially one year) can significantly affect the turnover rate calculation. 

Finally, like Method 1, this approach will not yield the same numeric value for an annual calculation as summing 12 monthly calculations.

Method 3: Using the SHRM and CIPD Definitions

Method 3 helps in scenarios where companies are looking at a running total throughout the year. In the case where turnover is either very consistent or cyclically predictable, Method 3 helps build useful visualizations that represent how an enterprise is tracking against projections or goals for the year. This method matches the monthly and yearly turnover definitions defined by SHRM (USA) and CIPD (UK). It has the added benefit that the turnover values for the year or quarter will always match the sum of the monthly values.

In Method 3, we define the sub-time periods (usually monthly) and use the formula in Method 1. Then, as we move through the primary time period (usually yearly), we add the percentages together. This creates a visualization that is mathematically very similar to Method 2 at the conclusion of the primary time period. It serves as an excellent way to track toward turnover goals or annual projections. Formulaically this would be represented as:

Monthly Turnover Rate = [Population Exiting During the Month] / [Average of the Population at the Start and End of the Month]

Annual Turnover Rate = Sum of all 12 monthly turnover calculations

The Correct Definition for Retention Rate

The first point we need to understand is that retention is not 100% - [Turnover Rate]. Many companies incorrectly represent retention by simply stating that if turnover is 20%, then retention is 80%. No matter which turnover method is chosen above, that statement is not correct.

Retention is the measure of the population that exists as active at the start date and is still active at the end date. Retention is a number that, if explained well to business partners, can be insightful and complementary to turnover.

Determining retention is easy from a formulaic perspective but may not be as clear if there are movements in and out of the data cut such as department to department moves or seasonal help such as interns. As with any of the approaches, understanding your population and your audience is key to providing meaningful insights.

To determine retention, we need two pieces of data. We need individually identified workers at the start of the time period and we need to know which of those identified workers still exist at the end of the time period.

Retention helps eliminate the mathematical confusion that hiring or transfers bring into the equation by simply asking, “which employees are still here after (x) periods of time?” Retention helps unmask issues that turnover doesn’t expose. For example. if a company has a robust, quick hiring process or material growth, the retention measurement can show insights that turnover would not reveal.

Formulaically, we take the ID’s of all individuals at the start date and look up which ID’s exist at the end date. Then divide the end date number by the start date number and multiply by 100.

For example, suppose we select a time period of one year and we have 125 employees at the start of the year. At the end of the year, we find that we have 150 employees, but 35 of those were hired during the year.

We can determine the number of employees that are still with the company by calculating:

Number Retained = [The number of employees at the end of the year] – [Number of new hires during the year]

= 150 – 35

We can then calculate the retention rate as:

Retention Rate = 115 / 125 * 100

or 92 %

For our landscape company, the owner may find it extremely valuable to know the number of employees that return from year to year. For example, knowing the percentage of staff that is retained from one season to the following year’s season will provide insight into the hiring needs of the company for future seasons. 

In Summary

Based on considerations such as business culture, HRIS system capabilities and configurations, available analytic tooling and cyclical/seasonal hiring, one of the three turnover definitions listed above may be more appropriate than the others and will provide the insight you need.

Using more than one definition for turnover is not advisable. Using a consistent and well communicated method of turnover in conjunction with a retention measurement will provide analysts and business leaders with insights that are both actionable and understandable.