Thanks for this great article! This is indeed an important subject these days and as you pointed out, there is no good reliable source of these benchmarks. But I’m curious about your closing statement about ‘Turnover Calculation’.
As far as I’ve seen companies use either the average or ending headcount or starting headcount for numerators as they like/according to their availability of data.
Why would you say that starting headcount is the preferred way? Is there any literature or articles recommending on over the other that I should read?
I think there are two reasons why companies started using the average of beginning and ending headcount instead:
1) Quick quits who might have started after the beginning of the period
2) Acquisitions and internal reorgs or shuffles of teams which will impact the headcount numbers and turnover numbers of the departments
Also usually the termination numbers are usually just the count of terminations recorded in a period on the HRIS system. So while we know that there might be X terminations in a month instantly, knowing the number of terminations ( employees who have left) out of employees from the beginning of the period usually requires some manual work.
I’m sure most if not all of the sources that were mentioned in the article just considered all terminations in the month instead of the terminations of only the employees who were active at the beginning of the period.
Probably there is no similar phenomenon of acquisitions in the customer churn example that you have mentioned but I’m curious to know how (if) they manage the quick quits problem with customer data using starting count of the period.
Hope I’m not wasting your time, just trying to learn some good practices from you.
Not at all! I am actually enjoying the engagement and the stimulating conversation on this subject and others. I actually think that the best way to learn is to speak about this stuff with each other! So thank you!
The post actually got a bit of heat on LinkedIn as a lot of people are using averages in the denominator, but it sounds like not a lot people actually know why they are doing that. Haha!
I think your points are interesting!
1) I can see your point about people who have started after and then left. Though, I would say the bias in the denominator is going to be minimal given there are likely only a few people like this in the organization. One way of addressing it is shortening the period. Usually, we think about turnover in Years, but we could also look at it in months and estimate it this way. Note that averaging here would also not address the problem if they left in the middle of the period.
2) I am totally okay with all the changes. I think acquisition is probably the biggest culprit in here. But, if I wanted to see turnover, I would want to see only the headcount before the acquisition. After all, I wouldn't go to the board with turnover numbers that consider the acquisition. I would rather show a story of turnover before and after acquisition. Similar to other metrics.
Thanks for this great article! This is indeed an important subject these days and as you pointed out, there is no good reliable source of these benchmarks. But I’m curious about your closing statement about ‘Turnover Calculation’.
As far as I’ve seen companies use either the average or ending headcount or starting headcount for numerators as they like/according to their availability of data.
Why would you say that starting headcount is the preferred way? Is there any literature or articles recommending on over the other that I should read?
Hey Raj,
This is an awesome question and it has inspired me to write this LinkedIn post:
Thank you for the inspiration and what do you think?
Post:
--
Silly way to calculate turnover:
Averaging starting and ending headcount for your denominator.
There is no reason to do this:
• Logically, turnover is how many people have left since the start of the period
• Turnover rate is the % of people who left since the start of the period
Mathematically: People leaving / Starting HC *100%
You should not use ending headcount in the denominator because you will ask people who have left how many are still here. Which makes no sense.
You wouldn't calculate your customer churn rate with averages.
So, why would you do it with employees?!
I think there are two reasons why companies started using the average of beginning and ending headcount instead:
1) Quick quits who might have started after the beginning of the period
2) Acquisitions and internal reorgs or shuffles of teams which will impact the headcount numbers and turnover numbers of the departments
Also usually the termination numbers are usually just the count of terminations recorded in a period on the HRIS system. So while we know that there might be X terminations in a month instantly, knowing the number of terminations ( employees who have left) out of employees from the beginning of the period usually requires some manual work.
I’m sure most if not all of the sources that were mentioned in the article just considered all terminations in the month instead of the terminations of only the employees who were active at the beginning of the period.
Probably there is no similar phenomenon of acquisitions in the customer churn example that you have mentioned but I’m curious to know how (if) they manage the quick quits problem with customer data using starting count of the period.
Hope I’m not wasting your time, just trying to learn some good practices from you.
Hey Raj,
Not at all! I am actually enjoying the engagement and the stimulating conversation on this subject and others. I actually think that the best way to learn is to speak about this stuff with each other! So thank you!
The post actually got a bit of heat on LinkedIn as a lot of people are using averages in the denominator, but it sounds like not a lot people actually know why they are doing that. Haha!
I think your points are interesting!
1) I can see your point about people who have started after and then left. Though, I would say the bias in the denominator is going to be minimal given there are likely only a few people like this in the organization. One way of addressing it is shortening the period. Usually, we think about turnover in Years, but we could also look at it in months and estimate it this way. Note that averaging here would also not address the problem if they left in the middle of the period.
2) I am totally okay with all the changes. I think acquisition is probably the biggest culprit in here. But, if I wanted to see turnover, I would want to see only the headcount before the acquisition. After all, I wouldn't go to the board with turnover numbers that consider the acquisition. I would rather show a story of turnover before and after acquisition. Similar to other metrics.
Thank you for your thoughtful comments!!!!