Attrition Rate for employees, given that there were none for some months

[314ql]

Hello!
I am trying to calculate attrition. It is a simple formula: I need to calculate the sum of people who left company and divide it by the average number of employees during this time.
I did it, but I am not so sure what to do if e.g. in one division there were no employees during first months.
So now I am calculating attrition for the whole 2022 year and I noticed one division in which from January to August there were no employees.
The average therefore is very low and even one person who left the company gives weird results.
What I am asking is what would you do? Would you exclude the months where there were no employees from calculating the average?

 

[JeffM]

To give you really good advice, I'd need to know who the audience is and how the data is going to be used.

If your audience is comfortable with math, I'd just say that results are not meaningful for divisions that had fewer than x man-months in the year. If you have a two-person division, you can only get three possible answers: 0%, 50%, or 100%. It is unlikely that any of those numbers says much useful.

What you can do is to show the rate for the company and for the larger divisions where the numbers are likely to be meaningful. If you are worrying about seasonality, you can follow the same logic: if there are months with no employees, aggregate into longer periods such as quarters in addition to the years as a whole. Always start with the largest aggregate. Stop breaking the aggregate down once the base becomes too small to make much sense from the results.

You need to remember that averages are not always informative. And the smaller the numbers that you are working with, the less need for averages. If I am trying to make sense of thousands of numbers, averages, deviations, ratios etc. are essential. If I am trying to under stand what eight numbers are telling me, mushing them up into an average is going to result in a number that may hide what is important. Averages, ratios, etc are tools; use them when they are helpful but not otherwise.