How to Calculate Employee Turnover (And What to Do With the Number)

Almost every company that measures turnover measures a single number, and that number almost never helps decide anything.

Not because it is miscalculated. It blends things that are not alike: the person who left for a better job, the one you let go in the first month, and the one who stuck around three years and quit the day someone else got promoted. All three land in the same denominator, and all three mean different things.

The formula

Turnover rate for a period is:

(Separations in the period ÷ Average headcount for the period) × 100

Average headcount is the headcount at the start plus the headcount at the end, divided by two. If you started the year with 80 people and ended with 100, your average is 90. If 18 people left during the year, your annual turnover is 20%.

Two details that move the result more than they look like they should:

• Use average headcount, not ending headcount. If you grew fast, dividing by the ending figure gives you an artificially low number. It is the most common mistake.
• Define “separation” before you count. Do finished internships count? Project contracts? Pick a rule and hold it, because the value of the number lives in comparing it to itself over time.

The three cuts that actually help

The global number is for the board report. To make decisions you need to split it.

Voluntary vs involuntary. If someone left because they wanted to, you have a retention problem. If you let them go, you have a selection problem. Two different problems, two different teams accountable, and adding them together erases the information.

Early turnover. Separations within the first six or twelve months, measured separately. It is the most brutal metric in HR. An early exit is almost always a hiring decision that went wrong, and the cost is already sunk. If your overall turnover is 18% but half of it happens before the one-year mark, your problem is not climate. It is who you are letting in.

Regretted turnover. Of the people who left, how many did you want to keep? This cut takes an opinion, not a data point. Almost nobody tracks it. It is the most valuable of the three. A company at 25% turnover where almost none of it is regretted is doing better than one at 12% where the best people are the ones leaving.

What to compare against

The next question is always “is 20% a lot?” It depends on the sector, the geography, and the kind of role. Any benchmark that ignores those three will make you feel good or bad for no reason.

Two comparisons do hold up. Against yourself, quarter over quarter, with the same definition. And across teams inside your own company in the same period. If one team doubles everyone else’s turnover, with the same labor market and the same employer brand, the variable that changed is inside that team.

The cost, to size the conversation

Replacing a person does not cost what the search costs. According to SHRM, total replacement cost can reach up to twice the role’s annual salary, depending on the position.

That figure includes what nobody puts on a spreadsheet: the time of the team doing interviews, lost productivity while the seat is empty, the new person’s ramp, and the effect on everyone watching that empty seat. If you have 90 people and 20% turnover, that is 18 replacements a year.

You can run that math on your own numbers in the ROI calculator, or see the full breakdown in the real cost of a bad hire.

What to do with the number

Turnover is a symptom, and it arrives late by definition. By the time you measure it, the person is gone.

If most of it sits in early turnover, the problem is at the entrance. The lever there is not more benefits or better climate. It is measuring fit before the offer, against the role and against the culture the person will actually meet. A candidate who is flawless for the job can still be a seven-month exit if what your company rewards is not what keeps that person going.

If most of it sits with tenured people, the problem is development or management, and an entry assessment will not help. What helps is knowing which of the people already inside can take the next seat, which is what we covered in internal mobility.

Both cases start the same way: by splitting the number. A flat 20% says nothing. A 20% with half of it before the one-year mark, concentrated in two teams, says exactly where to look.