[Turnover and attrition](https://www.teampredict.ai/glossary#turnover) # What is employee turnover? Also called employee churn, labour turnover Definition Employee turnover is the movement of people out of a company over a set period, counted as the employees who leave and are replaced. Last updated October 13, 2026 ## Employee turnover counts the people who leave and get replaced. Turnover is a flow, not a snapshot. Headcount tells you how many people work here today. Turnover tells you how many walked out of the door during a month, a quarter or a year, whatever happened to the total. That is why a company can grow and still have high turnover. If 30 people leave and 40 join, headcount rises by 10 while 30 roles were vacated, recruited for and ramped up again. ## Turnover matters because every exit carries a cost. It is the headline number executives ask for, and it sizes every downstream cost: recruiting, onboarding, lost output and lost institutional knowledge. The total on its own is a blunt instrument. It mixes resignations with layoffs and regretted exits with welcome ones, so the useful work starts when you split it, as the sections below show. ## You measure turnover as separations over average headcount. Separations during the period divided by average headcount for the same period, expressed as a percentage and usually annualized. Formula Turnover rate = separations ÷ average headcount × 100 Average headcount is usually the headcount at the start plus the headcount at the end, divided by two (see [headcount](https://www.teampredict.ai/glossary/headcount)). A [separation](https://www.teampredict.ai/glossary/separation) is any end of employment in the period. Decide up front whether contractors, interns and internal transfers count, and keep that rule fixed. ## Here is a worked example with round numbers. Take an invented company that starts the year with 200 employees and ends it with 220. During the year 30 people leave. Average headcount is (200 + 220) ÷ 2 = 210. Turnover is 30 ÷ 210 × 100, about 14.3% for the year. _Figure: An invented company starts the year with 200 employees and ends it with 220, and 30 people leave during the year. Average headcount is 200 plus 220, divided by 2, which is 210. Turnover is 30 divided by 210, about 14.3%. Split by who chose, 20 resignations are 9.5% voluntary turnover and 10 dismissals are 4.8% involuntary turnover._ Now split the 30. If 20 resigned and 10 were dismissed, voluntary turnover is 20 ÷ 210, about 9.5%, and involuntary turnover is 10 ÷ 210, about 4.8%. Those two numbers lead to very different conversations. ## Turnover splits into types that tell different stories. - [Voluntary turnover](https://www.teampredict.ai/glossary/voluntary-turnover): the employee chose to leave. - [Involuntary turnover](https://www.teampredict.ai/glossary/involuntary-turnover): the company ended the employment. - [Regretted attrition](https://www.teampredict.ai/glossary/regretted-attrition): you would have paid to keep the person. - [Functional turnover](https://www.teampredict.ai/glossary/functional-turnover) and [dysfunctional turnover](https://www.teampredict.ai/glossary/dysfunctional-turnover): whether the exit left the organization better or worse off. _Figure: Every departure answers three separate questions. Who chose: voluntary if the employee did, involuntary if the company did. Did you want to keep them: regretted or non-regretted. What did it do to the organization: dysfunctional if it damaged it, functional if it left it better off._ ## Good and bad depend on who left, not only how many. Published turnover benchmarks vary enormously by industry, role, seniority and region, and rarely say which population they measured. Your own trend over time is the more reliable comparison. A good sign is a stable or falling voluntary rate with a small regretted share. A bad sign is a rising regretted share, even while the overall rate stays flat, because that is the version that hurts. ## These mistakes make turnover numbers misleading. - Comparing a monthly or quarterly rate with an annual one without annualizing it first, which makes the shorter period look far too low. - Dividing by the start-of-year or end-of-year headcount instead of the average, which skews the rate in a company that is growing or shrinking. - Reporting one company-wide figure. Turnover is often local to a team or a manager, so read it per team as well. - Changing what counts as a separation between periods, so the trend measures the rule change instead of the people. - Mixing layoffs into a retention report, which makes a restructuring look like a retention failure. ## Work out your own turnover rate. Enter your headcount at the start and end of the period and the number of separations, and the free calculator runs the formula and shows the annualized figure. [Calculate your turnover rate](https://www.teampredict.ai/tools/employee-turnover-rate-calculator) Free, no signup required. ## These related terms are in the glossary too. - [Turnover rate](https://www.teampredict.ai/glossary/turnover-rate) The turnover rate is the percentage of your workforce that left during a period. - [Attrition](https://www.teampredict.ai/glossary/attrition) Attrition is headcount lost when employees leave and the role is not backfilled, so the position disappears from the org chart. - [Employee churn](https://www.teampredict.ai/glossary/employee-churn) Employee churn is an informal synonym for employee turnover, borrowed from subscription metrics where churn means customers lost. - [Staff turnover](https://www.teampredict.ai/glossary/staff-turnover) Staff turnover is the same measure as employee turnover, and is the more common phrasing in the UK, Ireland, India, Australia and much of the Commonwealth. - [Separation](https://www.teampredict.ai/glossary/separation) A separation is any end of the employment relationship, voluntary or involuntary. It is the numerator in most turnover formulas. - [Cost of turnover](https://www.teampredict.ai/glossary/cost-of-turnover) The cost of turnover is the full cost of replacing one leaver: recruiting spend, interviewing time, onboarding, the new hire's ramp period, manager time and any temporary cover. ## These pages on this site go deeper. - [What is employee turnover](https://www.teampredict.ai/blog/what-is-employee-turnover) About TeamPredict TeamPredict is retention early-warning software that flags resignation risk from public LinkedIn signals, so you have lead time to keep the people you want to keep. [See how TeamPredict works](https://www.teampredict.ai/retention-software) [Back to the HR glossary](https://www.teampredict.ai/glossary#employee-turnover), where all 52 terms are defined.