Employee turnover is the rate at which people leave your organisation over a given period. It is the most quoted number in HR and one of the most inconsistently calculated, which is why two organisations reporting 18 per cent can be in completely different situations — and why benchmarking against a sector average is far less useful than most people assume.
This guide covers the definition, the formula, the variants that are actually worth calculating, and the mistakes that make the headline figure misleading.
Table of Contents
What employee turnover means
Employee turnover measures departures as a proportion of your workforce over a period, usually a year. It covers everyone who left, whatever the reason.
Two distinctions do most of the analytical work:
- Voluntary — the employee chose to leave. Resignations, and usually retirements.
- Involuntary — the organisation ended it. Dismissals, redundancies, end of fixed term.
Turnover is also distinct from attrition, though the words get used interchangeably. Attrition usually implies the role is not being backfilled — the headcount shrinks. Turnover implies replacement.
How to calculate employee turnover
The standard formula:
Turnover rate = (leavers during the period ÷ average headcount during the period) × 100
Average headcount is normally (headcount at start + headcount at end) ÷ 2.
Worked example. You started the year with 240 employees, ended with 260, and 47 people left.
- Average headcount = (240 + 260) ÷ 2 = 250
- 47 ÷ 250 = 0.188
- × 100 = 18.8 per cent employee turnover
If your headcount changed sharply mid-year, the two-point average distorts things. Use a monthly average — add the twelve month-end headcounts and divide by twelve — which is more work and considerably more accurate during growth or restructuring.
For a monthly rate, use the same formula over one month. Do not multiply a monthly rate by twelve to annualise it; seasonality makes that unreliable in most sectors.
The employee turnover variants worth calculating
The headline rate is the least informative version. Four cuts earn their keep.
Voluntary turnover
Resignations only, over average headcount. This is the number that reflects choices people made about you. A restructure inflating the headline figure tells you nothing about whether people want to work here.
Regretted turnover
Voluntary leavers you would have preferred to keep. It requires a judgement call per leaver, which is why most organisations skip it, and it is the single most decision-useful figure on this list. A 22 per cent rate that is mostly unregretted is a functioning organisation. A 9 per cent rate concentrated in people you were about to promote is an emergency.
Early-tenure turnover
Leavers with under twelve months’ service, as a proportion of starters. High early-tenure turnover points at recruitment, onboarding or a job advert that oversold the role — a different problem with a different fix, and one the annual figure buries completely.
Stability index
(Staff with one or more years’ service ÷ staff in post a year ago) × 100
This catches something turnover hides: churn concentrated in a small number of posts. One role filled and vacated four times generates the same turnover count as four separate people leaving four separate jobs, but the stability index tells them apart.
Employee attrition rate vs turnover rate
Turnover and attrition get used interchangeably, but they answer different questions. Turnover counts everyone who leaves, whether or not you replace them. Attrition usually means people who leave and are not replaced, so headcount shrinks: retirements, roles absorbed into a restructure, or posts left empty during a hiring freeze.
Attrition rate = (leavers not replaced during the period ÷ average headcount) × 100
Worked example: with an average headcount of 200 and 30 leavers in the year, turnover is 15 per cent. If 12 of those roles were never refilled, the attrition rate is 6 per cent.
Track both if headcount is changing. A falling attrition rate alongside steady turnover means you’re replacing people rather than shrinking, which is a recruitment cost question. Rising attrition with flat turnover usually reflects a deliberate decision to reduce headcount, not a retention problem.
- Use turnover to understand retention and the cost of replacing people.
- Use attrition to understand how headcount is changing over time, for example during a hiring freeze.
- Report them separately. Mixing them makes a restructure look like a retention crisis, or hides one.
What is a good employee turnover rate?
CIPD analysis puts average turnover for UK workers at around 34 per cent — roughly 27 per cent moving to a new employer and 7 per cent not working a year later. Sector variation around that is enormous: hospitality and retail routinely run several times higher than professional services or the public sector.
Which makes the honest answer unsatisfying. There is no universal good rate. What there is:
- Your own trend, which is the only genuinely comparable benchmark you have.
- Your sector, treated as a sanity check rather than a target.
- The distribution, which matters more than the average. Turnover concentrated in one team, one manager or one tenure band is a specific solvable problem. Evenly spread turnover is a cultural one.
Zero is not the goal either. Some turnover is healthy — it brings in new thinking and moves people out of roles they have outgrown. An organisation with 2 per cent turnover is often one where nobody can leave rather than one where nobody wants to.
Five mistakes in employee turnover maths
- Using end-of-year headcount instead of the average. In a growing organisation this understates turnover, sometimes substantially.
- Changing the population between years. Including contractors one year and not the next produces a trend that is entirely artefact.
- Pooling voluntary and involuntary. A redundancy programme makes your engagement look catastrophic in a metric that was never measuring it.
- Reporting small teams as percentages. One leaver from a team of six is 17 per cent turnover and one person. Report counts at that scale.
- Comparing to a benchmark with an undefined method. If you cannot see how a published figure was calculated, it is not a comparison.
What the number cannot tell you
Employee turnover is a lagging indicator, and a slow one. A resignation in March reflects a decision taken in January, following a frustration that built through the previous autumn. By the time it lands in your dashboard, the events that caused it are six months cold and the people who could have described them have gone.
It also cannot see the people who have disengaged and stayed — the quiet quitting population, who cost you real output while leaving your turnover rate untouched.
So calculate it properly, segment it hard, and treat it as the scoreboard rather than the game. The leading indicators — whether people feel recognised, whether they can see a future here, whether raising a problem achieves anything — move months earlier and are the only ones you can still act on.
Wotter does both halves of that. Our leaver analysis segments employee turnover by team, manager, tenure and demographic, and sits alongside continuous engagement data — so the scoreboard comes with the story of how it got that way.
For UK benchmarking, the CIPD’s employee turnover data hub and its turnover and retention factsheet are the most reliable free sources, and both are explicit about their methods.
Frequently asked questions
Common questions on calculating turnover and interpreting the result. Turnover is the lagging indicator. Wotter is an employee engagement platform — we track the leading ones, while they can still be acted on. Anything else, drop us a line at hello@wotter.group.
