![\"Diagram](\"/article-assets/tasso-di-turnover/tasso-di-turnover-tre-formule.jpg\")
The question almost always comes up after the fact, when someone asks \"so what's that as a percentage?\" — and the percentage, on its own, doesn't answer it.
\nThe employee turnover rate is the ratio between the staff movements in a period and the headcount taken as a reference, multiplied by one hundred: in the most common version, leavers in the period divided by headcount, times one hundred [1].
\nIt applies to a professional practice with two team members, a family business with twelve employees and an organization of a hundred: the figure changes, not the way it gets misread.
\nIn the first quarter of 2026 alone, 1,527,000 employment relationships ended in Italy [3]: a phenomenon measured precisely at the national level and rarely with the same rigor inside a single company.
\n\n\nThe employee turnover rate doesn't tell you whether there is a problem: it tells you how many people have walked through the door. In a small team its unit of measure is not the percentage point, it's the person.
\n
This article walks through the decisions hidden in the formula, the three ways to calculate it, an example on a small headcount, the four checks that come before any conclusion and the mistakes that make the number useless.
\nThe employee turnover rate is a ratio: the numerator is a count of staff movements, the denominator is a reference headcount, and the result is multiplied by one hundred.
\nTreccani defines it as a \"staff rotation rate that indicates the flow of people into and out of the company's workforce\" [1].
\nThe part people take for granted is the denominator, and it's the part that decides the result.
\nThe standard formulas don't use just one: the overall rate is calculated on the average annual headcount, while the hiring and separation rates are calculated on the headcount at the start of the year [1].
\nTwo identical companies using different denominators get different numbers, and both call them \"turnover rate.\"
\nEven before the denominator, you have to decide what goes into the numerator.
\nThe end of a fixed-term contract, a retirement, a resignation and a dismissal are four different events, which the formula adds up as if they were the same.
\nAdding them up is legitimate; doing it without knowing how many there are of each type, much less so.
\nTwo closely related terms need to be kept apart, because in everyday business language they are swapped all the time.
\nThe replacement rate puts hires over leavers and tells you whether the headcount is rebuilding itself, not how much it rotates [1].
\nAnd turnover, on its own, still means annual revenue in many business conversations [1].
\nAn employee turnover rate is a number that means something only if three elements are stated together: the time window, the types of movement counted and the headcount used as the denominator.
\nWithout those three elements stated, it isn't an indicator, it's a percentage.
\nThe three figures you're looking for have three formulas, and between them not only the numerator changes but the denominator too.
\nIn the standard formulation, outgoing turnover is the number of leavers in the year divided by the headcount at the start of the year, times one hundred; incoming turnover is the number of hires in the year divided by the same start-of-year headcount, times one hundred; overall turnover is the sum of hires and leavers divided by the average annual headcount, times one hundred [1].
\nTwo different denominators inside three formulas that are presented together.
\nThe practical consequence is that the overall rate is not the sum of the other two: anyone who checks it with a calculator finds a gap and doesn't know what to attribute it to.
\nThe fix is made once and holds for the following years.
\nYou choose a single reference headcount and calculate all three rates on it.
\nThe most stable choice is the average headcount for the period, that is the average of the headcount at the start of the period and at the end of the period, because it doesn't shift depending on whether hires happened in January or in November.
\nWith a single denominator the numbers add up: the incoming rate plus the outgoing rate gives the overall rate, and the gap disappears.
\nThe cost is stated up front: a rate calculated this way is not comparable with an industry table built on the standard formula.
\nThe gain is that the company's three numbers become comparable with one another and over the years, which is what they're really for.
\nThere's a fourth indicator, often more telling than the first three: the replacement rate, that is the number of hires divided by the number of leavers, times one hundred [1].
\nAbove one hundred the headcount grows, below one hundred it thins out, around one hundred it replaces itself.
\nReplacing is not standing still: it means that every year a share of the work changes hands, with the onboarding time that comes with it.
\n
\n\nThe employee turnover rate doesn't tell you whether there is a problem: it tells you how many people have walked through the door. In a small team its unit of measure is not the percentage point, it's the person.
\n
Take a small-town machine shop — an explicitly hypothetical case — that starts the year with nine people and ends it with eleven.
\nThe average headcount for the period is ten.
\nDuring the year two people leave and four join.
\nWith the average headcount as the single denominator: outgoing turnover 20%, incoming turnover 40%, overall turnover 60%.
\nThe 60% doesn't describe a hemorrhage: it describes a company that grew by two people.
\nThe overall rate adds hires and leavers together, so it rises even when the company is hiring — the first example of what the formula doesn't tell you.
\nThe second shows up when you move a single person.
\nIf one person had left instead of two, outgoing turnover would have been 10% instead of 20%.
\nOne person is worth ten percentage points.
\nWith a headcount of ten the smallest step of the ratio is one tenth, that is ten points; with a headcount of twenty it's five points; with a headcount of a hundred it's one point.
\nIt's not a flaw in the calculation: it's the denominator that sets the resolution at which the percentage can move.
\nThe point applies to most of Italy's business landscape, where the average company size is 3.9 employees against 5.0 for the EU average, within the harmonized European scope of industrial and service activities that makes the two values comparable (2022 data) [2].
\nWith denominators of that order of magnitude, a gap of twenty or thirty points from one year to the next can correspond to a single person who changed jobs.
\nThe business owner who splits their time between the office and the shop floor already knows this without calculating anything: they know who left, and they know why the line ran slower the following Monday.
\nThe rate adds no information about the individual case: it's there to compare this year with the previous three, calculated the same way — something memory does badly and a spreadsheet does well.
\nThere is no value above which the employee turnover rate becomes a problem.
\nThe very source that provides the formulas says so: an \"optimal\" value doesn't exist, because rates vary with the company's industry, labor market conditions, company size and the contract types considered [1].
\nWhat does exist is a threshold within the calculation itself, and it comes from the denominator: as long as the gap between two periods stays below the smallest step of the ratio — one person — it isn't a signal, it's noise.
\nAbove that threshold the number deserves four checks, in this order.
\nOnce the four checks are done, the rate tells you where to look.
\nIt doesn't tell you why: the reasons people leave are a separate investigation, and they have a dedicated page on the causes of employee turnover, which also covers the costs of a departure and the levers to reduce it.
\nHere we stay with measurement: no section of this article deals with causes, and no conclusion about why should be drawn from the number.
\nThe investigation starts from what was recorded while it was happening, and the place where those notes accumulate is the problem log: a department that starts moving before the others has almost always left entries open for months.
\nThe distinction between departures that take value away and departures that free it up is set out in the glossary entries dysfunctional turnover and functional turnover.
\n![\"Diagram](\"/article-assets/tasso-di-turnover/tasso-di-turnover-quattro-controlli.jpg\")
An employee turnover rate gets spoiled more often when it's compared than when it's calculated.
\nFive recurring mistakes, all fixable without rewriting the data history.
\nThe last mistake is the most expensive, because it leads you to decide on a comparison that doesn't exist.
\nA universal benchmark doesn't exist: rates vary with the company's industry, labor market conditions, company size and the contract types considered [1].
\nThe figures used in this article — ten points per person on a headcount of ten — are the arithmetic of the ratio, not parameters to adopt.
\nThe ISTAT and INPS data cited measure employment relationships and job positions across Italy as a whole: they describe the context in which a company operates, not a point of comparison for the individual company.
\nThe machine shop example is explicitly hypothetical and serves to show how the denominator works, not a real case.
\nWhere seasonal or on-call contracts are structural, the rate is higher by construction and comparing it with companies that don't use them loses meaning.
\nFinally: the calculation says nothing about the causes of departures, and the formulas described here don't claim to.
\nDivide the number of people who left in the period by the reference headcount and multiply by one hundred.
\nIn the standard formulation the denominator is the headcount at the start of the year [1].
\nIf you also calculate incoming and overall turnover, you get numbers that are consistent with each other only by using the same denominator for all three.
\nIncoming turnover counts only hires; overall turnover adds hires and leavers together [1].
\nOverall turnover therefore rises even when the company grows, and a high value doesn't in itself mean the workforce is emptying out.
\nNo: an \"optimal\" value doesn't exist, because rates depend on industry, the labor market, company size and contract types [1].
\nThe useful comparison is with your own data from previous years, calculated with the same method.
\nThe average of the headcount at the start of the period and the headcount at the end of the period.
\nThe scope of people counted in the headcount must match the scope of the movements put in the numerator.
\nWith the same formula, replacing the annual window with the shorter one and using the average headcount for the same interval.
\nTurnover calculated over three months can't be compared with an annual figure without rescaling it, and it is much more affected by seasonality than the twelve-month figure.
\nCalculating the employee turnover rate means making three decisions before picking up the calculator: which time window, which movements to count, which headcount to put in the denominator.
\nThe three versions — incoming, outgoing, overall — add up only if they rest on the same reference headcount.
\nIn small teams the percentage moves one person at a time, and that is the real threshold below which movement isn't a signal.
\nWhen movement exceeds that threshold, look at four things before concluding anything: the department, the tenure of those who left, the season and the role of the person.
\nA reasonable cadence is once a year for the company, with a department-level reading for the same period.
\nThe mistakes that make the number useless almost all concern the comparison, not the calculation.
\nThe employee turnover rate is not a judgment: it's the count of how many people walked through the door in a period, and in a small team it moves one person at a time, not one point at a time.
\nEverything else follows from there.
\nA denominator chosen once and stated next to the formula, the same time window every year, and next to the percentage the answer to four questions: which department, with how much tenure, in which season, and who.
\nThe reasons people leave are a different investigation, covered in the causes of employee turnover, while this indicator's place among the other people metrics is in people management.
\nThe other indicator that calls for the same care in stating the scope before the result is the absenteeism rate.
\nThe value of a well-calculated rate shows up in the third year.
\nAt that point the December discussion about how much worse it went than the year before loses its subject: there is a series of numbers built the same way, and the series speaks for itself.
\nAnd when a department starts moving before the others, you notice while it's happening: in time to ask questions, rather than after the fact.
\n\n[1] Istituto della Enciclopedia Italiana — Treccani, \"Turnover\", Dizionario di Economia e Finanza, 2012. Available at: https://www.treccani.it/enciclopedia/turnover_(Dizionario-di-Economia-e-Finanza)/
\n[2] ISTAT, \"Industria e servizi — Strutture produttive\", Noi Italia 2025 (2022 data). Available at: https://noi-italia.istat.it/pagina.php?id=3&categoria=10&action=show&L=0
\n[3] INPS, \"Osservatorio sul mercato del lavoro: i dati di marzo 2026\", Coordinamento generale statistico attuariale, published June 25, 2026. Available at: https://www.inps.it/it/it/inps-comunica/notizie/dettaglio-news-page.news.2026.06.osservatorio-sul-mercato-del-lavoro-i-dati-di-marzo-2026.html
\n[4] ISTAT, \"Il mercato del lavoro — II trimestre 2026\", press release, 2026. Available at: https://www.istat.it/comunicato-stampa/il-mercato-del-lavoro-ii-trimestre-2026/
","headings":[{"level":2,"text":"The two decisions that define the employee turnover rate calculation","id":"the-two-decisions-that-define-the-employee-turnover-rate-calculation"},{"level":2,"text":"Deriving incoming, outgoing and overall turnover from a single headcount","id":"deriving-incoming-outgoing-and-overall-turnover-from-a-single-headcount"},{"level":2,"text":"In a small team the turnover rate moves ten points per person","id":"in-a-small-team-the-turnover-rate-moves-ten-points-per-person"},{"level":2,"text":"The four checks to run before reading an employee turnover rate","id":"the-four-checks-to-run-before-reading-an-employee-turnover-rate"},{"level":2,"text":"The most common mistakes in calculating the employee turnover rate","id":"the-most-common-mistakes-in-calculating-the-employee-turnover-rate"},{"level":2,"text":"Limits and conditions of applicability","id":"limits-and-conditions-of-applicability"},{"level":2,"text":"FAQ","id":"faq"},{"level":3,"text":"How do you calculate the outgoing turnover rate?","id":"how-do-you-calculate-the-outgoing-turnover-rate"},{"level":3,"text":"What's the difference between incoming turnover and overall turnover?","id":"whats-the-difference-between-incoming-turnover-and-overall-turnover"},{"level":3,"text":"Is there a normal employee turnover rate?","id":"is-there-a-normal-employee-turnover-rate"},{"level":3,"text":"What goes into the average headcount?","id":"what-goes-into-the-average-headcount"},{"level":3,"text":"How do you calculate turnover on a monthly or quarterly basis?","id":"how-do-you-calculate-turnover-on-a-monthly-or-quarterly-basis"},{"level":2,"text":"Key takeaways","id":"key-takeaways"},{"level":2,"text":"Conclusion","id":"conclusion"},{"level":2,"text":"Sources and references","id":"sources-and-references"}],"tldr":"Two departures in a year out of ten people: is that a problem or just normal attrition?","tldrItems":null}