Attrition vs Turnover: Definitions, Formulas, and What Small Businesses Should Track

Turnover and attrition both describe employees leaving a company, but they measure different things. Turnover tracks all departures, including those where the company hires a replacement. Attrition tracks departures where the position is not refilled, meaning the company gets smaller. The distinction matters because each points to a different problem and a different response.

This guide covers what each term means, how to calculate both, why the math works differently at small companies, and which metric matters more when you are running one.

What Is Employee Turnover?

Turnover has two subtypes that require different responses. Voluntary turnover means the employee chose to leave: they resigned, retired, or moved to another company. Involuntary turnover means the company initiated the separation: the employee was terminated for performance, laid off due to restructuring, or let go during a reduction in force.

Within voluntary turnover, there is a further distinction: regrettable vs non-regrettable. A high performer who leaves for a competitor is regrettable turnover. An underperformer who resigns before a performance improvement plan is non-regrettable. Tracking this distinction prevents you from treating all departures as equally problematic.

What Is Employee Attrition?

Common attrition scenarios: an employee retires and the team absorbs their responsibilities, a role is automated and the position is eliminated, a department is restructured and two roles are consolidated into one, or the company is downsizing and reduces headcount through natural departures rather than layoffs. Attrition is often a deliberate strategy in larger organizations that want to reduce costs without the disruption and legal risk of layoffs.

Attrition vs Turnover: The Key Difference

The one-line summary: turnover is about the door revolving (people come and go, headcount stays constant). Attrition is about the door closing (people leave, headcount drops). Both start with an employee departure. The difference is what happens next: replacement or elimination.

Attrition vs Retention: Where Each Metric Sits

Retention, not attrition, is the opposite of employee turnover. Turnover counts who left, retention counts who stayed, and one year produces both figures: a 20-person team that loses 4 people reports 20% turnover and roughly 80% retention. Attrition sits inside turnover, as the slice of exits where the role went away too.

The word roughly is doing real work in that sentence. Retention is normally calculated against the people on the books when the period opened, with anyone hired since then excluded, while turnover divides by average headcount. So the two rarely sum to exactly 100. The retention rate guide works through the formula.

How to Calculate Turnover Rate and Attrition Rate

The formulas are structurally identical. The difference is the numerator: turnover counts all departures (replaced and unreplaced), attrition counts only departures where the position was not refilled. Both use average headcount as the denominator, calculated as (headcount at start of period + headcount at end of period) / 2.

Which Departures Count, and Which Do Not

Both formulas fall apart if the numerator is defined loosely, and most small businesses define it loosely. A separation is an employee coming off your payroll for any reason: a resignation, a retirement, a discharge for performance, a layoff, the end of a fixed-term engagement, job abandonment, or a death in service. The Bureau of Labor Statistics sorts these into three buckets in its JOLTS series (quits, layoffs and discharges, and other separations), and borrowing that structure keeps your categories stable from one year to the next instead of drifting every time someone new builds the spreadsheet.

What does not belong in the numerator is the part people get wrong. An employee promoted into a different role or moved between locations has not separated: they are still on your payroll, so a company-wide rate should ignore the move entirely, even though a department-level rate would count it as an exit from that department. Employees out on FMLA, parental, medical or military leave are still employed, so they stay out of separations and stay in headcount. Independent contractors and 1099 workers never enter either side of the equation because they were never employees, and the same holds for staffing-agency temps who sit on the agency's payroll rather than yours.

A few cases need a written rule rather than a fresh judgment call each time. A temp who converts to a permanent role is a hire, not a separation. A boomerang employee who leaves in March and returns in September is one separation and one hire; if they leave again in December, that is a second separation. Seasonal staff will swamp the rate if you fold them into the same figure as your year-round team, so either exclude them and say so in the footnote, or report a seasonal rate separately. Furloughs are the genuinely ambiguous case: if you expect people back and they remain on the books, most employers leave them out, but the only wrong answer is changing your treatment between periods.

Getting Average Headcount Right

The denominator is where a growing company quietly flatters its own numbers. The shortcut most people use is a two-point average: headcount on the first day of the period plus headcount on the last day, divided by two. That is accurate when team size is roughly flat and misleading the moment it is not, because it assumes growth arrived evenly across the period when it usually arrived in one or two bursts.

The defensible method is to record headcount at the end of every month and average the twelve figures. Take a company that opens January with 12 employees, holds at 12 through September, then hires hard in the fourth quarter to finish December at 30. The two-point average is (12 + 30) / 2 = 21. The twelve-month average is (12 x 9 + 18 + 24 + 30) / 12 = 15. With four separations during the year, the two-point method reports 4 / 21 = 19.0% and the monthly method reports 4 / 15 = 26.7%. The second number is the honest one, because for most of the year those four people were leaving a team of twelve, not a team of twenty-one.

Annualizing carries the same trap. Multiplying a quarterly rate by four, as in the table above, assumes departures spread evenly across the calendar, which is exactly what they do not do in a seasonal business or in any company that runs performance reviews and bonus payouts at a fixed time of year. If you have twelve months of data, calculate over twelve months directly rather than scaling up a quarter. And do not panic if the result exceeds 100%: in high-churn industries it routinely does, and it simply means the average position turned over more than once during the year.

A Worked Example: One Full Year at a 22-Person Company

Take a professional services firm that opens the year with 22 employees and closes with 21, with month-end headcounts averaging 22. Six people separated: two resigned for jobs elsewhere, one retired, one relocated with a spouse, one was terminated for performance, and one role was eliminated when the firm dropped a service line. Five of the six positions were refilled and the eliminated role was not. The firm made five hires during the year, one of whom quit in week seven.

The headline is 27.3%, and it is the figure a lender or an acquirer will ask for. It is also the least useful number in the table. The retirement and the spousal relocation were never preventable, the performance termination was a management decision working as designed, and the eliminated role was a strategy change. Strip those out and the firm has a regrettable turnover rate of 9.1%: two people it wanted to keep and lost. That is the number that should drive a conversation about pay bands or career paths.

The 20% at the bottom deserves its own alarm. A departure in week seven means the firm paid to source, interview and onboard someone who produced almost nothing, and at five hires a year it takes only one more such exit to turn a bad hire into a hiring pattern. Note too how small attrition is next to turnover here, 4.5% against 27.3%. That ratio is typical at small business scale, and it is the arithmetic reason this guide keeps steering you toward turnover.

Here is that comparison as a blank sheet. It is deliberately one page: the same period, the same average headcount, and one line splitting your separations into positions you refilled and positions you did not. Everything else on it exists so that the two percentages at the bottom can be read next to each other without an argument about what went into either.

The refilled line is the whole distinction. Record it at the moment you decide, not at year end, because "we never got round to posting the role" and "we eliminated the role" look identical in a spreadsheet twelve months later and mean opposite things.

Why the Math Works Differently at Small Business Scale

Most turnover and attrition benchmarks are built for companies with hundreds of employees, where the numbers are statistically meaningful. At a small company, the math behaves differently, and understanding these quirks prevents overreaction or underreaction to the numbers.

The most actionable metric for a small business is 90-day retention rate: what percentage of new hires are still employed after 90 days. According to SHRM, up to 20% of all employee turnover happens inside the first 45 days. If your 90-day retention rate is below 80%, your onboarding process is the first place to investigate.

Segmenting the Rate So It Points at a Cause

A single company-wide percentage tells you that people left. It does not tell you who, when in their tenure, or from which team, and those three cuts are where the cause usually sits. The cleanest one at small scale is tenure, and the cleanest way to run it is by cohort rather than by rate. Take everyone hired in a given quarter and check how many are still employed at 90 days, at six months, and at one year. The denominator is a fixed group of named people, so the number cannot swing just because you hired three more staff in the meantime.

If you prefer a rate by tenure band, the denominator has to move with the band. First-year turnover is the number of first-year employees who left divided by the average number of employees who were in their first year, not divided by total headcount. Using total headcount as the denominator for every band is the most common spreadsheet error in small-company turnover reporting, and it makes early-tenure churn look a fraction of its real size.

By team or manager, skip the percentage entirely. A six-person team that loses three people produces a 50% rate that reads as noise to anyone who knows the denominator, when in fact three exits under one manager is the loudest signal in your data. Report counts and let them speak. Then attach a reason code to every exit at the moment it happens, drawn from a short closed list you do not revise: compensation, career growth, manager or team, workload or schedule, relocation or personal, performance, role eliminated. Free-text exit notes do not aggregate. Seven coded exits do, and after a year they tell you which of the retention strategies below is worth your budget.

Turnover Benchmarks by Industry

These benchmarks come from the JOLTS program at the Bureau of Labor Statistics, which publishes annual separations and quits by industry. Across the whole economy in 2025, total separations averaged 3.3% of employment a month and quits averaged 2.0%, which annualize to roughly 40% and 24%.

Two cautions apply before you hold your own number up against any row. JOLTS counts every separation, voluntary or not, so it sits well above the voluntary-only figures most HR content quotes as industry averages. Its rates are also weighted by employment, which means large employers move them far more than a 20-person firm ever could.

Industry rates are not fate, either. According to Gallup, 52% of voluntarily exiting employees say their manager or their organization could have done something to keep them, and 51% say that in the three months before they left, nobody talked to them about their job satisfaction or their future.

When Attrition Is a Deliberate Strategy

Letting headcount fall through attrition instead of cutting it through layoffs is a genuine option, and it is worth doing the arithmetic before you rely on it. Start with voluntary turnover. A 40-person company running 15% voluntary turnover loses about six people a year, roughly one every two months. If the plan is to shed four positions, a full hiring freeze gets there in about eight months, assuming the rate holds. It rarely holds precisely, which is why an attrition-based headcount plan needs a fallback date and a decision made in advance about what happens if the departures do not arrive on schedule.

The bigger problem is that attrition is untargeted. You do not choose who leaves; the labor market does, and the people with the most options leave first. A freeze long enough to remove four positions tends to remove them from wherever your most marketable employees happen to sit, not from the functions you meant to shrink. It also concentrates knowledge risk: in a company of 40, one exit can take the only person who knows how the payroll cycle or the renewal process actually runs.

There is a compliance dimension too. Choosing not to backfill a role is a business decision, but the pattern those decisions form is reviewable. If the positions you quietly eliminate skew toward older workers or toward any protected class, the analysis runs the same way it would for layoff selections.

A remark like "we can just wait for him to retire" is precisely the contemporaneous evidence that turns a budget decision into an age claim. Write down the business rationale for each position you decide not to fill, and write it at the time you decide.

The federal WARN Act reaches only employers with 100 or more employees, and only for a plant closing or a mass layoff, so most small businesses sit outside it. Several states run mini-WARN statutes at lower thresholds, New York from 50 employees and Iowa from 25. Check yours before eliminating several positions at once.

One last piece of housekeeping: when a position disappears, its work does not. Pull the job description for the eliminated role, list the recurring duties, and assign each one to a named person with an estimate of the hours it adds to their week. "The team will absorb it" is how planned attrition becomes unplanned attrition three months later.

How to Reduce Both Turnover and Attrition

Since attrition at small companies is usually a subset of turnover (departures you chose not to replace), the strategies for reducing both are largely the same. Focus on the drivers that cause people to leave in the first place.

The highest-ROI intervention for most small businesses is structured onboarding. According to SHRM, newly hired employees who completed a structured onboarding process are 58% more likely to still be with the company three years later.

A process is what makes a number like that reachable, and a good intention is not. New hire retention work needs named owners and dates on a calendar, which is what FirstHR automates: AI-generated plans, e-signature, task assignments, training delivery, and check-in scheduling.

Frequently Asked Questions