Attrition Rate: Formula, Benchmarks and How to Cut It

Attrition rate is the percentage of employees who leave an organisation over a defined period, calculated as the number of exits divided by the average headcount for that period, multiplied by 100.

The formula is simple. Almost every disagreement about attrition numbers inside a company comes from the denominator, the period, or from mixing voluntary and involuntary exits, not from the arithmetic. This guide sets out the calculation precisely, then covers what the number actually tells you and what to do about it.

High attrition means constant rehiring. Goodspace fills replacement roles in 7 days on a flat success fee. See how it works

What is the attrition rate formula?

The standard formula is:

Attrition rate = (Number of employees who left during the period / Average number of employees during the period) × 100

Where average headcount is:

Average headcount = (Headcount at start of period + Headcount at end of period) / 2

A worked example. A company starts the year with 180 employees, ends with 220, and 45 people left during the year.

Average headcount = (180 + 220) / 2 = 200
Attrition rate = (45 / 200) × 100 = 22.5%

Note what happens if you use ending headcount instead of average: 45 / 220 = 20.5%. And starting headcount: 45 / 180 = 25%. Same reality, three different numbers. In a growing company, using ending headcount flatters the figure, which is why the average is the standard and why you should ask which denominator someone used before comparing.

What are the different types of attrition?

Reporting a single blended number hides the only information that is actionable.

Voluntary attrition. The employee chose to leave. This is the number that measures your organisation.

Involuntary attrition. Terminations, redundancies, non-confirmation after probation. This measures your hiring and performance management, not your retention.

Regrettable attrition. Voluntary exits of people you wanted to keep. This is the number that actually matters and almost nobody tracks it.

Non-regrettable attrition. Voluntary exits of people whose departure is acceptable or welcome.

Early attrition. Exits within 90 or 180 days of joining. Usually a hiring or onboarding failure rather than a retention failure, because something was misrepresented or the onboarding failed.

If you track one additional metric beyond overall attrition, make it regrettable attrition. A company at 25% overall attrition where almost all of it is non-regrettable is in a very different position from one at 12% where every leaver was someone they wanted to keep, and the blended number cannot distinguish them.

How do you calculate monthly and annualised attrition?

For monthly attrition, use monthly exits over average monthly headcount:

Monthly attrition = (Exits in month / Average headcount in month) × 100

To annualise a monthly figure, multiply by 12. Be careful with this: annualising a single month with unusual activity, such as the month after appraisals or bonus payout, produces a badly misleading figure. Use a rolling twelve month calculation instead:

Rolling 12-month attrition = (Exits in last 12 months / Average headcount over those 12 months) × 100

The rolling calculation is the one to report to leadership, because it smooths the seasonality that is real in Indian companies, where exits cluster after appraisal cycles and bonus payouts.

What is a good attrition rate?

There is no universal answer, and any benchmark quoted without a sector, a level and a geography attached is not useful. Attrition varies enormously by industry, by role type and by seniority.

What is more useful than a benchmark:

Compare against yourself over time. Your trend is more informative than someone else's number.

Segment before you judge. Attrition concentrated in one team, one manager, one level or one location is a specific problem with a specific cause. A blended company number tells you nothing about where to act.

Distinguish healthy from unhealthy. Some attrition is necessary. Zero attrition means no new perspectives and no exit for poor performers.

Watch early attrition separately. High attrition at 90 days is not a retention problem, it is a hiring accuracy problem, and it is fixed at the job description and interview stage rather than with engagement initiatives.

The segmentation point is the practical one. When a company reports 20% attrition, the useful next question is always: where? The answer is usually that three teams account for most of it.

What does attrition actually cost?

The visible cost is recruitment fees. The larger costs are invisible and rarely counted.

The components:

Replacement cost. Agency fees or internal recruiting cost, advertising, verification. See cost per hire.

Vacancy cost. Output not produced while the seat is empty. In India this is longer than people assume, because notice period plus hiring cycle commonly means three to five months from resignation to a productive replacement.

Ramp cost. The new person is paid full salary while producing partial output for weeks or months.

Handover and knowledge loss. What the leaver knew and did not write down.

Team productivity drag. Colleagues absorbing the work, plus the morale effect, which is real when exits cluster.

Management time. Exit process, hiring process, onboarding.

A simple way to estimate it for your own company: take the total cost to hire and onboard one person in a given role, add the salary paid during the ramp period where output is partial, and add the cost of the seat being empty. For most professional roles the total lands well above a single month's salary and often well above a quarter's. Run the number for one role in your company rather than using a generic multiplier, because it is the number that makes retention investment legible to finance.

How do you find out why people are leaving?

Exit interviews are the standard tool and the least reliable one, because people leaving rarely tell the whole truth to the employer they are leaving.

Better sources, in rough order of usefulness:

Stay interviews. Ask current employees what would make them leave, and what nearly made them leave in the past year. Far more actionable than an exit interview, because the person is still there and the answer is not filtered through a desire to leave cleanly.

Segmented attrition data. Which team, which manager, which tenure band, which level. Patterns are usually visible in the data before anyone articulates them in an interview.

Tenure curve analysis. When do people leave? Clustering at 12 months suggests a hiring or role-clarity problem. Clustering at 24 to 36 months often suggests a progression problem.

Exit interviews conducted by someone independent, and ideally a follow-up call three months after the person has left, when they have no reason to be diplomatic.

Internal transfer requests. People asking to move internally are telling you about a problem in their current team before they resign.

That last signal is the most underused. A cluster of transfer requests out of one team is the clearest available early warning, and it usually arrives months before the resignations.

How do you reduce attrition?

Act on the segment, not the average. The interventions that work are specific to the cause, and generic engagement programmes applied to a specific problem waste money and credibility.

Matching cause to action:

Pattern in the data	Likely cause	What to do
High attrition in one team	Manager	Address the manager, directly
Clustered at 90 days	Hiring misrepresentation or onboarding	Fix the job description and the first 30 days
Clustered at 18 to 30 months	No progression path	Career framework, internal mobility
Concentrated at one level	Compensation compression	Benchmark that level specifically
After appraisal cycles	Rating or increment dissatisfaction	Review calibration and communication
Across the board	Compensation below market, or culture	Benchmark, then investigate

Broader levers covered in the employee retention guide: compensation benchmarking, career pathing, manager capability, internal mobility, and honest hiring.

One point worth stating plainly: manager quality is the most common single driver of team-level attrition, and it is the one most companies address last. If your data shows one team consistently losing people while comparable teams do not, that is the finding, and no engagement survey is needed to confirm it.

Rehiring constantly while you fix retention? Goodspace fills replacement roles in 7 days so vacancies do not compound the problem.

Conclusion

Get the calculation right first, because most internal arguments about attrition are denominator arguments. Use average headcount, use a rolling twelve month window for reporting, and never compare your number to an external benchmark without checking which denominator the other party used.

Then stop looking at the blended number. Segment by team, manager, tenure and level, and start tracking regrettable attrition specifically, because a high overall rate made up of non-regrettable exits is a different situation entirely. And check the tenure curve, because when people leave tells you why more reliably than any exit interview will.

FAQs About Attrition Rate

What is the attrition rate formula? Attrition rate equals the number of employees who left during a period divided by the average headcount for that period, multiplied by 100. Average headcount is the start-of-period headcount plus end-of-period headcount, divided by two.

How do you calculate attrition rate with an example? A company starting the year with 180 employees, ending with 220, and losing 45 people has an average headcount of 200 and an attrition rate of 45 divided by 200, times 100, which is 22.5%.

Should I use starting, ending or average headcount? Average headcount is standard. In a growing company, using ending headcount flatters the number and using starting headcount inflates it. The same 45 exits can produce 20.5%, 22.5% or 25% depending on the denominator chosen.

What is regrettable attrition? Voluntary exits of people you wanted to keep. It is the most useful attrition metric and the least commonly tracked. A company at 25% attrition that is mostly non-regrettable is in a far better position than one at 12% where every leaver was someone they wanted.

What is a good attrition rate? There is no universal figure, and benchmarks quoted without sector, level and geography are not useful. Compare against your own trend over time, and segment by team, manager, tenure and level, because that is where the actionable information is.

How do I annualise a monthly attrition rate? Multiply the monthly figure by 12, but be cautious. Annualising a single unusual month, such as the one after appraisals, is misleading. A rolling twelve month calculation is the better figure to report because it smooths seasonality.

What does attrition actually cost? Replacement cost, vacancy cost while the seat is empty, ramp cost while the new person produces partial output, knowledge loss, team productivity drag, and management time. In India the vacancy cost is larger than expected because notice period plus hiring cycle is commonly three to five months.

Are exit interviews reliable? Not particularly, because people leaving rarely give the full picture to the employer they are leaving. Stay interviews with current employees, segmented attrition data, tenure curve analysis and internal transfer requests are all more actionable signals.

Additional Resources

Your own HRIS exit data segmented by team, manager, tenure band and level, which is where the actionable pattern lives
Your internal transfer request log, the earliest available warning of a team-level problem
Your 90 day attrition figures, which measure hiring accuracy rather than retention

Backfilling roles while you fix the cause?

Attrition creates vacancies faster than most teams can fill them, and open seats compound the pressure on everyone remaining. See how Goodspace fills roles in 7 days.