Attrition and retention are the same story told in opposite directions. Attrition counts the people who left; retention counts the people who stayed. Track only one and you get half the picture usually the half that makes leadership feel better. This guide explains what is attrition, what is retention rate, how to calculate both correctly, and what the benchmark ranges actually look like across industries in 2026. Fabric works with recruiting teams whose retention numbers are shaped by hiring quality upstream, so the framing here is deliberately practical: what to measure, what to compare it against, and what to do when the number is wrong.
What is attrition?
Attrition is the gradual reduction of a workforce when employees leave and are not immediately replaced. It is measured as a percentage of the workforce over a defined period, most often a rolling twelve months. The definition matters: attrition is specifically about net headcount loss over a window, which is why it is often used interchangeably and incorrectly, with turnover. SHRM defines attrition as "gradual workforce reduction through employee retirements, resignations, deaths or elimination of positions without immediately filling them," and that "without immediately filling" clause is what separates it from turnover, which counts every exit regardless of backfill. Attrition splits cleanly into two categories. Voluntary attrition is people leaving on their own: resignations, retirements, career pivots. Involuntary attrition is the company initiating the exit: terminations, layoffs, non-renewals, role eliminations. Reporting a single blended number without the split hides which lever is broken.
What is retention rate?
Retention rate is the percentage of employees who stayed with the company across a defined period. Where attrition asks "how many left?", retention asks "how many are still here?" same window, same population, opposite framing. If you started the year with 100 employees and 88 of the original 100 are still on payroll twelve months later, your annual retention rate is 88%. Retention is a stayer-based metric: it looks only at people who were already there at the start of the period, ignores anyone hired mid-period, and asks what share of that original cohort remains. That's what makes it useful. It isolates whether the people you already had chose to stay, independent of how much hiring you did on top. A rising headcount can mask a retention problem for months if you only watch total employee count. The metric that answers what is retention rate honestly is the one calculated on the original cohort, not on the rolling total.
Attrition vs. retention rate: the actual difference
The two numbers are related but not perfect mirrors, a common mistake is assuming retention rate equals 100% minus attrition rate. That equation only holds when you're counting the same population the same way, which is rarely how organisations actually report. Attrition typically uses average headcount across the period as its denominator and counts all exits (including people who were hired and left within the same window). Retention typically uses starting headcount and counts only the original cohort who remained. The denominators differ, the numerators differ, and so the two figures don't sum to 100. Below is the practical difference:
| Aspect | Attrition rate | Retention rate |
|---|---|---|
| What it counts | Employees who left | Employees who stayed |
| Denominator | Average headcount across the period | Starting headcount at period open |
| Population | Everyone employed during the period | The original cohort only |
| What it hides | Fast churn on new hires (partially) | Any exits among people hired mid-period |
| Best used for | Diagnosing exit volume and cost | Diagnosing whether tenured employees stay |
The honest way to read them together: if attrition is high and retention is also high, you have a churn problem concentrated in new hires. If attrition is high and retention is low, the problem is broader and tenured people are leaving too. If both are healthy, the number is real not an artefact of measurement.
How to calculate attrition rate
The standard formula for attrition rate is:
Attrition rate = (Number of employees who left during the period ÷ Average number of employees during the period) × 100
Average headcount is usually calculated as (starting headcount + ending headcount) ÷ 2, or as the average of monthly headcounts for a more accurate figure on volatile teams. A worked example: a customer support team starts the year with 120 agents and ends with 96, after 42 exits and 18 new hires. Average headcount is (120 + 96) ÷ 2 = 108. Attrition rate = (42 ÷ 108) × 100 = 38.9%. Common mistakes when calculating: using starting headcount instead of average headcount inflates the rate on growing teams and understates it on shrinking teams; counting internal transfers as exits (they aren't the person didn't leave the company); mixing voluntary and involuntary numbers into a single blended figure when leadership wants to know which lever to pull.
How to calculate retention rate
The standard formula for retention rate is:
Retention rate = (Number of original employees still employed at period end ÷ Number of employees at period start) × 100
Note the strict cohort framing: the numerator counts only people who were on the payroll at the start of the period and are still on it at the end. Anyone hired mid-period does not enter either the numerator or the denominator. A worked example: an engineering team starts the year with 45 engineers. Twelve months later, 39 of those original 45 are still on the team. Retention rate = (39 ÷ 45) × 100 = 86.7%. New hires who joined mid-year and are still there don't count; new hires who joined and left don't count either. This is deliberate — the metric answers a specific question ("did the people we already had stay?") and dilutes if you let the definition drift.
Benchmark table: attrition and retention by industry
Attrition rates vary by an order of magnitude across industries, so a "good" number depends entirely on what you're being compared against. The U.S. Bureau of Labor Statistics publishes total separations rates in the JOLTS series that broadly track industry-level attrition; the ranges below combine those figures with SHRM's guidance on typical hiring benchmarks and are intended as a directional check, not a substitute for pulling your own industry cut.
| Industry / role type | Typical annual attrition | Typical annual retention |
|---|---|---|
| Leisure and hospitality | 60-80% | 20-40% |
| Retail (front-line) | 55-75% | 25-45% |
| BPO / contact centres (voice) | 30-60% | 40-70% |
| Professional and business services | 15-25% | 75-85% |
| Financial services | 12-18% | 82-88% |
| Technology (product engineering) | 10-20% | 80-90% |
| Healthcare (clinical staff) | 20-30% | 70-80% |
| Government and public sector | 5-12% | 88-95% |
Two caveats. First, "typical" spans a very wide band inside each industry: an early-stage tech startup and a mature enterprise both fall under "technology" but rarely share attrition profiles. Second, benchmarks include voluntary and involuntary exits together. A company doing a large layoff will show elevated attrition that isn't a retention problem in the usual sense.
What drives high attrition
Exit interview data across the sources cited in this piece converges on a fairly consistent short list of drivers. SHRM's coverage of attrition causes and the standing findings in LinkedIn's Talent Solutions research both point to the same clusters: compensation misalignment relative to market, limited career progression, poor manager relationships, workload and burnout, and increasingly mismatch between the role that was sold in the interview and the role the person actually landed in. That last one is where hiring quality feeds directly into attrition. When candidates are screened loosely on the front end, or when interviews test for signals that don't match day-to-day work, early attrition on new hires runs high. Fabric's Interview Engine screens, scores, and shortlists candidates against role-specific technical and behavioural criteria before recruiters spend panel time on them. The recruiter or hiring manager still makes the hire decision, but they make it on a shortlist that is closer to the actual job. It's a signal for your team to weigh, not an automatic reject.
How to bring attrition down (without gaming the number)
Three moves account for most of the durable reduction, in roughly this order:
- Fix the front door. Early attrition (people leaving within their first 12 months) is disproportionately caused by mis-hires. Tighten the screening bar, add a realistic job preview during the interview, and measure 6-month and 12-month retention on new hires specifically, not just blended attrition.
- Fix compensation banding. Below-market pay is the driver mentioned most often in exit interviews. A once-a-year comp review is not enough in a fast-moving market; quarterly banding checks on the highest-attrition roles catch the problem before people are already interviewing elsewhere.
- Fix the manager layer. People leave managers, not companies. The cliché is repeated because the data keeps confirming it. Investing in first-line manager training, and being willing to move poor people-managers out of people-management roles, moves attrition meaningfully.
What doesn't move attrition durably: ping-pong tables, one-off retention bonuses paid to whoever threatens to leave, and rebranding "turnover" as "attrition" in the annual report. Real reductions come from fixing the upstream causes.
FAQ
What is attrition in the workplace?
Attrition in the workplace is the gradual reduction of a workforce when employees leave through resignation, retirement, death, or role elimination and are not immediately replaced. It is measured as a percentage over a defined period, most often annually.
What is 80% attrition?
An 80% attrition rate means that, over the measurement period, 80 out of every 100 employees left the company and were not replaced by the end of the period. Rates that high are typically seen in high-volume, low-tenure roles such as BPO customer support, seasonal retail, and gig operations.
What is an example of attrition?
A software team of 50 engineers ends the year with 42 people because 8 resigned, retired, or moved internally without backfill. The team's annual attrition rate is 8 divided by the average headcount of 46, or about 17%.
What is attrition in HR?
In HR, attrition is the metric that tracks how many employees leave and are not replaced within a defined window. HR teams split it into voluntary attrition (resignation, retirement) and involuntary attrition (termination, role elimination) so the causes can be diagnosed separately.
What is attrition in BPO?
Attrition in BPO refers to the rate at which agents leave voice, chat, and back-office contact centre roles. It is one of the highest-attrition segments in the labour market sustained annual rates of 30-45% are common, and 60%+ is not unusual on entry-level voice programmes.
What is considered a high attrition rate?
There is no universal cutoff. High is defined relative to your industry and role type. As a rough guide: under 10% annual attrition is low across most industries, 10-20% is typical for professional services and tech, and anything sustained above 25% signals a problem outside high-turnover sectors like BPO, retail, and hospitality.