The HR and people analytics glossary

Employee turnover is the movement of people out of a company over a set period, counted as the employees who leave and are replaced.

Why it matters: It is the headline number executives ask for, and it sizes every downstream cost: recruiting, onboarding, lost output and lost institutional knowledge.

How it is measured: Separations during the period divided by average headcount for the same period, expressed as a percentage and usually annualized.

Attrition is headcount lost when employees leave and the role is not backfilled, so the position disappears from the org chart.

Why it matters: The distinction matters in planning: turnover creates a role to fill, attrition shrinks the team. Many HR teams use the two words interchangeably, so agree internally on which one you mean before comparing numbers with anyone.

How it is measured: Leavers during the period divided by average headcount for the same period.

The turnover rate is the percentage of your workforce that left during a period.

Why it matters: It is the comparable version of the raw leaver count: five departures means something very different in a team of twenty than in a company of five hundred.

How it is measured: Separations divided by average headcount, multiplied by 100. Annualize a monthly or quarterly figure before comparing it to an annual one, or the number reads far too low.

Voluntary turnover is a departure the employee chose: a resignation, a retirement, or a move to another employer.

Why it matters: This is the part of turnover a retention program can actually move. Reporting it separately from involuntary exits stops a restructuring from looking like a retention failure, and the reverse.

Involuntary turnover is a departure the company initiated: a termination, a performance exit, or a layoff.

Why it matters: Mixed into one total it hides the real story. A quarter of involuntary exits after a restructure is a plan working, the same number of voluntary exits is a problem.

Regretted attrition is a departure the company did not want: someone whose performance, skills or relationships you would have paid to keep.

Why it matters: It is the number leaders should act on. A flat overall turnover rate can hide a rising regretted share, which is the version that hurts.

How it is measured: The manager tags each exit as regretted or not, ideally at the exit interview, then you track regretted exits as a share of all exits.

Non-regretted attrition is a departure the company is neutral about or relieved by, usually a poor fit or a sustained performance problem.

Why it matters: Counting it in the same bucket as regretted attrition makes a healthy team look unstable and pushes managers to keep people they should let go.

Functional turnover is a departure that leaves the organization better off, typically a low performer or a poor culture fit choosing to move on.

Why it matters: It reframes the turnover conversation from volume to composition. The question is not only how many people left, it is which people.

Dysfunctional turnover is a departure that damages the organization: a high performer, a specialist, or someone hard to replace at any price.

Why it matters: This is the class of exit worth building an early-warning process around, because the replacement cost and the ramp time are both at their highest.

New-hire turnover is the share of a hiring cohort that leaves within a set early window, usually the first ninety days or the first year.

Why it matters: It points at hiring and onboarding rather than long-term management, so it separates a selection problem from a retention problem.

How it is measured: Hires from a cohort who left inside the window, divided by the size of that cohort.

Staff turnover is the same measure as employee turnover, and is the more common phrasing in the UK, Ireland, India, Australia and much of the Commonwealth.

Why it matters: Worth knowing when you read benchmarks or vendor material: a UK report on staff turnover and a US report on employee turnover are measuring the same thing.

Employee churn is an informal synonym for employee turnover, borrowed from subscription metrics where churn means customers lost.

Why it matters: The borrowed word carries a borrowed assumption. Customer churn is largely a pricing and product story, people churn is mostly a manager, growth and workload story, so the playbooks do not transfer.

A separation is any end of the employment relationship, voluntary or involuntary. It is the numerator in most turnover formulas.

Why it matters: Getting the definition wrong quietly breaks every rate you publish. Decide up front whether interns, contractors, fixed-term staff and internal transfers count, then apply it the same way every period.

The cost of turnover is the full cost of replacing one leaver: recruiting spend, interviewing time, onboarding, the new hire's ramp period, manager time and any temporary cover.

Why it matters: It converts a percentage into a budget line, which is what turns a retention project into a funded one. Build the figure from your own salaries and time, not a headline multiple from an article.

The cost of a bad hire is what a hire who does not work out costs you in total, from sourcing and onboarding through lost productivity, manager time, exit and re-hiring.

Why it matters: It is the number that justifies slowing a hiring process down. It also sets a fair ceiling on what a better assessment or a longer interview loop is worth paying for.

Employee retention is a company's ability to keep the people it wants to keep, for as long as the business needs them.

Why it matters: It is the outcome every term on this page eventually rolls up to. Framed properly it is about the people you want to keep, not about keeping everyone.

The retention rate is the percentage of employees present at the start of a period who are still employed at the end of it.

Why it matters: It is the mirror of turnover but not its exact complement, and reading it as one is the most common reporting mistake in this area.

How it is measured: Employees present at both the start and the end of the period, divided by the headcount at the start, multiplied by 100. Anyone hired during the period is excluded from both sides, which is why retention rate and turnover rate rarely add to 100.

New-hire retention is the share of a hiring cohort still employed after a set milestone, commonly ninety days, six months or one year.

Why it matters: It is the fastest feedback loop you have on recruiting quality, job-description accuracy and onboarding, because you get an answer within months rather than years.

Average tenure is how long employees have been with the company, measured across a team, a level or the whole organization.

Why it matters: It is a stability read that also flags concentration risk: a short-tenure team is still learning, a team where all the knowledge sits with two long-tenured people is exposed if either leaves.

How it is measured: Report the median alongside the mean. A handful of very long-tenured founders or early employees pulls the mean up and hides a young, fast-moving population underneath it.

A stay interview is a structured conversation with a current employee about what keeps them here and what would make them consider leaving.

Why it matters: It gathers the same information as an exit interview while you can still act on it. Run with the right questions it is the highest-return retention habit a manager has.

An exit interview is a structured conversation with a departing employee about why they are leaving and what would have changed their mind.

Why it matters: It is the cheapest source of honest feedback you will ever get, and the least useful if nobody aggregates the answers. One interview is an anecdote, fifty tagged interviews are a pattern.

An exit survey is the written, usually anonymous form version of the exit interview, built from consistent questions so answers can be counted across many leavers.

Why it matters: Consistent questions are what make exit data analyzable. Free-text notes from twenty different managers cannot be compared, the same eight rated questions can.

A counteroffer is a revised offer, usually pay or title, made to an employee who has already resigned in order to keep them.

Why it matters: It treats the symptom at the worst possible moment and at the highest price. Knowing your own policy before a resignation lands is the point, because deciding under pressure produces the expensive answer.

A boomerang employee is a former employee who is rehired by the same company after working somewhere else.

Why it matters: Ramp time is short and the culture fit is already known, which makes alumni one of the cheapest talent pools you have. It only works if exits are handled well in the first place.

An EVP is the whole deal an employer offers in exchange for work: pay, growth, flexibility, the work itself, the people and the mission.

Why it matters: It is the thing a competing offer is compared against. If you cannot state yours in a sentence, your recruiters and your managers are each improvising a different one.

Internal mobility is people moving into new roles inside the company, sideways or upward, instead of leaving to get the same move elsewhere.

Why it matters: A large share of resignations are moves the person would have taken internally if the path had been visible. Internal mobility turns a departure into a transfer.

How it is measured: Internal hires as a share of all roles filled in the period. Track it alongside regretted attrition, because the two tend to move in opposite directions.

Flight risk is the likelihood that a specific employee leaves in the near future, based on observable signals rather than a manager's hunch.

Why it matters: It exists to prioritize. A manager has time for a handful of real retention conversations a month, and flight risk decides which handful. It is a prompt to talk to someone, never a verdict about them.

How it is measured: Score the known departure signals for one person and read them together rather than one at a time. A single signal is noise, several arriving in the same few weeks is a pattern.

A resignation signal is an observable change that tends to precede someone giving notice, such as a public profile refresh, a withdrawal from optional work, or a shift in visible career activity.

Why it matters: Signals buy time. The gap between the first signal and the resignation letter is the only window in which a retention conversation can still change the outcome.

A predictive attrition model estimates which employees are most likely to leave in a coming period, from patterns that preceded past departures.

Why it matters: A model ranks attention, it does not make decisions. Treat the output as a queue for manager conversations, and never as an input to pay, promotion or performance decisions.

Open to Work is a public LinkedIn setting that tells recruiters a person is open to new roles. Shown publicly, it is one of the most direct job-search signals a profile can carry.

Why it matters: It is public and deliberate: the person chose to broadcast it. Treated as a prompt for a genuine conversation about growth or workload, it is useful. Treated as a disciplinary trigger, it destroys the trust that retention depends on.

Quiet quitting is doing the job as written and no more: no extra hours, no optional projects, no volunteering beyond the role.

Why it matters: It is usually disengagement showing up as withdrawn effort rather than as a resignation, and it often arrives months before one. It can also be a healthy boundary, so the response is a conversation, not a policy.

Quiet firing is making someone's job worse until they resign: withheld growth, shrinking scope, cancelled one-on-ones, no feedback.

Why it matters: It converts a decision you should be making openly into a resignation, and it is visible to everyone else on the team, which damages the retention of the people you did want to keep.

Quiet hiring is covering new work by moving or stretching existing employees instead of opening a role.

Why it matters: It is a legitimate way to fill a gap quickly and a reliable way to burn people out when the stretch has no end date, no title change and no pay change attached.

Moonlighting is holding paid work outside your main job, whether freelance, contract or a second employer.

Why it matters: It can signal a pay gap, an engagement gap, or simply a side interest. Which one it is decides whether it is a retention issue at all, so it is worth asking before assuming.

A toxic work environment is a workplace where the day-to-day norms, usually set by managers, reliably damage the people in it.

Why it matters: It is among the most consistent drivers of voluntary exits, and it is a local condition. One team can be toxic inside an otherwise healthy company, which is why turnover has to be read per manager, not only company-wide.

Employee engagement is how committed and involved people are in their work and their company, beyond simply doing the tasks assigned to them.

Why it matters: It is the condition that sits upstream of retention, output and referrals. It is also frequently confused with satisfaction: a satisfied employee is comfortable, an engaged one is invested.

eNPS is a single-question engagement measure: how likely an employee is, on a zero to ten scale, to recommend the company as a place to work.

Why it matters: One question gets answered, so you get a trend rather than a survey nobody finished. The trade-off is that it tells you the direction, never the reason, so it needs a follow-up question next to it.

How it is measured: The percentage of promoters (scores of 9 or 10) minus the percentage of detractors (scores of 0 to 6). Passives (7 and 8) count in the base but not in the score.

A pulse survey is a short, frequent survey, often a handful of questions sent monthly or quarterly, used to track sentiment between full engagement surveys.

Why it matters: Frequency is the point: an annual survey tells you about a problem long after the people affected have made their decision.

Survey fatigue is the fall in response rate and answer quality that follows asking people the same questions too often, especially when nothing visibly changes afterwards.

Why it matters: It is the reason survey-based retention data decays. Once the people most unhappy stop responding, the scores improve while the underlying situation does not.

Absenteeism is unplanned absence from work beyond normal, approved leave.

Why it matters: It is one of the few disengagement signals that lands in a system you already have. A rise concentrated in one team is worth reading as a management signal rather than a personal one.

How it is measured: Unplanned absence days divided by scheduled working days for the same population and period.

A one-on-one is a recurring private conversation between a manager and a direct report about the work, the obstacles and the person's growth.

Why it matters: It is where nearly every retention signal surfaces first. A cancelled one-on-one is not a saved half hour, it is the main channel where you would have heard about a problem.

A skip-level meeting is a conversation between an employee and their manager's manager, without the direct manager present.

Why it matters: It surfaces the problems people will not raise with the person who causes them, which is exactly the class of problem that ends in a resignation.

People analytics is the use of workforce data to answer business questions about hiring, retention, performance and organizational design.

Why it matters: Done well it replaces opinion about the workforce with evidence. Done badly it produces dashboards nobody acts on, so start from a decision somebody has to make.

Retention metrics are the recurring measures a people team reports on keeping staff: turnover and retention rates, regretted attrition, tenure, new-hire retention and engagement scores.

Why it matters: A small set reported consistently beats a large set reported once. Fix the definitions first, because a metric that changes meaning between quarters is worse than no metric.

A leading indicator moves before the outcome you care about. In retention, engagement scores, absenteeism and flight risk signals lead, while the turnover rate lags.

Why it matters: A lagging metric can only confirm what already happened. If your retention reporting contains only lagging metrics, every meeting about it is a post-mortem.

Headcount is the number of employees on the payroll at a point in time. Average headcount is the average across a period, and it is the denominator in most turnover and retention formulas.

Why it matters: The denominator decides the answer. Using a start-of-year headcount in a fast-growing company understates turnover, using an end-of-year headcount overstates it, and the average is the fair middle.

How it is measured: Add the starting and ending headcount for the period and divide by two. In a fast-growing company, average the monthly headcounts instead.

Span of control is the number of direct reports a manager has.

Why it matters: It quietly sets the ceiling on management quality. Past a certain width, one-on-ones get cancelled and growth conversations stop happening, and both show up in turnover a few quarters later.

How it is measured: Total employees divided by total managers, then look at the distribution rather than the average. The averages usually look fine while a few managers carry impossible spans.

Cohort analysis groups employees by a shared starting point, such as hire month, level or location, and follows each group over time.

Why it matters: It is how you tell a real change from a mix change. A company-wide retention drop is often one bad hiring cohort or one team, and only cohorts make that visible.

A benchmark is a reference value from outside your company, used to judge whether your own number is high or low.

Why it matters: Turnover benchmarks vary enormously by industry, role, seniority and region, and published figures rarely state which population they measured. Your own trend over time is a more reliable comparison than a benchmark whose definition you cannot check.

Succession planning is identifying who could step into critical roles, and closing the gaps before the seat is empty.

Why it matters: It converts a key departure from a crisis into a plan. The exercise also surfaces single points of failure while you still have time to fix them.

The 9-box grid is a talent review tool that places each employee on a three by three matrix of current performance against future potential.

Why it matters: It forces a shared, explicit judgment in a talent review instead of nine managers each using a private definition of top talent. The boxes are the start of a conversation, not a label to keep.

An HRIS (Human Resource Information System) is the system of record for employee data: contracts, roles, reporting lines, pay and leave.

Why it matters: It is where headcount, tenure and separation dates come from, so its data quality caps the accuracy of every metric on this page.