HR Metrics

HR metrics are the handful of ratios that describe whether an organisation is keeping the people it hires, hiring them at a sensible cost and speed, and getting them to work. There are hundreds in the textbooks. About eight are used by anyone who has to defend a budget: turnover, voluntary turnover, first-year turnover, retention, cost per hire, time to hire, offer acceptance and absence. Each has a standard formula, and each is routinely calculated wrong in ways that make comparison with anyone else's number meaningless.

The calculator below produces all eight from the numbers a payroll or HR system already holds, with the 2025 benchmarks alongside. The rest of the page gives the formulas, the definitional choices that change the result, what the benchmarks actually measure, and the metrics that look like insight and are not.

HR metrics calculator

The formulas and the choices inside them

Two of these are worth expanding. Retention is not the mirror of turnover: a team that started the year with 100 people, hired 30 and lost 20 of the new hires has 20 per cent turnover on average headcount and 100 per cent retention of the original staff. Both numbers are true and they describe different problems, which is why first-year turnover exists as a separate measure. Cost per hire, under the SHRM and ANSI standard, includes internal costs that most organisations never tally: the recruiter's salary share, the hiring manager's interview hours, the referral bonus, the applicant tracking system. An organisation that reports $1,800 per hire is usually reporting agency fees.

What the 2025 benchmarks measure

• Turnover. The SHRM 2025 benchmarking reports put median voluntary turnover at 9 per cent and the average at 13, across more than 2,000 organisations, with the gap showing how much a minority of high-churn employers pull the average up. The Job Openings and Labor Turnover Survey gives a different cut: the 2025 annual average total separations rate was 3.3 per cent of employment per month, of which quits were 2.0, layoffs and discharges 1.1. Multiply a monthly JOLTS rate by twelve and you get economy-wide turnover of roughly 40 per cent, which is not comparable to SHRM's survey of established employers. Use SHRM for a company benchmark and JOLTS for the labour market context.
• Cost per hire. SHRM 2025: $5,475 average for non-executive hires, $35,879 for executives, using the SHRM and ANSI cost standard. Earlier SHRM editions gave $4,700 and $4,129; the rise reflects both inflation and more complete cost counting.
• Time to fill. SHRM 2025: median 44 days for non-executive roles, 45 for executive, measured from requisition to acceptance. Technology and healthcare roles run longer; hourly and retail roles shorter.
• Absence. The Bureau of Labor Statistics reported a 3.2 per cent absence rate and 1.7 per cent lost worktime rate for full-time workers in 2025. The absence rate counts workers absent in the survey week; the lost worktime rate counts hours, and is the closer match to a days-lost calculation. Our absenteeism guide goes deeper.

Benchmarks describe where the middle of a distribution sits. They do not say what is good for a particular business. A software company with 5 per cent voluntary turnover may be losing exactly the people it cannot replace; a call centre at 35 per cent may be running at the industry's floor. The comparison that matters is the same metric, defined the same way, in your own organisation over time.

Cutting the numbers so they say something

A company-wide turnover rate hides the team with the problem. The cuts that regularly find it are by manager, by tenure band, by location, by job family and by month of the year. Voluntary turnover concentrated in one manager's team, in the first year, or in the weeks after annual bonuses pay out, each points to a different cause and a different fix. Regretted versus non-regretted turnover, where the manager records at exit whether they would have kept the person, separates the loss that matters from the loss that was overdue.

Exit and stay interviews are where the numbers acquire reasons. Our exit interview questions and stay interview questions pages cover both, including why exit answers are systematically flattering. Engagement survey results cut the same way, by manager and tenure, predict the turnover figure six months out better than the turnover figure predicts itself; the survey builder is the place to start.

Hiring metrics cut by source: cost per hire and first-year turnover by agency, job board, referral and internal move. Referrals are usually cheapest and stay longest; agencies usually cost most and stay shortest; the numbers will tell you whether that holds for you, which is the point of tracking them.

Metrics that look like insight

• HR-to-employee ratio. Widely benchmarked, meaningless without knowing what HR does in the organisation. A company that outsources payroll and benefits has a different ratio from one that does not, and neither number says anything about quality.
• Revenue per employee. A finance metric that moves with pricing, capital intensity and outsourcing far more than with anything HR controls.
• Training hours per employee. Measures attendance at training, not learning. If a metric is needed, track the skill or certification the training was for.
• Engagement score as a target. Once a manager's bonus depends on the team's engagement score, the score measures the manager's ability to campaign for it. Use it as a diagnostic, cut by team, and never as a target.
• Productivity from activity data. Keystrokes, active hours and application time measure presence at a computer. Our employee productivity formula guide covers what to measure instead, and the monitoring laws that apply to collecting the data at all.
• Any metric with no owner and no decision attached. If nobody can say what they would do differently if the number moved, stop calculating it. The eight in the calculator each attach to a decision: who to keep, how to hire, where the manager problem is, and whether absence is a cost worth acting on.

Key takeaways

• Eight metrics do most of the work: turnover, voluntary turnover, first-year turnover, retention, cost per hire, time to hire, offer acceptance and absence rate.
• Turnover is leavers over average headcount; retention is original staff still employed. They are not mirror images, and first-year turnover is the gap between them.
• Cost per hire under the SHRM and ANSI standard includes internal costs. SHRM's 2025 average is $5,475 for non-executive hires; figures far below that are usually agency fees alone.
• SHRM 2025 median voluntary turnover is 9 per cent (average 13); median time to fill 44 days; BLS absence rate 3.2 per cent. Definitions differ, so compare your own trend first.
• Cut every metric by manager, tenure, location and source. The company average hides the team with the problem.
• Drop metrics with no owner and no decision attached. HR-to-employee ratio, training hours and activity-based productivity are the usual candidates.