HR metrics are numbers that describe the workforce and the work HR does: how many people you have, how many leave, how long hiring takes, what it costs and whether people are paid and treated consistently. This guide gives the formula for each common metric, says what the result tells you, walks through one quarter of numbers for a 200-person company and ends with a dashboard template you can copy.
The formulas are standard definitions, and the example figures are made up so you can check the arithmetic yourself. If you've got an HRIS, you'll find most inputs already there. Where an outside source sets the definition, it's linked. We don't quote industry benchmarks, because they change and they're tied to your sector and size, which isn't something a general guide can know.
- Start from the question leadership asks (are we losing people, are we hiring fast enough, is pay consistent), then pick 6 to 10 metrics that answer it.
- Most rate metrics need a denominator. Turnover, hires and separations are divided by average headcount, so fix that definition once and keep it; you'll thank yourself in a year.
- Write down the formula, the source system, the owner and the review cadence for each metric. A number without a definition can't be compared next quarter, and that's the usual reason trends get disputed.
- Small teams produce noisy rates, and that's not a flaw in the formula. One departure in a team of 20 moves annual turnover by 5 points, so read counts next to percentages.
- Selection metrics such as the adverse impact ratio are compliance tools. Treat a result below 80% as a prompt to investigate, not as a verdict.
What Are HR Metrics and How Do They Differ from HR Analytics?
HR metrics are quantified measures of workforce outcomes and HR processes, such as turnover rate, time to fill and cost per hire. A metric is a single calculated number, but it says little alone. A KPI is a metric you've tied to a target and an owner. HR analytics is the work of combining metrics to explain why a number moved and what will happen next.
Analytics is usually described in four types, which also answers a common search question:
- Descriptive analytics reports what happened, for example 9 people left last quarter.
- Diagnostic analytics explains why, for example that 5 of the 9 reported to the same manager.
- Predictive analytics estimates what's likely next, for example which roles are at risk of leaving.
- Prescriptive analytics recommends an action, for example which fix to try first.
Five common examples are turnover rate, time to fill, cost per hire, absence rate and eNPS. Each is a count or ratio with a stated denominator, and each has a limit: turnover ignores who left, time to fill ignores how well the hire works out, and eNPS says little without the reasons behind the scores.
Most teams need descriptive and diagnostic work long before predictive models, so don't start with the model. You'll rarely have enough clean data for one anyway, and it's the basics that fail first. The metrics below are all descriptive building blocks.
Timing matters because people rarely say they're unhappy until they have decided to go. Gallup surveyed 717 U.S. employees who had voluntarily left a job in the previous 12 months. 42% said their manager or organization could have done something to prevent the departure, and 45% said no manager or other leader had proactively discussed their job satisfaction, performance or future in the three months before they left. A turnover rate tells you the loss happened. It doesn't tell you the conversation was missing, and that gap is what most teams could fix.
For organizations that report workforce data to boards or investors, ISO 30414 is the international guideline for internal and external human capital reporting. Its contents differ between editions, so check the edition you're reporting under before adopting its metric list.
Which HR Metrics Should You Track?
Track 6 to 10 metrics (that's plenty), each chosen because it answers a question someone will act on, rather than because a tool offers it. A metric nobody can act on is decoration, not insight, and a long list makes the few that matter harder to see.
| Question you're asked | Metrics that answer it | Where the data usually lives |
|---|---|---|
| Are we losing people we want to keep? | Turnover rate, voluntary turnover rate, 90-day retention, regretted attrition | HRIS |
| Are we hiring fast enough and at a sensible cost? | Time to fill, cost per hire, offer acceptance rate | Applicant tracking system, finance |
| Are new hires set up to succeed? | Onboarding completion before day one, time to productivity | Onboarding checklist, manager survey |
| Is the team engaged and developing? | eNPS, review completion rate, training hours per employee | Survey tool, performance system, LMS |
| Is pay consistent and affordable? | Compa-ratio, overtime rate, revenue per employee | Payroll, finance |
| Are we treating people consistently and lawfully? | Adverse impact ratio, representation, compliance training completion | ATS, HRIS |
Aim for one or two metrics per question, which is enough to show a trend without burying the reader. If you can't name who reads the number and what they'd do differently, drop it.
HR Metrics Formulas by Category
Each table gives the formula and what the result tells you; it's meant to be skimmed. Rates are multiplied by 100 to give a percentage, though a few ratios (compa-ratio) stay as decimals. Where a government source sets the definition, the table says so, and you'll find the link there.
Headcount and Turnover Metrics
The denominator for most rates in this section is average headcount: (headcount at the start of the period + headcount at the end) / 2. The Bureau of Labor Statistics computes its hires and separations rates by dividing the number of hires or separations by employment, and defines total separations as quits plus layoffs and discharges plus other separations.
| Metric | Formula | What it tells you |
|---|---|---|
| Average headcount | (Start headcount + end headcount) / 2 | The denominator for the rates below |
| Turnover rate | Separations in the period / average headcount × 100 | How much of the workforce left, voluntary and involuntary together |
| Voluntary turnover rate | Resignations in the period / average headcount × 100 | Departures the company could have influenced |
| Retention rate | (End headcount − new hires in the period) / start headcount × 100 | Share of the starting group still employed |
| 90-day retention | New hires still employed at day 90 / new hires in the cohort × 100 | Whether hiring and onboarding hold up in the first quarter |
| Regretted attrition | Departures you wanted to keep / average headcount × 100 | Whether you're losing the right or the wrong people |
To compare periods of different length, annualize, but be careful: a quarterly rate multiplied by 4 gives an approximate yearly rate. Say that you did it, because a single bad quarter overstates the year, and recalculate on full-year departures and average headcount when you can, since the two results can differ. Readers who don't know it's annualized will compare it with last year's full figure.
There's no universal good turnover rate. It varies by industry, role mix and company size, so compare against your own last four quarters first, then against a peer group that defines turnover the way you do.
Regretted attrition needs a rule written before anyone leaves, and it shouldn't be left to whoever's asked, for example a rating threshold or a critical-skill flag. If managers decide afterwards, the number follows opinion, and it's hard to defend.
Recruiting Metrics
| Metric | Formula | What it tells you |
|---|---|---|
| Time to fill | Days from approved requisition to offer accepted, averaged over filled roles | How long a vacancy stays open |
| Time to hire | Days from candidate's application to offer accepted, averaged | How fast the process moves a candidate |
| Cost per hire | (Internal recruiting costs + external recruiting costs) / hires in the period | What each hire costs in fees, ads, tools and recruiter time |
| Offer acceptance rate | Offers accepted / offers extended × 100 | Whether the offer, the pay and the process win candidates |
| Applicants per opening | Applications received / open roles | Demand for the role and the quality of the posting |
| Quality of hire | Defined by you, often a blend of 6-month performance rating, retention and manager satisfaction | Whether hires turn out well; there is no single standard formula |
Define the start and end dates once; it's a small step that saves arguments later. Time to fill and time to hire are often used as if they mean the same thing, yet they differ by the time before a candidate applies. Decide which one you report and label it.
Onboarding and Offboarding Metrics
These metrics aren't in most generic HR metrics lists, yet they're where cross-team work shows up. They depend on IT, finance and managers rather than on HR alone. For the process behind them, see our guide to the employee onboarding process.
| Metric | Formula | What it tells you |
|---|---|---|
| Onboarding completion before day one | Pre-start tasks done before the start date / pre-start tasks assigned × 100 | Whether the laptop, accounts and paperwork are ready when the person arrives |
| Time to productivity | Days from start date to the milestone you define, such as first independent task or quota attainment | How long a new hire takes to contribute; the milestone is your choice |
| New-hire survey score | Average score on a 30, 60 or 90 day pulse survey | How the person experiences the first months |
| Exit interview completion rate | Exit interviews held / voluntary departures × 100 | Whether you collect leaver feedback consistently; see exit interview questions |
| Access removal time | Hours from last day to last system account disabled | Whether offboarding closes security gaps |
The first metric is the easiest to improve, because it's a count of tasks with owners. A new-hire checklist makes it measurable; our new hire onboarding template has one you can copy.
Performance, Engagement and Development Metrics
| Metric | Formula | What it tells you |
|---|---|---|
| eNPS | % promoters (scores 9 and 10) − % detractors (scores 0 to 6) on "How likely are you to recommend this company as a place to work?" | A single engagement signal that can be repeated every quarter |
| Review completion rate | Reviews submitted on time / reviews due × 100 | Whether the performance process is actually running |
| Goal attainment rate | Goals marked met / goals set × 100 | How realistic goals are; treat very high and very low values with suspicion |
| Training hours per employee | Total training hours / average headcount | Learning activity, not learning quality |
| Internal mobility rate | Internal moves and promotions / average headcount × 100 | Whether people can grow without leaving |
eNPS is a convention borrowed from Net Promoter Score, so report the question wording with the number. A complete picture of performance tools is in our guide to performance management software.
Pay and Cost Metrics
| Metric | Formula | What it tells you |
|---|---|---|
| Compa-ratio | An employee's pay / midpoint of the pay range for the role | Where someone sits in their range; 1.00 is the midpoint |
| Revenue per employee | Revenue / average headcount | Output per head; compare only with your own history or close peers |
| Overtime rate | Overtime hours / total paid hours × 100 | Staffing strain or scheduling problems |
| Labor cost ratio | Total labor cost / revenue × 100 | How much of revenue goes to people |
| HR cost per employee | Total HR function cost / average headcount | What it costs to run HR; compare with your own history |
Time and Attendance Metrics
| Metric | Formula | What it tells you |
|---|---|---|
| Absence rate | Employees with an absence in the period / total employees × 100 | How common absence is. The BLS Current Population Survey defines it as workers with absences divided by total full-time wage and salary employment |
| Unscheduled absence rate | Unplanned absence days / scheduled workdays × 100 | Sudden gaps that scheduling can't absorb |
| Schedule adherence | Hours worked on schedule / hours scheduled × 100 | Whether rosters match what happens |
| Coverage | Shifts staffed to plan / shifts planned × 100 | Whether you can run the operation as designed |
Hourly and shift teams read these closely, since they can't ignore gaps in coverage. See employee scheduling software and workforce management software for the tools that produce the data.
Diversity and Compliance Metrics
| Metric | Formula | What it tells you |
|---|---|---|
| Representation | Employees in a group / total employees × 100, by level | Who is in the workforce and at which levels |
| Selection rate | People selected from a group / applicants from that group × 100 | The hiring or promotion rate for that group |
| Adverse impact ratio | Selection rate of a group / selection rate of the group with the highest rate × 100 | A screening test for uneven selection |
| Pay gap | (Median pay of group A − median pay of group B) / median pay of group A × 100 | Whether pay differs between groups, before controlling for role and tenure |
| Compliance training completion | Employees who completed required training / employees required × 100 | Whether mandatory training is done |
The adverse impact ratio comes from the U.S. Uniform Guidelines on Employee Selection Procedures. Under 29 CFR 1607.4(D), a selection rate for any race, sex or ethnic group that is less than four-fifths (80 percent) of the rate for the group with the highest rate is generally regarded as evidence of adverse impact. It's a rule of thumb for flagging, not a legal conclusion, although small samples can mislead. Involve counsel before acting on a result; don't treat the ratio as a finding.
HR Metrics Example: One Quarter for a 200-Person Company
The numbers below are invented to show the arithmetic.
| Input | Value |
|---|---|
| Headcount at start of quarter | 196 |
| Headcount at end of quarter | 204 |
| New hires | 17 |
| Separations (6 resignations, 3 involuntary) | 9 |
| Roles filled and days to fill | 41, 52 and 37 days |
| Total recruiting cost | $68,000 |
| Offers extended and accepted | 21 and 17 |
| Hires from last quarter's cohort / still employed at day 90 | 14 / 12 |
| Metric | Calculation | Result |
|---|---|---|
| Average headcount | (196 + 204) / 2 | 200 |
| Turnover rate | 9 / 200 × 100 | 4.5% for the quarter |
| Annualized turnover | 4.5% × 4 | About 18% |
| Voluntary turnover rate | 6 / 200 × 100 | 3.0% for the quarter |
| Time to fill | (41 + 52 + 37) / 3 | 43.3 days |
| Cost per hire | $68,000 / 17 | $4,000 |
| Offer acceptance rate | 17 / 21 × 100 | 81.0% |
| 90-day retention | 12 / 14 × 100 | 85.7% |
Here are two readings from these numbers, one on rates and one on sample size. First, 4.5% in a quarter sounds modest, but annualized it's 18%, so say which one you mean. Second, the 90-day figure rests on 14 people: two early leavers moved it by 14 points. Show the counts beside the percentages.
Now a selection check, which isn't a turnover metric but uses the same arithmetic. Suppose 40 applicants from group A and 30 from group B were interviewed, and 20 from group A and 9 from group B advanced. The selection rates are 50% and 30%. The adverse impact ratio is 30 / 50 × 100 = 60%, which is below 80%, so the stage deserves a review of the criteria and the process.
How to Set Up HR Metrics Reporting
Reporting works when definitions are written down, not when the dashboard looks good; it doesn't depend on the tool. Work through these steps in order; it's quicker than fixing a dashboard later.
Start from the question
List the three or four questions leadership asks every quarter. Pick the metrics from the table above that answer them and drop the rest; you don't need them all.
Write the definition
For each metric, write the formula, the time window, who counts as an employee and the date field used. Decide whether contractors and interns are in or out.
Name the source of truth
Each number comes from one system: the HRIS for headcount and separations, the applicant tracking system for time to fill, payroll for pay. Don't rebuild the same number in two spreadsheets; they'll drift apart.
Assign an owner and a cadence
One person is responsible for each number and for explaining a move, so nobody says it's someone else's. Monthly works for most counts; engagement and pay metrics are fine quarterly.
Add a review trigger
Decide ahead of time what change prompts a conversation, so you aren't improvising, such as voluntary turnover above your own twelve-month average by a set margin. Write the threshold next to the metric.
Common HR Metrics Mistakes
Most bad HR reporting comes from a short list of avoidable errors, not from bad intent. Each one has a simple fix.
- Changing definitions midyear. Although it feels harmless, if time to fill starts at requisition approval in Q1 and at posting date in Q2, the trend is meaningless. Version your definitions so you can say what changed and when. It's a five-minute habit.
- Reading rates on small samples. One resignation in a 20-person team is 5 percentage points of turnover. Show counts and avoid ranking small teams; they're too noisy to compare.
- Mixing voluntary and involuntary exits. A restructuring spikes total turnover without saying anything about retention. Report them separately, since they've different causes and different fixes.
- Treating training hours as learning. Hours measure activity, however, not skill. Pair them with a check that skills were used, because that's what the training was for.
- Collecting sensitive data without a purpose. Diversity, health and absence data are personal, but teams often gather them first and ask why later. Collect only what a decision needs, don't hoard the rest, restrict who can see it and check local privacy rules.
HR Metrics Dashboard Template
You'll want this in a spreadsheet, so copy this table and fill one row per metric you've chosen. The last two columns are the ones most dashboards leave out, and they're the ones that make a number actionable.
| Metric | Formula | Source system | Owner | Cadence | Review trigger |
|---|---|---|---|---|---|
| Turnover rate | Separations / average headcount × 100 | HRIS | People Ops | Monthly | Above your 12-month average by the margin you set |
| Voluntary turnover rate | Resignations / average headcount × 100 | HRIS | HR business partner | Monthly | Two or more resignations from one team in a quarter |
| Time to fill | Days from requisition approval to accepted offer | ATS | Talent acquisition | Monthly | Open longer than your target for a role family |
| Cost per hire | (Internal + external recruiting cost) / hires | ATS and finance | Talent acquisition | Quarterly | Rises two quarters in a row |
| Onboarding completion before day one | Pre-start tasks done / assigned × 100 | Onboarding checklist | People Ops | Per cohort | Any hire starting with open IT tasks |
| 90-day retention | Still employed at day 90 / cohort × 100 | HRIS | People Ops | Quarterly | Any early leaver without an exit conversation |
| eNPS | % promoters − % detractors | Survey tool | Head of People | Quarterly | Drop of 10 points or more |
| Adverse impact ratio | Group selection rate / highest selection rate × 100 | ATS | People Ops with counsel | Per hiring stage, twice a year | Below 80% |
- Every metric has a written formula and time window
- Headcount excludes or includes contractors by one stated rule
- Counts appear next to every percentage
- Each metric has a named owner and a source system
- Review triggers are written down before the numbers come in
- Sensitive metrics have restricted access
Where The People Ops Manager Fits
HR Metrics FAQ
What are HR metrics and what are some examples?
HR metrics are calculated measures of workforce outcomes and HR processes. Common examples are turnover rate, time to fill, cost per hire, offer acceptance rate, absence rate and eNPS. The formulas are in the tables above.
What are the four types of HR analytics?
Descriptive (what happened), diagnostic (why it happened), predictive (what's likely to happen) and prescriptive (what to do about it). Most teams get more value from descriptive and diagnostic work first than from predictive models.
What's the difference between an HR metric and an HR KPI?
A metric is any calculated number. A KPI is a metric tied to a target, an owner and a decision. Turnover rate is a metric; voluntary turnover held under a target set by leadership and owned by a named person is a KPI.
How many HR metrics should you track?
Most companies do well with six to ten, picked for the questions leadership actually asks. Add more only when a specific decision needs them.
What are the 7 C's of HR?
There's no single standard list under that name, and sources define it differently. Check which framework a source means before using it, and don't treat it as a set of measurable metrics.
How do you calculate employee turnover rate?
Divide the number of separations in the period by average headcount and multiply by 100. Average headcount is the start and end headcount added together and divided by two. For an annual figure from a quarterly one, multiply by four and say that you did.