Guide

HR KPIs and metrics every company should track

HR metrics turn people management from anecdotes into decisions. The trap is tracking too many: a dashboard of thirty numbers is a dashboard nobody reads. The metrics below are the working core — each has a simple formula, a clear question it answers, and a decision it should trigger. Track these consistently for a year and most HR debates in the company become short.

Retention metrics

Turnover rate = departures in period ÷ average headcount × 100. The overall health number. Compare against your industry, not against zero — retail and hospitality run structurally higher than office work.
First-year turnover = departures within 12 months of hire ÷ hires × 100. The most diagnostic number in HR: high first-year turnover means hiring or onboarding is broken, not culture.
Regrettable turnover — departures you would have prevented. A 15% turnover rate means something entirely different if the leavers are your best people versus your weakest.

Hiring metrics

Time to hire = days from job opening to accepted offer. Long times cost output and lose candidates to faster competitors; sudden increases usually mean an undefined role or a slow decision-maker.
Cost per hire = total recruiting cost (ads, agencies, referral bonuses, recruiter time) ÷ hires. Watch the trend and the per-channel breakdown more than the absolute number.
Offer acceptance rate — below roughly four in five, your offers, salary positioning, or process experience have a problem candidates can see and you cannot.

Productivity and reliability metrics

Time to productivity — days from start date until the employee performs the core role independently; define the milestone per role. This is the metric structured onboarding moves.
Absenteeism rate = unplanned absence days ÷ scheduled days × 100. Spikes cluster around teams and managers before they show anywhere else.
Training completion and certification currency — for regulated roles, the percentage of staff with mandatory certificates valid today. This number is either 100% or a finding waiting for an inspector.

Making the numbers trustworthy

Every metric above is only as good as the records underneath it — turnover requires accurate start and end dates, first-year turnover requires them per person, certification currency requires expiry dates recorded at all. This is the practical case for structured HR records: in EmployDB, employment events carry effective dates and certificates carry their documents, so the metrics compute from facts rather than from someone’s reconstruction. Review the numbers quarterly, segmented by team; averages hide exactly the problems metrics exist to find.