Guide

Recruitment metrics: time to hire, cost per hire, and the ones that matter

Recruitment metrics exist to answer three questions: how fast do we hire, what does it cost, and does it work? Most recruiting dashboards answer the first two and dodge the third. The set below covers all three with formulas and — the part usually missing — the decision each number should trigger. Track them per role and per quarter; recruiting numbers only mean anything as trends and comparisons.

Speed metrics

Time to hire = days from job opening to accepted offer. The candidate-experience number: long times lose the best people to faster offers. Rising trend → find the stalled stage with a funnel review.
Time to fill = days from opening to start date. The operations number: it is what the business feels, since the work is uncovered until the start date.
Stage timing — days spent in each stage (screening, interviews, decision, offer). This is where the fix lives: overall time to hire tells you there is a delay; stage timing tells you whose calendar it is hiding in. The decision stage is the usual culprit.

Cost and source metrics

Cost per hire = total recruiting costs (ads, agencies, referral bonuses, tools, recruiter time) ÷ hires in period. Watch the trend; the absolute number varies legitimately by role seniority.
Source of hire — which channel produced each hire. Almost every company that starts tracking this discovers it is overspending on its weakest channel.
Source quality — first-year retention and performance by source. Referrals typically win; the interesting finding is which paid channel earns its cost. This metric requires connecting recruiting data to HR records — which is an argument for keeping both structured.

Quality metrics — the ones that make it honest

Offer acceptance rate = offers accepted ÷ offers made. Below roughly 80%, candidates are discovering something at offer time they dislike — pay positioning, process experience, or a gap between posting and reality.
Funnel conversion — applicants → screened → interviewed → offered → hired. Anomalies localize problems: many applicants but few screen-passes means a mistargeted posting; many finalists but few offers means an indecisive panel.
First-year retention of hires (quality of hire’s honest proxy) — a fast, cheap process that produces hires who leave in month eight is an expensive process wearing a good dashboard. This is the metric that keeps the others truthful.

Making the numbers cheap to produce

Every metric above requires only that dates and outcomes be recorded when they happen: opening dates, stage transitions, sources, offer dates, start dates — and then employment outcomes from the HR side. Companies that keep structured records get these numbers as filtered views; companies that do not, reconstruct them annually in a spreadsheet archaeology project and quietly stop. One process note: verification is a common hidden stage-delay, worth timing separately — and worth compressing; on EmployDB, candidates with verified histories turn the verification stage from days of employer callbacks into a same-day consent grant, which shows up directly in your time-to-hire trend.