Recruiting without metrics is vibes with a careers page. You do not need a forty-KPI dashboard. You need a short stack that answers: How fast do we hire, what does it cost, do offers stick, which channels work, are hires good, and do they stay?
These six cover that loop. Define them the same way every quarter or the chart is fiction.
Time-to-fill
What it is: Calendar days from req approval (or open date) to accepted offer. Some teams also track time-to-start.
Why it matters: Empty seats have a cost. Slow loops lose candidates. This is the speed metric executives actually ask for.
Watch for: Gaming by delaying req open until a candidate is ready, or pausing without stopping the clock. Split time-to-fill from time-in-process (candidate entered funnel → offer) so you can see recruiter speed vs approval drag.
Healthy use: Trend by job family. A 60-day average that hides 20-day support hires and 120-day specialist searches is useless.
Cost-per-hire
What it is: Total recruiting cost for a period ÷ number of hires. Include job ads, agency fees, tools, events, referral bonuses, and (if you are serious) recruiter labor allocation.
Why it matters: Channels look cheap until you count agency spend. Internal recruiters look expensive until you count what agencies would have billed.
Watch for: Excluding agency fees from “our” cost-per-hire, or counting only media spend. Also separate executive and volume roles or one VP search wrecks the average.
Healthy use: Compare channel cost against quality and retention, not against a vanity low number.
Offer acceptance rate
What it is: Accepted offers ÷ offers extended (in a period).
Why it matters: Low acceptance means late surprises: pay, level, role reality, process length, or competing offers. High acceptance with bad quality means you are closing the wrong people.
Watch for: Verbal offers that never go written, or counting reneged accepts as accepts. Track decline reasons in the same report.
Healthy use: Segment by level and whether the band was shared early. Surprises kill accept rates.
