Recruiting Metrics That Actually Matter (And the Ones That Don't)

1. The Metrics That Predict Hiring Health

These are your leading indicators. They move weeks before a missed hiring goal shows up, which means they are the only numbers that give you time to fix anything.

• ✓
  Pipeline velocity per stage. Not overall time to hire — the days candidates spend in each stage. A 40-day process usually hides one stage eating 12 of those days. If resumes sit unreviewed for 6 days or scheduling adds 9, you know exactly where to intervene, which is the whole point of reducing time-to-hire without cutting evaluation.
• ✓
  Pass-through rates between stages. What percentage of screened candidates reach a first interview, and what percentage of finalists get offers? Healthy screen-to-interview conversion typically runs 25 to 40 percent for well-targeted roles. A 5 percent rate means your top of funnel is broken; a 90 percent rate means your screen is not filtering at all.
• ✓
  Offer-accept rate and qualified-candidates-per-opening. Aim for 85 to 90 percent on offers; below 80 percent signals a compensation, speed, or expectations problem that no amount of sourcing fixes. And track how many candidates per req actually meet the bar — 4 to 6 qualified candidates per opening is a workable pipeline; 40 unqualified applicants is not.

2. Quality of Hire: The Metric Everyone Wants, Few Measure

Ask any talent leader which metric matters most and they will say quality of hire. Ask how they measure it and the room goes quiet. It lags by months and requires data from outside the ATS, but it is measurable — you just have to build the loop deliberately.

• ✓
  Track 90-day and 1-year retention per role family. A hire who leaves inside 90 days is almost always a screening or expectations failure, not a performance one. If more than 10 percent of hires in a role family wash out in the first quarter, the problem is upstream in your funnel, not in the people.
• ✓
  Survey hiring managers at day 90 with one question. “Knowing what you know now, would you make this hire again?” on a 1-to-5 scale. It takes 30 seconds, response rates are high because managers have opinions, and averaged across a quarter it becomes your most honest quality signal.
• ✓
  Correlate performance against screening scores. Pull your first-performance-review outcomes and line them up against the scores candidates received at screening. If your top-rated candidates are not becoming your top performers, your screening criteria are measuring the wrong things — and you now have the evidence to change them.

3. Funnel Diagnostics: Find Where and Why Candidates Drop

Aggregate conversion tells you that candidates are disappearing; diagnostics tell you where, why, and whose problem it is. This is the difference between a report and an instrument.

• ✓
  Separate rejections from withdrawals at every stage. Candidates you reject and candidates who ghost you are opposite problems. High rejection rates at screen mean poor top-of-funnel targeting; high withdrawal rates mid-process almost always mean you are too slow — strong candidates are accepting offers elsewhere while yours are waiting on interview three.
• ✓
  Measure source quality, not source volume. The job board that sends 300 applicants and 0 hires is worse than the referral channel that sends 12 and 3. Track hires and onsite-stage candidates per source, per 100 applicants. Most teams that run this analysis end up cutting a third of their sourcing spend with no loss in hires.
• ✓
  Watch screen-to-interview conversion as your calibration gauge. This single ratio tells you whether the recruiter's screen and the hiring manager's bar agree. If a manager rejects 70 percent of the candidates a screen advances, the two are working from different definitions of the role — a criteria problem that consistent, structured screening (increasingly done with AI-assisted resume screening) surfaces within two weeks.

4. The Vanity Metrics to Stop Reporting

Some metrics are worse than useless: they actively push teams toward bad behavior. If a number cannot change a decision, it is decoration — and if it rewards the wrong decision, it is damage.

• ✓
  Raw applicant counts. “We received 1,200 applications” measures the reach of your job post, not the health of your hiring. Worse, celebrating it incentivizes vague, keyword-stuffed postings that attract volume over fit — which then buries your team in screening work that produces nothing.
• ✓
  Cost-per-hire as a target rather than a constraint. Driving cost-per-hire down pushes teams toward cheap, high-volume sourcing and away from the expensive channels that produce people who stay. A $6,000 hire who leaves in 5 months costs vastly more than a $10,000 hire who stays 3 years. Report cost-per-successful-hire, adjusted for 1-year retention, or do not report it at all.
• ✓
  Activity metrics: calls made, emails sent, resumes reviewed. These measure effort, not effect, and any team managed on them learns to manufacture activity. Fifty outreach emails with a 2 percent response rate is not twice as good as twenty-five with a 20 percent response rate — it is dramatically worse, and only outcome metrics will tell you that.

5. Build a Metrics Cadence That Changes Behavior

The metrics themselves are half the system. The other half is the rhythm in which they are reviewed — because a number nobody looks at until the quarterly business review is a number nobody can act on.

• ✓
  Run weekly pipeline reviews on leading indicators; save quality for quarterly retros. Stage velocity, pass-through rates, and aging candidates belong in a 30-minute weekly review where you can still change the outcome. Retention, hiring-manager satisfaction, and source quality move slowly — review them quarterly, when a full cohort of data makes the trend real.
• ✓
  Look at per-role views, not just aggregates. A company-wide 35-day time to hire can hide a 70-day engineering pipeline averaged against a 15-day support one. Aggregates are for the board deck; per-role and per-recruiter views are where the actual problems live. Segment before you conclude anything.
• ✓
  Assign one owner per bottleneck, with a date. “Scheduling is slow” is an observation; “Priya owns cutting stage-3 scheduling from 8 days to 4 by March 15” is a plan. Every metric that misses its threshold in the weekly review should leave the meeting attached to a name. Unowned metrics regress to decoration within a month.

Get These Metrics Without Building a Spreadsheet Empire

The reason most teams report vanity metrics is not that they prefer them — it is that the good metrics are painful to assemble by hand from ATS exports and calendar archaeology. As AI reshapes every stage of talent acquisition, analytics is the layer where the payoff compounds fastest.

Daisy Recruiter ships with recruitment analytics built in: stage-by-stage drop-off, source quality ranked by outcomes rather than volume, and time-to-hire tracking per role — out of the box, from the first application you process. No exports, no pivot tables, no waiting until the quarter is over to find out where your funnel was leaking.