Hiring Funnel Conversion Metrics: The Numbers That Actually Tell You What's Broken
The five conversion rates that tell you everything
Track these per role family and per quarter:
- Application → screen call. What % of applicants are reaching a recruiter screen.
- Screen → technical assessment/interview. What % advance from recruiter to first technical stage.
- Technical → onsite loop. What % advance to the final loop.
- Onsite → offer. What % of onsite candidates get an offer.
- Offer → accept. What % of offers are accepted.
The product of these five is your overall funnel efficiency. The individual rates tell you where the leak is.
Rough benchmarks for engineering hiring
For senior engineering roles in 2026, healthy ranges look approximately like:
- App → screen: 5–15% (varies enormously with sourcing channel quality)
- Screen → technical: 50–70%
- Technical → onsite: 25–40%
- Onsite → offer: 40–60%
- Offer → accept: 75–90%
These are rough. Your numbers will vary. What matters is direction and consistency, not hitting an absolute target.
What each leak usually means
Low app → screen
You are either getting too many low-quality applicants (sourcing problem) or recruiters are over-rejecting at resume stage (calibration problem). Audit by re-screening 20 rejected applications blindly and see how many you'd actually call now.
Low screen → technical
Recruiter screens are filtering on the wrong signal, or the role is mis-marketed. Listen to a sample of screen calls. The filter criteria are usually different from the criteria the hiring manager actually cares about.
Low technical → onsite
Either the technical screen is too hard for the level (calibrate against your existing team — if 30% of your current engineers would fail it, the bar is too high), or the screen tests the wrong skill. Re-examine what the screen is filtering for vs. what the role requires.
Low onsite → offer
The most expensive leak. You have invested 4+ hours of interviewer time per onsite candidate before deciding "no." Either your top-of-funnel is letting through too many marginal candidates, or your loop's calibration is off and you'd actually offer if you re-debriefed cleanly. Often both.
Low offer → accept
Compensation gap, decision velocity, or candidate-experience problem. Run an exit survey on declined offers — three direct questions, no narrative. Compensation and timeline are usually the answer.
What "low" means
In every case: low relative to your historical baseline. A 50% offer→accept rate is fine if it's stable; it's catastrophic if it dropped from 85% last quarter. Track trends, not absolutes.
The dashboard
Most ATS tools surface these but bury them in reports nobody opens. Pull them weekly into a one-page dashboard. Color-code anything outside one standard deviation of the trailing 4-quarter average. Review with the hiring leadership team in 15 minutes.
ClarityHire's analytics surfaces stage conversion per role family and flags movements outside trend, so the leak shows up before the quarter-end review.
What this enables
You stop debating "why are we struggling to hire" abstractly and start saying "our screen-to-technical rate dropped from 65% to 40% — what changed?" Concrete questions, concrete answers, fixable problems.