Recruiting Funnel Metrics Analyst

What it does

It models the recruiting funnel (sourced → applied → screen → interview → offer → hire) as a flow system and turns numbers into judgement. For each consecutive stage pair it computes a yield ratio (passed/entered) and drop-off, expressing total funnel conversion as the product of stage rates (SHRM + LinkedIn Talent Solutions recruiting-funnel canon). On the speed side it separates Time-to-Fill (req open → offer accept) from Time-to-Hire (first contact → offer accept) — conflating them is the single most common interpretation error. Stage aging is measured by median to resist outlier reqs, and offer accept rate is benchmarked against the 80%/85% thresholds.

The primary bottleneck is chosen via Theory of Constraints (Goldratt): the stage with the highest (absolute candidate loss × longest wait) is the constraint. Fixing only that stage lowers total TTH; investing elsewhere is waste. For each bottleneck it forms a root-cause hypothesis (screen too strict? job ad attracting the wrong pool? interview scheduling dragging? offer not competitive?), ranks interventions with RICE (Reach × Impact × Confidence / Effort), and flags the speed-quality trade-off with Quality-of-Hire proxies (90-day turnover, accept rate). Where data allows, a Four-Fifths Rule equal-opportunity signal is checked. The result is summarised as a 0-100 weighted Funnel Health Score with a verdict.

When to use it

For questions like "why isn't this role filling", "where are candidates dropping", "why is interview-to-offer low", "how do I cut TTH", "which source is most efficient". For TA teams, HR Directors, recruiters and hiring managers who have an ATS/Excel/CSV export but want a numeric answer to "what single move?".

Method / frameworks

Yield Ratio funnel, Time-to-Fill vs Time-to-Hire (SHRM), Cost-per-Hire (ANSI/SHRM-09-2012), Quality-of-Hire proxies, Theory of Constraints, Four-Fifths Rule (EEOC), and RICE prioritization. Benchmarks are calibrated at runtime via web research per role/sector/country; missing evidence is written to gaps rather than invented.