← The American Hiring System Solutions

Part of the Labor Market Infrastructure Initiative program

Quality-of-Hire Index

A shared, audited definition of hire quality — plus a consistency test any employer can run against their own AI screening vendor — replacing time-to-fill and cost-per-hire as the metric that actually gets managed.

Published August 1, 2026·Last revised August 1, 2026

Mission Alignment

Running the same AI screening tool twice on identical candidate data produces only 14% shortlist overlap — arguably the single most damning statistic in the whole research base behind this topic. It validates both the applicant's sense that rejection feels arbitrary and the employer's failure to get anything reliable for what they're paying for. Meanwhile, when firms dropped formal degree requirements from postings — a widely-praised, bipartisan-endorsed reform — the measured effect on actual hiring was about 0.14 percentage points, because most companies changed the posting without changing the evaluation mechanism underneath it. Both findings point to the same root cause: there's no shared, accountable metric for whether a hiring decision was actually good, so reforms optimize whatever gets measured instead — usually time-to-fill or cost-per-hire.

Problem Statement

Employers have no standardized way to know whether their screening process, human or AI, is actually finding better candidates — only whether it's fast and cheap. That makes "quality of hire" an aspiration mentioned in every hiring strategy deck and tracked, rigorously, in almost none.

Prior Research Findings

  • AI screening tools produce only 14% shortlist overlap when the same tool is run twice on identical candidate data — the screening layer both employers and candidates depend on is close to noise-dominated (recruiting industry data, 2026).
  • Dropping formal degree requirements from job postings moved the needle on actual hiring by about 0.14 percentage points — roughly 97,000 workers out of 77 million annual US hires — because only the 37% of companies that also changed their evaluation process saw a measurable effect (Burning Glass Institute + Harvard Business School, Feb 2024).
  • The cost of a bad hire runs 50–213% of annual salary depending on level (SHRM-derived estimates), which is measured and acted on constantly — while the cost of a false negative, a good candidate wrongly rejected, is structurally unmeasurable within the normal operation of a business. This solution exists specifically to close that measurement asymmetry.

Scope Boundary

In Scope

  • A published, shared quality-of-hire methodology (retention, performance-review outcomes, promotion rate at fixed intervals) employers can adopt in place of time-to-fill/cost-per-hire
  • A standardized same-data repeat-run consistency test any employer can commission against their own AI screening vendor
  • An anonymized, opt-in benchmark registry so quality-of-hire has an industry reference point the way cost-per-hire does today

Out of Scope

  • Building or selling a screening or AI hiring tool — this audits and measures existing tools, it doesn't replace them
  • Individual candidate scoring or ranking, which is exactly the kind of opaque, unaudited mechanism this solution is meant to hold accountable, not reproduce

Product Components

Quality-of-Hire Definition

  • A shared, published methodology combining retention, performance-review outcomes, and promotion rate at fixed post-hire intervals
  • Designed to be adoptable alongside existing HR systems rather than requiring a new platform migration

Screening Consistency Audit

  • A standardized same-data, repeat-run test any employer can commission against their own AI screening vendor
  • Produces a public or private overlap-consistency score directly comparable to the 14% baseline already documented industry-wide

Benchmark Registry

  • Anonymized, opt-in employer benchmark data, giving quality-of-hire the kind of shared reference point cost-per-hire and time-to-fill already have

Phased Milestones

  1. Publish v1 methodology for public comment
  2. Pilot the consistency audit with 2–3 employers across different AI screening vendors
  3. Publish the first anonymized benchmark report
  4. First screening vendor voluntarily publishes its own consistency score

Open Research Questions

  • What's the right minimum measurement window (90 days? a full year?) for quality-of-hire outcomes to be meaningful without being too slow to actually inform hiring decisions?
  • Will screening vendors participate in an audit that could produce a public score worse than the already-documented 14% baseline?

Success Metrics

  • Number of employers adopting quality-of-hire as a tracked KPI alongside or in place of time-to-fill and cost-per-hire
  • At least one independently-audited screening-tool consistency score published and publicly comparable to the 14% baseline

Revision History

DateChanges
August 1, 2026First published

Discussion