Goal of the analysis:
Measure the share of hires who voluntarily self‑identify as members of underrepresented groups (as defined and permitted by local law), understand stage‑by‑stage representation through the funnel, and remove process barriers so all candidates have equitable access and experience. Diversity Hiring Rate is a monitoring metric—not a selection criterion—and must be managed in strict compliance with applicable laws. Executives use this analysis to broaden outreach, improve candidate experience, strengthen fairness and objectivity (structured interviews, validated assessments), and align sourcing investments to build stronger, more diverse teams without compromising quality, merit, or equal opportunity.
Data required:
- HRIS/ATS hires and pipeline:
- Requisition metadata: job family/level, location, business unit, internal vs external, backfill vs net new.
- Candidate stage timestamps: applied, screen, HM review, interview rounds, assessment, offer, accept/decline/withdraw/expire.
- Hires by cohort (start date or offer accept date).
- Voluntary self‑identification (handled compliantly):
- Self‑ID demographics permitted in jurisdiction (e.g., gender, race/ethnicity, disability status, veteran status); “prefer not to say” indicator.
- Self‑ID participation rate (% of applicants/hires who provided data).
- Source/channel and campaign data:
- Primary/assist source (career site, referral, job board, social, agency/RPO, campus, internal mobility, talent pool), campaign IDs.
- Decision artifacts and assessments:
- Structured screening outcomes, assessment pass/fail, interview ratings; rejection/withdrawal reasons.
- Market availability and benchmarks:
- External labor market availability by role/geo (e.g., public statistics) to create an “availability index.”
- Compliance and policy context:
- Local legal guidance on data use, storage, and reporting; adverse impact testing policies; equal opportunity and pay equity policies.
Detailed step-by-step instruction on how to conduct the analysis:
- Define scope, cohorts, and lawful categories.
- Choose cohort window (e.g., hires by start month for last 6–12 months). Segment by job family/level and location, given legal and market differences.
- Define which self‑ID categories can be reported in each jurisdiction; aggregate where sample sizes are small to protect privacy.
- Clarify that metrics are used for monitoring and process improvement, not decision‑making.
- Extract and prepare data.
- Pull ATS candidate stage histories and hire outcomes; join to voluntary self‑ID data and to source/channel. Ensure one terminal outcome per candidate per requisition.
- Compute self‑ID participation rates and tag “unknown/declined to state.”
- Normalize stage names/time zones; deduplicate candidates across reqs where necessary.
- Compute headline metrics.
- Diversity Hiring Rate (%) = Hires in defined group ÷ Total hires in cohort × 100 (per role family/geo).
- Diversity Pipeline Rate by stage = Candidates in defined group at stage ÷ Total at stage × 100.
- Adverse Impact Ratio (AIR) = Selection rate for group ÷ Selection rate for group with highest selection rate (flag if < 0.8 per common 4/5ths guideline; consult legal).
- Availability Index = (Group share of hires) ÷ (Group share in external labor market for the role/geo).
- Stage-by-stage pass-through and velocity.
- Compute conversion rates Applied→Screen, Screen→Interview, Interview→Offer, Offer→Accept by group; add time-in-stage medians to detect friction.
- Compare Offer Acceptance Rate and Offer speed by group to identify experience or competitiveness gaps.
- Segment to localize patterns.
- Slice by job family/level, location, remote vs on‑site, source channel, recruiter, and hiring manager (aggregate appropriately to protect privacy).
- Identify where top‑of‑funnel mix is thin vs where conversion gaps occur.
- Diagnose drivers and potential barriers.
- Top‑of‑funnel: JD clarity, outreach breadth, employer brand, channel mix. Check channel diversity yield (where permissible).
- Process: time‑in‑stage, missed SLAs, scheduling friction, interview bloat, assessment validity, panel composition (ensure structured and consistent).
- Offer: compensation competitiveness versus market midpoint, flexibility (remote/hybrid), benefits clarity, offer speed.
- Data quality: low self‑ID participation can distort rates; track and improve participation ethically.
- Compare to market availability and internal baselines.
- Contrast hiring shares with external availability to avoid unrealistic targets; use Availability Index to identify over/under‑performance relative to market.
- Track internal trends (rolling 12 months) to assess improvement independent of mix shifts.
- Scenario and what‑if (within legal bounds).
- Model impact of broadening outreach (channels/partnerships), inclusive JD rewrites, structured interviews, assessment calibration, and faster offers on funnel mix and acceptance.
- Estimate Time to Fill and Cost per Hire effects alongside Diversity Hiring Rate changes to balance objectives.
- Integrity, privacy, and fairness checks.
- Report only aggregated, minimum‑N compliant views; suppress cells with small counts. Use self‑ID data solely for analytics/compliance, not selection.
- Run adverse impact tests with legal/HR; validate that any assessment correlates with job performance and is role‑relevant.
Format of the output of analysis:
- Executive scorecard: Diversity Hiring Rate by role family/geo, self‑ID participation %, Availability Index, Adverse Impact Ratio by stage, trend vs prior period.
- Funnel charts: stage representation and pass‑through by group (aggregated), with time‑in‑stage overlays.
- Heatmaps: Diversity Hiring Rate and pipeline rates by role × location and by source channel (where permissible); self‑ID participation heatmap.
- Offer panel: Offer Acceptance Rate and offer speed by group (aggregated), comp vs market midpoint view.
- Scenario/ROI deck: expected changes in funnel mix, TTF, and CPH from inclusive process and outreach improvements.
How to interpret results:
- Low Diversity Hiring Rate with low top‑of‑funnel representation: Outreach and awareness challenge—broaden sourcing, partnerships, and refine JDs; check employer brand and apply experience.
- Healthy top‑of‑funnel but lower pass‑through for a group: Investigate process factors (speed, scheduling, assessment validity, interview structure); run adverse impact tests with legal.
- Offer declines higher for a group: Review offer speed, compensation alignment, and flexibility/benefits communication; ensure consistent candidate experience.
- Very low self‑ID participation: Data may be unreliable—improve communication on privacy and purpose; make self‑ID voluntary, easy, and clearly separated from selection.
- Availability Index < 1.0 with compliant process metrics: May reflect tight market availability; set targets relative to external benchmarks rather than absolute shares.
- AIR < 0.8 at a stage: Potential adverse impact—pause for review with legal/HR; validate job‑relatedness of screens/assessments and check for process friction.
Steps a company can take to improve on this measure:
- Inclusive outreach (lawful and broadening):
- Expand channels and partnerships (universities, professional associations, community groups); optimize programmatic ads to reach wider audiences.
- Refresh employer brand and career site with inclusive language, realistic job previews, and pay transparency where legal.
- Job design and postings:
- Simplify must‑have vs nice‑to‑have; remove non‑essential requirements; use inclusive JD language; ensure accessibility and mobile apply.
- Fair, predictive selection:
- Adopt structured interviews with anchored rubrics; validate assessments for job relevance; consider anonymized resume screens where feasible.
- Train interviewers on bias mitigation and consistent evaluation; limit interview bloat and tighten SLAs.
- Candidate experience and speed:
- Enable self‑scheduling; provide clear timelines and feedback; send offers within 24–48 hours post‑final interview for competitive roles.
- Compensation and flexibility:
- Align bands to market; offer equitable flexibility (remote/hybrid, schedules) where feasible; communicate benefits clearly and consistently.
- Data, privacy, and governance:
- Increase self‑ID participation through clear privacy messaging; secure storage and access controls; minimum‑N suppression; legal review of analytics and reporting.
- Monthly DEI–TA huddles to review funnel metrics, AIR results, and action progress; publish aggregated dashboards.
- Example scenarios:
- If top‑of‑funnel share is low for a role, add targeted partnerships and inclusive JD rewrites; expect Applied→Screen up and Diversity Hiring Rate to trend toward market availability.
- If Interview→Offer shows AIR < 0.8, audit interview rubrics and assessment validity; implement structured interviews and interviewer calibration; re‑test AIR after two cycles.
- If offer declines are concentrated for a group, improve offer speed and clarify benefits/flexibility; monitor Offer Acceptance Rate by cohort.
Benchmark comparisons:
General benchmarks (directional):
- Diversity Hiring Rate targets vary materially by role, geography, and external availability; many organizations set goals relative to market availability rather than absolute levels.
- Self‑ID participation: mature programs sustain ≥80–90% participation (voluntary) enabling reliable analytics.
- Adverse Impact Ratio: common practice flags AIR < 0.8 for deeper review (informational; confirm with legal and statistical tests).
Constructing internal benchmarks:
- Track rolling 12‑month Diversity Hiring Rate and stage pass‑through by role/geo; publish quartiles and guardrails (e.g., AIR alerts, self‑ID participation <70% triggers comms improvements).
- Set targets relative to external availability (Availability Index trending toward ≥1.0) while maintaining equal opportunity and job‑related selection standards.
- Pair diversity metrics with Time to Fill, Offer Acceptance, retention, and performance to ensure improvements are sustainable and merit‑based.
- Rebaseline after process/tool/brand changes and legal guidance updates; codify effective, lawful practices in a DEI–TA playbook.
