All articles
6 min readBy CV Cleaner Pro Team

Anonymizing CVs for Blind Hiring: a Recruiter's Checklist

Blind hiring reduces bias only when anonymization is consistent. A 12-point checklist for recruitment agencies running structured anonymized CV pipelines.

Anonymizing CVs for Blind Hiring: a Recruiter's Checklist

Blind hiring works — when it's done consistently. The 2024 NBER review of 17 field experiments shows callback gaps narrow by 30–60% under structured anonymization. But the same studies note that inconsistent anonymization (some fields stripped, some not) can actually amplify bias because reviewers latch onto whatever signal remains.

If you're running blind hiring at a recruitment agency, you don't need a research paper — you need a checklist. Here's ours.

The 12-point anonymization checklist

#FieldAction
1Full nameReplace with Candidate #<hash>
2EmailStrip entirely
3PhoneStrip entirely
4Profile photoRemove all embedded images
5Street addressStrip to country only
6Date of birthStrip
7Nationality / citizenshipStrip
8Gender / pronounsReplace pronouns with "they/them"
9Marital statusStrip
10Graduation yearStrip (keep degree)
11Employer namesReplace with industry + size band
12LinkedIn / GitHub URLsStrip until shortlist stage

Rows 1–9 are table stakes. Rows 10–12 are the ones agencies most often miss — and they leak more bias signal than you'd think. Graduation year is a strong age proxy; employer prestige drives halo effects; LinkedIn URL leads reviewers straight back to a photo.

What to do instead of stripping

A field stripped to a blank doesn't make a CV less readable — but a field replaced with an abstracted signal keeps reviewers in flow. Examples:

  • Employer "Goldman Sachs" → "Tier-1 global investment bank, 40k+ employees"
  • Graduation year → "Education completed" / "In progress"
  • Location → "EU-based" / "US East Coast" / "Remote-only"

This is what we call structured anonymization: keep the category of information, strip the identifying token. Reviewers can still assess experience without being able to Google the candidate.

Where this breaks

Two failure modes we see repeatedly:

  1. Reviewers re-de-anonymize from context. "Worked on the Slack acquisition" → reviewer infers Salesforce, infers seniority band, infers compensation expectations. The fix is to also abstract project names above a threshold of public recognizability.
  2. Cover letters bypass the pipeline. Anonymization that only runs on the CV file leaves the cover letter as a back-channel. Process both, or don't accept cover letters at the shortlist stage.

GDPR and "right to be forgotten" notes

EU agencies should retain the anonymized version for the duration of the search and purge the original after a configurable window (we default to 90 days, see our GDPR page for details). Don't keep the original "just in case" — that defeats the consent basis you collected at upload.

CV Cleaner Pro applies the full 12-point checklist by default and lets you toggle each field per role. See it on the demo CV →

Clean every CV in your pipeline.

Detect prompt injection, strip hidden text, anonymize PII, reformat with your branding — in under 60 seconds per CV.

Start free