Recruitment advertising for frontline employers

We fill the top of the funnel for frontline employers. This page is about what happens to those applicants once they reach your shortlist.

Book a demo

Hiring guideWritten for frontline employers · 16 September 2026

Blind Hiring

Blind hiring means removing information that identifies a candidate — name, photograph, address, school, graduation year, sometimes employer — from applications before anyone screens them. The case for it is intuitive and the evidence for it is weaker and stranger than most articles admit: the most famous supporting study was about orchestra auditions rather than résumés and has been questioned on its statistics, while the largest field trial — more than 2,100 public servants shortlisting 16 candidates each — found de-identification made reviewers less likely to shortlist female and minority candidates, because they had been favoring them when names were visible. For frontline hiring specifically there is a further problem: the decision that most affects who gets hired happens long before the shortlist.

What blind hiring actually removes

In practice the term covers several different interventions of very different strength, which is part of why the evidence is muddled. From lightest to heaviest:

Name and photograph removed — the most common version, and the one most software does by default.
Address and contact details removed — which also removes the postcode, a strong proxy for a lot of things.
Education details removed — school name and graduation year, the latter being a close proxy for age.
Employer names removed — rarer, because it strips a great deal of the useful signal along with the identifying one.
Structured application replacing the résumé entirely — everyone answers the same questions in the same fields, which is a different and arguably larger intervention.
Work-sample or skills assessment first, with the application read only afterwards — the heaviest version, and the one with the clearest logic behind it.

When someone says blind hiring "works" or "does not work", the first question is which of these six they mean. The evidence below is not the same for all of them.

The evidence, honestly

The study everyone cites

The famous result is Goldin and Rouse's study of symphony orchestras adopting screened auditions, published in 2000, which reported that a screen substantially improved women's chances of advancing. It is a genuinely elegant natural experiment and it is the origin of almost every "blind hiring increases diversity" claim you will read.

Two things are worth knowing before you lean on it. First, it studied live auditions — a performance evaluated in real time — not résumé screening, and the two are not obviously the same task. Second, the statistical strength of the result has been questioned in subsequent re-examination, with critics arguing the estimates are fragile relative to how confidently they are repeated. That does not make the finding wrong. It does make it a thinner foundation than the volume of citation suggests.

The largest field trial found the opposite

The Behavioral Economics Team of the Australian Government ran a trial with over 2,100 public servants across 14 agencies, each shortlisting 5 of 16 hypothetical candidates for an executive-level role. Some reviewers saw names; some saw the same CVs with names swapped by gender; some saw CVs identified only by number.

The finding was that reviewers were already practising positive discrimination, and removing names took it away. With identities visible, reviewers were 2.9% more likely to shortlist female candidates and 3.2% less likely to shortlist male ones; minority males were 5.8% more likely to be shortlisted and minority females 8.6%; Indigenous female candidates were 22.2% more likely to be shortlisted. The authors concluded that de-identification would undermine efforts to promote diversity in that setting.

What that does and does not mean

It was a framed experiment with hypothetical candidates in a senior public-sector context where diversity targets were explicit and salient. It is not evidence that blind hiring harms diversity in a private frontline business with no such targets. It is strong evidence that the effect depends on what reviewers are already doing — and that you cannot assume the direction in advance.

What is not in dispute

Correspondence studies — sending matched fictitious applications that differ only by name — have repeatedly found callback differences by the name on the application. That literature is large and has been replicated across countries and decades. It establishes that names affect callbacks. It does not establish that removing names is the best available fix, which is a separate question.

The frontline problem: the shortlist is the wrong place to look

Most writing about blind hiring assumes a funnel that ends with a stack of résumés being read. Frontline hiring frequently does not work that way. The application is often four questions on a phone, there may be no résumé at all, and the person deciding is a store or plant manager who will speak to the candidate within the hour.

Which means the decisions that actually shape who gets hired sit earlier and later than the shortlist:

Who saw the job. Where you advertised, which audience the targeting reached, and who was excluded before anyone applied. This is the largest single determinant of the composition of your applicant pool and it is almost never discussed alongside blind hiring.
What the posting said. Requirements that are not genuine requirements — a degree for a job that does not need one, “3+ years” where six months would do, open availability, English fluency where the work is bilingual — remove whole populations before any screening happens.
How the application worked. A ten-field form on a phone excludes people differently from a four-question one. So does requiring an email address, a résumé upload, or a desktop browser.
The interview itself. Unstructured conversation is where most of the discretion in frontline hiring lives, and anonymising a résumé does nothing to it.

If you are going to spend effort on one of these, the posting and the interview structure are where the leverage is for frontline roles — not because blind screening is bad, but because for most frontline employers there is no screening stage substantial enough for it to change.

If you want to try it anyway, here is a defensible version

There is a reasonable case for trying it: it is cheap, it is reversible, and the correspondence-study literature gives it a plausible mechanism. If you do, do it in a way that produces evidence rather than a policy.

Pick the version that matches your process. If you have no résumé stage, anonymising résumés is theatre. A structured application where everyone answers the same questions is the version that fits most frontline hiring.
Anonymise what is a proxy, not just the name. Postcode, school and graduation year carry more than most people expect. Removing the name alone is the weakest intervention available.
Decide the criteria before you look. The mechanism only works if there is a standard to apply; anonymised résumés read against no criteria are just anonymous impressions.
Keep it to the screening stage. Nobody is proposing anonymous interviews, and the interview is where most discretion sits — so structure that too, or the gain is given straight back.
Measure it. Track the composition of applicants, shortlists, interviews and hires before and after. If you cannot see a change, you have learned something worth knowing.
Expect it to be uncomfortable. The APS trial found reviewers were favoring under-represented candidates and the blinding removed that. You may find the same, and you should decide in advance what you would do if you did.

Run it as a trial with a measurement plan and a date to review, not as a policy announcement. A practice adopted on principle cannot be abandoned on evidence.

What we can and cannot claim

We have no measured evidence of our own on blind hiring. Our data is top-of-funnel: what it costs to reach frontline candidates on social and what share of them apply. It has nothing to say about what happens to an application after it lands, and we are not going to borrow a percentage from a vendor survey to fill the gap.

What our data does support is the point about the top of the funnel. Across 891 managed campaigns, the single largest driver of campaign outcomes is the applicant conversion rate — what happens after the click — which explains around 70% of the variation in cost per applicant, against 30% for the ad itself. In other words, the posting and the application form do most of the work in determining who ends up in your pool. That is a finding about cost, not about fairness, but it points at the same place: the decisions that shape a candidate pool mostly happen before anybody screens anything.

The adjacent things worth doing regardless

None of these depend on the blind-hiring evidence being settled, and all of them are within a frontline employer's control.

Cut requirements that are not requirements. Every unnecessary line removes people, and the removal is not random. Our job description library goes through this line by line for 80 roles.
Structure the interview. Same questions, same order, criteria written down beforehand, scored independently afterwards. The interview scorecard template builds one.
Know which questions are unlawful and keep them out of the relaxed part of the conversation, which is where they surface. See illegal interview questions.
Advertise where the people are. A pool drawn from one referral network will look like that network. Paid social reaches people who are not searching, which is a different population from a job board.
Shorten the application. Four questions, mobile-first, no account required. It is the cheapest change on this list and the one with the most measurable effect.

If blind hiring turns out to help in your setting, it will help on top of these. If it turns out to do nothing, these still work.

Frequently asked questions

What is blind hiring?

Blind hiring is the practice of removing identifying information from applications before anyone screens them — typically the candidate's name and photograph, and sometimes their address, school, graduation year or employers. The aim is to have reviewers judge the content of an application rather than its author. Versions range from removing a name to replacing the résumé entirely with a structured application or a work-sample test.

Does blind hiring actually work?

The honest answer is that it depends on the setting and the evidence is mixed. Correspondence studies consistently show that names affect callback rates, which gives the idea a plausible mechanism. But the most cited supporting study concerned orchestra auditions rather than résumés and has been questioned statistically, and the largest field trial — over 2,100 Australian public servants — found de-identification made reviewers less likely to shortlist female and minority candidates, because they had been favoring them when names were visible.

What did the Australian Public Service blind recruitment trial find?

The Behavioral Economics Team of the Australian Government asked more than 2,100 public servants across 14 agencies to shortlist 5 of 16 hypothetical candidates, with some seeing names and some seeing de-identified CVs. Reviewers were 2.9% more likely to shortlist female candidates and 3.2% less likely to shortlist male ones when identities were visible; minority males were 5.8% more likely to be shortlisted, minority females 8.6%, and Indigenous females 22.2%. The authors concluded de-identification would undermine diversity efforts in that context.

What information should be removed in blind hiring?

More than the name, if you are doing it seriously. Photograph, address and postcode, school name and graduation year all act as proxies, and graduation year in particular is a close proxy for age. Employer names are sometimes removed too, though that strips useful signal along with identifying signal. Removing only the name is the weakest version of the practice.

Is blind hiring useful for frontline and hourly roles?

Often less than for professional hiring, because many frontline processes have no substantial résumé screening stage to blind. The application may be four questions on a phone and the manager may speak to the candidate the same hour. In that shape, the decisions that most affect who gets hired are where you advertised, what the posting required, how long the application was, and how the interview was run.

What is the difference between blind hiring and structured interviewing?

Blind hiring removes identifying information before screening. Structured interviewing standardizes the questions, the order and the scoring so candidates are compared on the same basis. They address different stages, and for frontline employers the interview is usually where more of the discretion sits — so blinding a résumé stage while leaving the interview unstructured gives the gain straight back.

Is blind hiring legal?

Removing identifying details from applications is a process choice rather than a regulated practice in most jurisdictions, and employers do it routinely. That said, hiring processes are subject to anti-discrimination and record-keeping law that varies by jurisdiction, and some jurisdictions have specific rules about what you may ask and when. This page is not legal advice — have counsel review any change to your process.

How do you measure whether blind hiring made a difference?

Track the composition of applicants, shortlists, interviews and hires for a period before the change and the same period after, and set the review date in advance. Run it as a trial with a measurement plan rather than announcing it as a policy, because a practice adopted on principle is very hard to abandon on evidence — and the Australian trial is a reminder that the effect can run in the direction you did not expect.

What should we do instead if we only have time for one change?

For most frontline employers, cut the requirements that are not genuine requirements from the job posting. Every unnecessary line — a degree the work does not need, a years-of-experience figure, open availability, English fluency where the work is bilingual — removes people before any screening happens, and the removal is not random. It is also free.

Who applies is decided before anyone screens anything.

We put frontline jobs in front of people who are not job hunting, and report cost per applicant. Median across 891 managed campaigns: $13.88.

Book a Demo

Sources read for this page: Goldin and Rouse (2000) on screened orchestra auditions, and the subsequent re-examination of its statistical strength; the Behavioral Economics Team of the Australian Government report on unconscious bias in Australian Public Service shortlisting (over 2,100 participants, 14 agencies); and the correspondence-study literature on name effects in callbacks. Campaign figures: Boostpoint 2026 Social Job Advertising Benchmark, 891 managed campaigns and 1,334 campaign-months. No effect-size claim is made for blind hiring; where we have no measured evidence, this page says so. Not legal advice.