What Research Actually Says About Résumés (And the Myths It Kills)

Almost everything circulating about résumés is folklore. Advice repeated until it became obvious, with nobody tracing it back to a source — because there often isn’t one. We did the opposite exercise: start from named, datable studies and look at what they actually establish. Some confirm the intuition. Others contradict it head-on.

Myth 1: “75% of résumés are rejected by a robot”

It is the most quoted statistic in the industry, and it has no source. No published study produces it. Its earliest trace is a vendor sales pitch from the early 2010s, passed from blog to blog until it became received wisdom.

More recent data points the other way: major applicant tracking systems do not auto-reject résumés. They filter, sort and prioritise — which is a different thing. A minority of employers configure automatic rejection, and almost always on hard, checkable criteria: a mandatory degree, a work permit, a legally required certification.

The real reason an application never gets read is far more mundane: volume. An attractive posting collects hundreds of applications within days, and nobody reads them all with equal attention.

What actually filters: the recruiter’s own search

Filtering is real, but it does not come from an autonomous robot. It comes from the recruiter typing keywords into their own tool. This is the most solid finding in the whole file, and it comes from genuine academic research.

Hidden Workers: Untapped Talent, led by Joseph Fuller and Manjari Raman at Harvard Business School with Accenture in 2021, surveyed 8,000 workers and 2,250 executives across the United States, United Kingdom and Germany. It establishes that 88% of employers agree qualified candidates get screened out because they do not match the exact criteria in a job description — rising to 94% for middle-skills roles.

“

Keywords do not matter because a robot eliminates you. They matter because a human searches with them, and what is not phrased in their terms does not appear in their results.

The practical consequence is precise: use the posting’s exact terminology wherever you genuinely did the work. Not a personal synonym, not an elegant paraphrase — the posting’s term. And once, well placed, beats five times repeated, because the person who opens the document will actually read it.

Myth 2: “one page, always”

The one-page rule may be the most universally repeated advice there is. The data contradicts it.

ResumeGo put 482 recruiting professionals — recruiters, hiring managers, HR, executives — through a screening simulation. Across the 7,712 résumés they chose, two-page versions won 2.3 times more often than one-page versions. The result holds at the bottom of the ladder too: for entry-level roles the gap is still 1.4 times in favour of two pages. Recruiters also spent markedly longer on them — around 4 minutes versus 2 and a half.

This result has to be read correctly, and this is where many go wrong. The protocol compared the same candidate in a short and a long version. The two-page version won because it carried more substance, not because it was longer. So the conclusion is not “make your résumé longer” — it is “do not amputate relevant content to fit a format”.

In Québec, university career services say the same. Polytechnique Montréal’s guide states that two-page CVs are the most common on the Québec market, and explicitly invites candidates to detail their experience rather than condense it.

The 7.4 seconds that decide

In 2018, TheLadders repeated an eye-tracking study it had first run in 2012. The first scan of a résumé is measured there at roughly 7.4 seconds, up from about 6 seconds six years earlier.

What matters more than the duration is the path. The eye follows an F-pattern: job title, employer, then a move right to check dates and progression, then a jump down to education. The résumés that survived that scan shared the same traits: simple layout, clearly titled sections, bold titles, bulleted accomplishments.

That also settles the bold debate. An ATS reads plain text: it never sees a font weight, and bolding a keyword changes nothing whatsoever in a matching score. Bold does not serve the machine — it serves the eye, and the eye decides in seven seconds.

Bullets: quantified, short, and all one form

On writing bullets, university career services and the skim-reading literature agree: one idea per bullet, a measurable result wherever one exists, two lines maximum. An overflowing bullet is not read more slowly — it is skipped.

One point often goes unnoticed and yet shows instantly: verb-form consistency. Polytechnique Montréal’s guide asks candidates to open every statement with an action verb or a noun, and above all to stay consistent. Its model CV is entirely in the infinitive.

The convention depends on the market, a nuance most tools ignore. In Québec, the infinitive is the taught form for describing tasks, with the past participle used to present achievements — the key being never to mix the two in one document. In North American English, the norm is the past tense: “Built”, “Led”, “Automated”. A résumé that alternates forms is spotted at a glance.

The summary: bullets in Québec, prose elsewhere

Here the conventions genuinely diverge, and following the wrong one is not neutral.

Polytechnique Montréal describes the summary as a list of 4 to 6 bullets, and insists on one point: it is not a general summary of your skills. Each bullet must be a specific asset, tied to a requirement of the posting or to what you know about the company.

The North American English consensus differs: two to three lines of prose establishing who you are, at what level and in what domain, before the experience bullets prove the rest. Both work — each in its own market.

What a résumé must never carry

Polytechnique Montréal’s guide is categorical about a list of details that should never appear, because they expose you to discrimination and say nothing about your ability to do the job:

• Your photo
• Your date and place of birth
• Your marital status
• Any reference to religion or politics
• Your citizenship or residency status — unless the role explicitly requires it

Two practical details from the same guide, often forgotten outside North America: the expected paper size is Letter, not A4; and local practice is to send the CV and the cover letter as a single PDF file.

What to take away

1. 1 No robot eliminates you automatically — but a recruiter filters with their own keywords, so use the posting’s wherever you genuinely did the work.
2. 2 Do not amputate relevant content to fit one page. One page if your background fits one page, two if you can fill them honestly.
3. 3 Write for a seven-second scan: clear hierarchy, bold titles, short quantified bullets.
4. 4 Hold a single verb form from the first bullet to the last — the one your target market expects.
5. 5 Match the summary to the market: four to six bullets in Québec, two to three lines of prose in North American English.
6. 6 Remove anything that exposes you to discrimination without proving a thing about your skills.

None of this takes special talent. It only takes writing for the person who will read — which most résumés, still today, do not do.