How long should a CV be? The evidence is thinner than you think

Two numbers that cannot both be true

Here are the two most cited pieces of research in CV advice.

A recruiter spends 7.4 seconds looking at your CV.

Recruiters spend four minutes and five seconds on a two page CV, and two minutes twenty four on a one page CV.

Both are quoted constantly, often in the same article, by people who have not noticed that they differ by a factor of about thirty. They cannot both describe what a recruiter does at their desk. It is worth seeing where each came from, because the answer explains the whole genre.

The seven seconds

The six seconds figure, later updated to 7.4, comes from Ladders, a job listings company.

The 2012 study is titled Eye Tracking Online Metacognition: Cognitive Complexity and Recruiter Decision Making. It was run by Will Evans, then Head of User Experience Design at TheLadders, over ten weeks, with thirty recruiters. It was never peer reviewed and never published in a journal. No university was involved.

The 2018 update, the one that produced 7.4 seconds, was a press release. It discloses no sample size, no method, no researcher, and no duration. Trade press covering it at the time reported the same absence. I could not find the sample size for the 2018 figure anywhere, and I do not believe it has ever been published.

So the single most quoted number in CV advice is: n=30 in the only version with a disclosed sample, run in house by a job board's design lead, never peer reviewed, and the version everybody actually quotes has no published method at all.

The four minutes

The length research is not better, it is just differently weak.

ResumeGo published a study in January 2022. 482 recruiters, hiring managers and HR professionals reviewed 7,712 CVs. Two page CVs were selected 2.3 times more often than one page ones. Broken down: 1.4 times at entry level, 2.6 at mid level, 2.9 at managerial level. Two page CVs scored 8.6 out of 10 against 7.1.

That is a much larger study than the eye tracking one, and it is where the four minute figure comes from.

The caveats matter as much as the numbers. It is a simulation. Participants knew they were in a study and were evaluating CVs as a task, not filling a real vacancy. It measures stated preference in a controlled setting, not callbacks, not interviews, not hires. It is not peer reviewed. And it was run by a company that writes CVs for a living, whose product is longer and more thorough than what most people write themselves.

The thirty-fold gap between the two studies is not a mystery to solve. It is the difference between how somebody reads a CV when a researcher is watching and how they read one at 4pm with ninety more in the queue. Which tells you that neither number describes your actual reader.

What is actually missing

I went looking for a peer reviewed study measuring one page against two pages with real callback or interview outcomes.

There does not appear to be one.

Not a weak one, not an old one. On a question that every job seeker asks and every advice site answers with total confidence, the academic literature appears to be silent. Every number you have read on CV length traces back to a company that sells something to job seekers.

That is worth sitting with before you spend an evening cutting a good paragraph to fit an arbitrary page count.

What is genuinely well evidenced

One thing in this area has been measured properly, and it is not about length.

Reducing discrimination against job seekers with and without employment gaps, published in Nature Human Behaviour in 2022, is real experimental callback research. Not a survey about what recruiters think they do, not a simulation, but measurement of what actually happened to applications.

Employment gaps reduce callbacks. That is a finding with a method behind it.

Notice the contrast. The thing nobody can evidence is the thing everybody worries about, which is formatting and length. The thing that has been measured is a fact about your history that no page count will fix, and that is worth addressing directly on the page rather than hiding.

So how long should it be?

Since the evidence will not settle it, here is what the evidence does support, stated honestly as inference rather than finding.

• Length is not the variable. No study has ever shown a page count causing an outcome. The two page result that exists is about a document that had more relevant substance in it, not about a document that was longer.
• A second page has to earn itself. The ResumeGo gap widened with seniority: 1.4 times at entry level, 2.9 at managerial. That is consistent with the obvious reading, which is that senior people have more that is worth reading, not that senior people benefit from more paper.
• Write for a skimming reader, whichever number is right. Both studies, however weak, agree the first pass is fast. The top third of the first page carries the decision. That is a design instruction, not a length instruction.
• Do not cut substance to hit one page. There is no evidence supporting the one page rule for anybody beyond the very start of a career, and the only study that touched it points the other way.
• Address the gap rather than disguising it. It is the one thing here with proper research behind it.

The other half of this subject is the belief that a machine is filtering you before any of this matters. The evidence for that is thinner than people think, and it is worth reading before you rewrite anything.

The working version of this is in The CV That Gets Read.