{"id":156979,"date":"2022-01-04T14:43:19","date_gmt":"2022-01-04T11:43:19","guid":{"rendered":"https:\/\/ceotudent.com\/from-2021s-perspective-5-secrets-of-effective-cv-preparation"},"modified":"2026-10-14T05:20:00","modified_gmt":"2026-10-14T02:20:00","slug":"from-2021s-perspective-5-secrets-of-effective-cv-preparation","status":"publish","type":"post","link":"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation","title":{"rendered":"5 Secrets of Effective CV Preparation: Which Ones Does the Research Support? (2026)"},"content":{"rendered":"<p><strong>TL;DR.<\/strong> The five secrets from the 2021 version and the famous &ldquo;7 seconds&rdquo; claim were tested against primary sources: only 1 of the 8 claims, that a resume (CV) should be free of spelling errors, is verified; 5 are partly supported and no source was found for 2. The &ldquo;7 seconds&rdquo; figure comes from a career site&rsquo;s eye-tracking study: the 2012 report has 30 recruiters and 6 seconds, the 2018 press release has 7.4 seconds; both measure the initial screening glance and neither is peer-reviewed. By contrast, in an experiment with 445 recruiters in Belgium, five spelling errors lowered the probability of being invited to an interview by 18.5 percentage points and two errors by 7.3 points (Sterkens et al., 2023). In a field experiment with 480,948 job seekers, those who received writing assistance were hired 8% more often (Wiles, Munyikwa and Horton, 2023). In a 9,022-application experiment in the United Kingdom, stating the number of years worked instead of employment dates increased callbacks by about 8% (Kristal et al., 2023). <strong>The CEO move:<\/strong> manage the CV as a decision document presented to a recipient who is short on time; one target role, evidence in the first third, a test before submission. <strong>The student move:<\/strong> treat every application round as data, run small single-variable trials and record the result.<\/p>\n<p>The stage where a machine is the first reader is covered in <a href=\"https:\/\/ceotudent.com\/en\/resume-writing-when-ai-reads-first-what-screening-systems-parse-2026\">the article on how to write a resume when AI reads it first<\/a> and is not repeated here. This piece is complemented by the articles on <a href=\"https:\/\/ceotudent.com\/en\/the-skill-audit-map-what-you-know\">the skill audit<\/a> and <a href=\"https:\/\/ceotudent.com\/en\/certificates-portfolios-degrees-what-signals-skill-2026\">the comparison of certificates, portfolios and degrees<\/a> for where the evidence that goes into a CV comes from, <a href=\"https:\/\/ceotudent.com\/en\/ai-fluency-interview-how-companies-test-ai-skills-2026\">the AI fluency interview<\/a> for the step after the CV, and <a href=\"https:\/\/ceotudent.com\/en\/professional-reputation-ai-era-build-authority-same-tools\">professional reputation<\/a> for the signals outside the document.<\/p>\n<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_84 counter-hierarchy ez-toc-counter ez-toc-grey ez-toc-container-direction\">\n<div class=\"ez-toc-title-container\">\n<p class=\"ez-toc-title\" style=\"cursor:inherit\">Table of Contents<\/p>\n<span class=\"ez-toc-title-toggle\"><a href=\"#\" class=\"ez-toc-pull-right ez-toc-btn ez-toc-btn-xs ez-toc-btn-default ez-toc-toggle\" aria-label=\"Toggle Table of Content\"><span class=\"ez-toc-js-icon-con\"><span class=\"\"><span class=\"eztoc-hide\" style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #999;color:#999\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewBox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #999;color:#999\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewBox=\"0 0 24 24\" version=\"1.2\" baseProfile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/span><\/a><\/span><\/div>\n<nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#How-do-the-five-secrets-of-2021-hold-up-against-the-research\" >How do the five secrets of 2021 hold up against the research?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#Does-a-recruiter-really-look-at-a-CV-for-7-seconds\" >Does a recruiter really look at a CV for 7 seconds?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#What-do-experiments-show-changes-callbacks-on-a-CV\" >What do experiments show changes callbacks on a CV?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#One-page-a-summary-paragraph-and-reverse-chronological-order-how-strong-is-the-evidence\" >One page, a summary paragraph and reverse chronological order: how strong is the evidence?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#Do-tailoring-to-the-posting-and-the-cover-letter-still-work-in-the-age-of-AI\" >Do tailoring to the posting and the cover letter still work in the age of AI?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#Who-reads-the-CV-first-in-2026\" >Who reads the CV first in 2026?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-7\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#Lead-like-a-CEO-how-is-a-CV-built-as-a-decision-document\" >Lead like a CEO: how is a CV built as a decision document?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-8\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#Learn-like-a-student-how-is-an-application-experiment-logged\" >Learn like a student: how is an application experiment logged?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-9\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#What-do-these-findings-not-mean\" >What do these findings not mean?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-10\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#Why-was-this-article-rewritten\" >Why was this article rewritten?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-11\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#Frequently-asked-questions\" >Frequently asked questions<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-12\" href=\"https:\/\/ceotudent.com\/en\/from-2021s-perspective-5-secrets-of-effective-cv-preparation\/#Sources\" >Sources<\/a><\/li><\/ul><\/nav><\/div>\n<h2 id=\"how-do-the-five-secrets-of-2021-hold-up-against-the-research\"><span class=\"ez-toc-section\" id=\"How-do-the-five-secrets-of-2021-hold-up-against-the-research\"><\/span>How do the five secrets of 2021 hold up against the research?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The old article listed five recommendations: opening with a summary that states the career objective; writing a short, one-page, error-free document; using a tidy design and reverse chronological order; tailoring the CV to the job posting; and adding a cover letter. The introduction also said that &ldquo;research&rdquo; shows recruiters look at a resume for only seven seconds. None of these had a source.<\/p>\n<p>The most comprehensive survey opened for this update is Risavy&rsquo;s 2017 review of resume research. The review&rsquo;s own finding: much of this literature is old. The studies it reports under length date from 1976-2001, and most of those under the objective statement date from 1984-1997. The table below combines this review, later experiments and the AI studies published after 2023.<\/p>\n<p><strong>CEOtudent analysis: evidence scorecard for the claims in the 2021 version<\/strong><\/p>\n<table>\n<thead>\n<tr>\n<th>Claim in the 2021 version<\/th>\n<th>What does the opened source say?<\/th>\n<th>Verdict<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Introduction: according to research, a recruiter looks at a resume for only 7 seconds<\/td>\n<td>The source is commercial. Ladders 2012 report: 30 recruiters, 6 seconds. 2018 press release: 7.4 seconds, sample not stated. What is measured is the initial screening glance.<\/td>\n<td>Partly<\/td>\n<\/tr>\n<tr>\n<td>1. Open with a summary that states the career objective<\/td>\n<td>Older studies find the objective statement important; the review says the effect needs to be demonstrated more conclusively (Risavy, 2017).<\/td>\n<td>Partly<\/td>\n<\/tr>\n<tr>\n<td>2a. A CV should usually be one page; it can run longer after 10 years of experience<\/td>\n<td>The consistent finding in the review is not to exceed two pages. In the only experiment opened, two pages were ranked more favorably (Blackburn-Brockman and Belanger, 2001). No source for the 10-year threshold.<\/td>\n<td>Not found<\/td>\n<\/tr>\n<tr>\n<td>2b. Zero spelling errors is among the most important secrets<\/td>\n<td>Five errors cost 18.5 points, two errors 7.3 points (Sterkens et al., 2023). Errors do as much damage as a lack of experience (Martin-Lacroux and Lacroux, 2017).<\/td>\n<td>Verified<\/td>\n<\/tr>\n<tr>\n<td>3. Tidy design, reverse chronological order<\/td>\n<td>A formal layout is preferred to a &ldquo;creative&rdquo; design (Arnulf et al., 2010). The preference for the chronological format rests on older studies. Writing the number of years instead of dates increased callbacks (Kristal et al., 2023).<\/td>\n<td>Partly<\/td>\n<\/tr>\n<tr>\n<td>4a. Tailor the CV to the posting<\/td>\n<td>Older studies recommend it; no recent experiment on CVs could be opened. On a freelancing platform, the link between tailoring and callbacks weakened by 51% after the AI tool (Cui et al., 2025).<\/td>\n<td>Partly<\/td>\n<\/tr>\n<tr>\n<td>4b. Add the keywords, stand out in the ATS<\/td>\n<td>No experiment measuring this could be opened. With human evaluators, a list of technical skills has no effect (Kessler et al., 2019). Keyword stuffing is a feature of the worst resumes (Ladders, 2018, commercial).<\/td>\n<td>Not found<\/td>\n<\/tr>\n<tr>\n<td>5. A cover letter is an expected formality for most office jobs<\/td>\n<td>The older literature supports the expectation (Risavy, 2017). After AI, the cover letter&rsquo;s power to differentiate declines (Cui et al., 2025).<\/td>\n<td>Partly<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><em>CEOtudent analysis. The Verdict column is an editorial assessment. &ldquo;Not found&rdquo; does not mean that no source exists, only that none turned up in the texts opened for this update.<\/em><\/p>\n<p>The count: 1 verified, 5 partly, 2 not found. Only one secret can be stated with the certainty of the old article.<\/p>\n<h2 id=\"does-a-recruiter-really-look-at-a-cv-for-7-seconds\"><span class=\"ez-toc-section\" id=\"Does-a-recruiter-really-look-at-a-CV-for-7-seconds\"><\/span>Does a recruiter really look at a CV for 7 seconds?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The claim traces back to a single company: Ladders, which describes itself as a career site for jobs paying $100,000 and above. According to the 2012 report, which could be opened, the eye movements of 30 professional recruiters were tracked over 10 weeks. The report says recruiters spent about 6 seconds on their initial &ldquo;fit \/ no fit&rdquo; decision, whereas research based on recruiters&rsquo; own self-reports had suggested 4-5 minutes per resume. About 80% of the review time went to six data points: name, current title and company, previous title and company, the start and end dates of the previous and current positions, and education. Beyond these, recruiters did little more than scan for keywords matching the job posting.<\/p>\n<p>The report for the 2018 update could not be opened because of a bot wall on the company&rsquo;s site; the document that was opened is the company&rsquo;s press release dated November 6, 2018. According to the release, the average initial screening time is 7.4 seconds; it does not say how many recruiters were tracked. The &ldquo;seven seconds&rdquo; in the old article is most likely a rounded version of this number.<\/p>\n<p>There are three caveats. First, the source is commercial: the 2012 report advises readers to have their resume rewritten by a professional and ends by promoting the company&rsquo;s own service of that kind. Second, what is measured is the initial screening glance, not the reading of the document. Third, a peer-reviewed article gives a different order of magnitude: Sterkens and colleagues (2023), citing Arnulf and colleagues, report the time employers spend screening a resume as &ldquo;about 45 seconds&rdquo; (only the abstract of the cited study could be opened). The two numbers differ by a factor of roughly 6 (CEOtudent calculation).<\/p>\n<p>What can be said with confidence: the first glance is short and concentrated on a few fields. &ldquo;7 seconds&rdquo;, however, is one company&rsquo;s small-sample measurement.<\/p>\n<h2 id=\"what-do-experiments-show-changes-callbacks-on-a-cv\"><span class=\"ez-toc-section\" id=\"What-do-experiments-show-changes-callbacks-on-a-CV\"><\/span>What do experiments show changes callbacks on a CV?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The studies below do not ask about preferences; they change one feature of the resume and measure the outcome.<\/p>\n<p><strong>Effects measured in resume experiments<\/strong><\/p>\n<table>\n<thead>\n<tr>\n<th>Study<\/th>\n<th>Design<\/th>\n<th>Feature changed<\/th>\n<th>Measured outcome<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Sterkens et al., 2023<\/td>\n<td>Belgium; 445 recruiters, 1,335 evaluations; scenario experiment<\/td>\n<td>0, 2 or 5 spelling errors<\/td>\n<td>5 errors: interview probability 18.5 percentage points lower. 2 errors: 7.3 points lower<\/td>\n<\/tr>\n<tr>\n<td>Wiles, Munyikwa and Horton, 2023<\/td>\n<td>Online labor platform; 480,948 job seekers; field experiment<\/td>\n<td>Writing assistance that flags errors<\/td>\n<td>Probability of being hired in the first month 8% higher; hourly wage 10% higher among those hired<\/td>\n<\/tr>\n<tr>\n<td>Kristal et al., 2023<\/td>\n<td>United Kingdom; 9,022 applications, 8 sectors; field experiment<\/td>\n<td>Number of years worked instead of employment dates<\/td>\n<td>Callbacks about 8% higher than for a resume without a gap and 15% higher than for a resume with a gap<\/td>\n<\/tr>\n<tr>\n<td>Kroft, Lange and Notowidigdo, 2012<\/td>\n<td>United States; about 12,000 resumes, 3,000 postings; field experiment<\/td>\n<td>Duration of current unemployment<\/td>\n<td>Callback rate falls from roughly 7% to 4% by the eighth month<\/td>\n<\/tr>\n<tr>\n<td>Bertrand and Mullainathan, 2003<\/td>\n<td>United States; about 5,000 resumes, more than 1,300 postings; field experiment<\/td>\n<td>Racial connotation of the name; resume quality<\/td>\n<td>50% more callbacks for white-sounding names. A higher-quality resume brings 30% more callbacks for these names<\/td>\n<\/tr>\n<tr>\n<td>Kline, Rose and Walters, 2022<\/td>\n<td>United States; more than 83,000 applications, 108 large employers; field experiment<\/td>\n<td>Racial connotation of the name<\/td>\n<td>24% of applications received a callback within 30 days. The rate is 2.1 percentage points lower for Black-sounding names<\/td>\n<\/tr>\n<tr>\n<td>Kessler, Low and Sullivan, 2019<\/td>\n<td>United States; 72 employers, 2,880 resume ratings; incentivized rating<\/td>\n<td>Internship quality; list of technical skills<\/td>\n<td>A prestigious internship adds 0.9-1.0 points on a 10-point scale. The skills list has no effect<\/td>\n<\/tr>\n<tr>\n<td>Cui, Dias and Ye, 2025 (preprint)<\/td>\n<td>Freelancing platform; AI cover letter tool<\/td>\n<td>Use of the tool<\/td>\n<td>Callbacks 3.56 percentage points higher (previous mean 7.02%); significant at the 10% level, not at the 5% level<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><em>Source: full texts of the studies named; values are as printed. The Wiles, Kroft, Bertrand, Kline and Kessler rows are taken from the NBER working papers. 2,880 is a CEOtudent total (1,840 + 1,040).<\/em><\/p>\n<p>Three conclusions follow from the table. First, the most clearly measured effects come from errors and content, not from decoration: spelling, the presentation of dates, the quality of experience. Second, a list is not evidence: in Kessler and colleagues, technical skills randomly added to a resume did not change employer interest; according to the authors, employers may treat a skills list as a costless claim. Third, the same resume gets different callbacks simply because the name changes; silence does not always say something about the quality of the document.<\/p>\n<p>For spelling errors, the loss per error is similar at both levels: 3.70 points with five errors and 3.65 points with two (CEOtudent calculation). According to Sterkens and colleagues, the effect of five errors is unlike that of any of the other candidate characteristics they examined. About half of the penalty is explained by perceptions of the candidate&rsquo;s mental ability (32.2%), conscientiousness (12.1%) and interpersonal skills (9.0%). The abstract of a study with 536 recruiters reports that spelling errors harm the chance of being shortlisted as much as a lack of experience does (Martin-Lacroux and Lacroux, 2017). The field experiment by Wiles and colleagues shows the same direction in real hiring; no evidence was found that employers were less satisfied. The effect was strongest among those with weak writing; resumes that were already well written benefited little from the assistance.<\/p>\n<h2 id=\"one-page-a-summary-paragraph-and-reverse-chronological-order-how-strong-is-the-evidence\"><span class=\"ez-toc-section\" id=\"One-page-a-summary-paragraph-and-reverse-chronological-order-how-strong-is-the-evidence\"><\/span>One page, a summary paragraph and reverse chronological order: how strong is the evidence?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p><strong>Length.<\/strong> No experiment supporting the &ldquo;one page&rdquo; rule could be opened. According to Risavy&rsquo;s review, the consistent finding in the literature is not to exceed two pages; the review also reports that in a 1987 survey 80% of recruiters considered one page appropriate. The abstract of the only experiment that could be opened points the other way: Blackburn-Brockman and Belanger (2001) mailed one-page and two-page resumes of four fictitious candidates to 570 recruiters at large accounting firms; the two-page resumes were ranked more favorably. The candidates came from a single sector and were new graduates. The 2018 Ladders release also recommends a firm two-page limit for experienced candidates. A defensible rule: irrelevant lines are cut, a line that carries evidence is not sacrificed for the sake of page count, and two pages are not exceeded.<\/p>\n<p><strong>Summary paragraph.<\/strong> The review says that the career or job objective statement has traditionally been considered important, but that new research is needed to demonstrate the effect more conclusively. The 2018 Ladders release says that the resumes attracting the most attention carry a detailed summary at the top of the first page; the same company&rsquo;s 2012 report, by contrast, had said that descriptive text beyond the six data points had almost no bearing on the initial decision. The evidence is old and inconsistent. The job of the summary is therefore not to describe an objective but to enable a decision: role, seniority and the single strongest piece of evidence.<\/p>\n<p><strong>Order and design.<\/strong> The review reports that the traditional chronological format is strongly preferred to functional and creative alternatives (studies dated 1991-1995). According to the abstract of Arnulf and colleagues (2010), 90 participants, half of them professional recruiters, ranked 12 candidates; the formal layout was preferred to the &ldquo;creative&rdquo; design. The most recent finding is the field experiment by Kristal and colleagues. All the resumes belonged to a fictitious mother with nine years of experience; the version that stated the number of years worked in each job instead of employment dates received even more callbacks than the resume without a gap. According to the authors, the effect comes from experience becoming more salient. The finding concerns a single candidate profile; it cannot be treated as a general rule.<\/p>\n<p>Unemployment experiments also show that dates are read closely. In Kroft and colleagues, callbacks decline over the first eight months. According to the abstract of Eriksson and Rooth (2014), long spells of unemployment in the past do not affect the employer&rsquo;s decision, whereas a current spell of unemployment lasting at least nine months is viewed negatively.<\/p>\n<h2 id=\"do-tailoring-to-the-posting-and-the-cover-letter-still-work-in-the-age-of-ai\"><span class=\"ez-toc-section\" id=\"Do-tailoring-to-the-posting-and-the-cover-letter-still-work-in-the-age-of-AI\"><\/span>Do tailoring to the posting and the cover letter still work in the age of AI?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The mechanism behind tailoring is strong. For the 2021 report by Harvard Business School and Accenture, 2,275 executives in the United States, the United Kingdom and Germany were surveyed. More than 90% of employers said they use a recruitment management system (RMS) to initially filter or rank candidates (94% at the middle-skills level, 92% at the high-skills level). Among employers, 88% acknowledge that qualified high-skills candidates are screened out because they do not match the exact criteria in the job description. Nearly half of companies screen out resumes with an employment gap longer than six months. A document that does not meet the posting&rsquo;s criteria can drop out before it reaches a human.<\/p>\n<p>The causal evidence, however, exists not for CVs but for proposal letters on a freelancing platform. Cui, Dias and Ye (2025, preprint) studied the AI cover letter tool that Freelancer.com launched on April 19, 2023. The tool increased the textual alignment of letters with the posting and increased callbacks. Over the same period, the link between alignment and callbacks weakened by 51% and the link with receiving an offer by 79%; employers shifted weight to past work ratings on the platform. Most drafts were sent with almost no editing, while the time spent editing was positively associated with hiring success. On the same platform, Galdin and Silbert (2025, preprint) found that employers placed a high value on applications written specifically for the posting before language models, and did not afterwards. In the model scenario in which the letter carries no signal at all, the most able fifth is hired 19% less often and the bottom group 14% more often.<\/p>\n<p>The upshot: tailoring is still necessary, but it no longer sets a candidate apart. When everyone can produce &ldquo;aligned&rdquo; text with a single click, value shifts to what is hard to imitate: verifiable results, dates, work that can be cited as a reference. The same applies to the cover letter: the older literature supports the expectation, but an unedited draft carries no signal.<\/p>\n<h2 id=\"who-reads-the-cv-first-in-2026\"><span class=\"ez-toc-section\" id=\"Who-reads-the-CV-first-in-2026\"><\/span>Who reads the CV first in 2026?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Official data says less than is commonly assumed about how widespread machine reading is.<\/p>\n<table>\n<thead>\n<tr>\n<th>EU-27, enterprises with 10 or more employees<\/th>\n<th>2021<\/th>\n<th>2023<\/th>\n<th>2024<\/th>\n<th>2025<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Using at least one AI technology<\/td>\n<td>7.65%<\/td>\n<td>8.06%<\/td>\n<td>13.48%<\/td>\n<td>19.95%<\/td>\n<\/tr>\n<tr>\n<td>Same indicator, 250 or more employees<\/td>\n<td>28.41%<\/td>\n<td>30.48%<\/td>\n<td>41.17%<\/td>\n<td>55.03%<\/td>\n<\/tr>\n<tr>\n<td>Using AI for human resources management or recruiting<\/td>\n<td>0.62%<\/td>\n<td>no data<\/td>\n<td>no data<\/td>\n<td>no data<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><em>Source: Eurostat, isoc_eb_ai dataset. Values are as printed in the dataset.<\/em><\/p>\n<p>The share of enterprises using AI rose about 2.6-fold in four years; among large enterprises it rose by 26.62 points and passed the halfway mark (CEOtudent calculation). But the indicator that separately measures AI use in recruiting was published only for 2021. Because no current official rate could be opened, this article gives no number in answer to the question &ldquo;what percentage of companies screen resumes with AI&rdquo;.<\/p>\n<p>The rhetoric of skills-based hiring also calls for caution. According to data from Indeed Hiring Lab (a commercial source), 52% of Indeed job postings in the United States mentioned no formal education requirement in January 2024 (48% in 2019). According to the 2024 report by researchers at the Burning Glass Institute and Harvard Business School, however, in roles that dropped the degree requirement the share of hires without a degree rose by an average of 3.5 percentage points; that corresponds to fewer than 1 in 700 hires. Evidence of skill is necessary, but it has not replaced the degree line.<\/p>\n<h2 id=\"lead-like-a-ceo-how-is-a-cv-built-as-a-decision-document\"><span class=\"ez-toc-section\" id=\"Lead-like-a-CEO-how-is-a-CV-built-as-a-decision-document\"><\/span>Lead like a CEO: how is a CV built as a decision document?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Through the CEO lens, a CV is not a life story but a decision document presented to a recipient who is short on time. It is managed like a product: it has a target, a value proposition and a pre-submission test.<\/p>\n<p><strong>CEOtudent editorial framework: one-page CV decision checklist<\/strong><\/p>\n<table>\n<thead>\n<tr>\n<th>Decision<\/th>\n<th>Question<\/th>\n<th>Finding it rests on<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Target<\/td>\n<td>Was this version written for a single role family?<\/td>\n<td>The first glance concentrates on a few fields (Ladders, 2012, commercial)<\/td>\n<\/tr>\n<tr>\n<td>First third<\/td>\n<td>Are the role, the seniority and the strongest evidence at the top of the document?<\/td>\n<td>The evidence for summaries is weak; their function is to enable a decision<\/td>\n<\/tr>\n<tr>\n<td>Evidence<\/td>\n<td>Does every experience line contain a number, a method or a named result?<\/td>\n<td>A skills list has no effect, the quality of experience does (Kessler et al., 2019)<\/td>\n<\/tr>\n<tr>\n<td>Dates<\/td>\n<td>Are the current status and the durations clear; has any gap been presented deliberately?<\/td>\n<td>Kroft et al.; Eriksson and Rooth; Kristal et al.<\/td>\n<\/tr>\n<tr>\n<td>Language<\/td>\n<td>Are the posting&rsquo;s actual terms met honestly?<\/td>\n<td>Systems filter by criteria (Harvard Business School and Accenture, 2021)<\/td>\n<\/tr>\n<tr>\n<td>Errors<\/td>\n<td>Has the text been checked by a spelling tool and a second pair of eyes?<\/td>\n<td>Sterkens et al.; Wiles et al.<\/td>\n<\/tr>\n<tr>\n<td>Length and layout<\/td>\n<td>Within two pages, with a plain format and standard headings?<\/td>\n<td>Risavy, 2017; Arnulf et al., 2010<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><em>CEOtudent editorial framework; the right-hand column shows the source each question rests on.<\/em><\/p>\n<p><strong>The 15-minute CV test.<\/strong> Before submission, with a timer running:<\/p>\n<ol>\n<li><strong>1 minute.<\/strong> The document is shown for 7 seconds to someone unfamiliar with it. Can they name the target role, the seniority and the most recent employer? (The duration is a strict limit, not a scientific threshold.)<\/li>\n<li><strong>3 minutes.<\/strong> Each line in the first third is put to one question: does this line change the decision? A line that does not is moved down or deleted.<\/li>\n<li><strong>4 minutes.<\/strong> The mandatory criteria in the posting are marked; for each one, a verifiable line is sought in the CV.<\/li>\n<li><strong>3 minutes.<\/strong> Experience lines that contain no number, method or result are counted; the weakest half are rewritten.<\/li>\n<li><strong>3 minutes.<\/strong> The text is run through a spelling tool, then read from the end to the beginning. Even two errors mean a measurable loss.<\/li>\n<li><strong>1 minute.<\/strong> The file name, the date format and the contact details are checked.<\/li>\n<\/ol>\n<p>The test is a checking tool, not a validated scale.<\/p>\n<h2 id=\"learn-like-a-student-how-is-an-application-experiment-logged\"><span class=\"ez-toc-section\" id=\"Learn-like-a-student-how-is-an-application-experiment-logged\"><\/span>Learn like a student: how is an application experiment logged?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Through the student lens, every application round is a data point. But the data is noisy: callback rates in the experiments are low and vary widely by market. The table below uses three published rates to calculate the probability of getting no callback at all in a round.<\/p>\n<p><strong>CEOtudent analysis: probability of zero callbacks by number of applications<\/strong><\/p>\n<table>\n<thead>\n<tr>\n<th>Callback rate per application<\/th>\n<th>10 applications<\/th>\n<th>20 applications<\/th>\n<th>30 applications<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>4% (Kroft et al., eighth month of unemployment)<\/td>\n<td>66.5%<\/td>\n<td>44.2%<\/td>\n<td>29.4%<\/td>\n<\/tr>\n<tr>\n<td>7% (Kroft et al., first month of unemployment)<\/td>\n<td>48.4%<\/td>\n<td>23.4%<\/td>\n<td>11.3%<\/td>\n<\/tr>\n<tr>\n<td>24% (Kline et al., large employers)<\/td>\n<td>6.4%<\/td>\n<td>0.4%<\/td>\n<td>0.03%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><em>CEOtudent analysis. Calculation: (1 &#8211; rate) to the power of the number of applications; applications are assumed to be independent. The rates come from US experiments with fictitious applications; they are not a personal forecast.<\/em><\/p>\n<p>How to read it: in the 4-7% band, a round of 10 applications passing in complete silence is as likely as a coin toss, or more likely. Rewriting the CV from scratch on the result of a single round is reacting to noise. The rules of the experiment:<\/p>\n<ul>\n<li>Only one variable changes in a round.<\/li>\n<li>The size of the round is written down in advance; at least 20 applications.<\/li>\n<li>The outcome measure is chosen in advance: a callback within 30 days.<\/li>\n<li>Untailored mass applications are not included in the experiment.<\/li>\n<\/ul>\n<p><strong>CEOtudent editorial framework: application experiment logging template<\/strong><\/p>\n<table>\n<thead>\n<tr>\n<th>Field<\/th>\n<th>Record<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Round and date range<\/td>\n<td>Round number, start and end<\/td>\n<\/tr>\n<tr>\n<td>Target role family<\/td>\n<td>A single role<\/td>\n<\/tr>\n<tr>\n<td>CV version and the one thing changed<\/td>\n<td>Example: three evidence lines instead of an objective statement<\/td>\n<\/tr>\n<tr>\n<td>Number of applications and channel<\/td>\n<td>Number set in advance; job posting, referral or direct outreach<\/td>\n<\/tr>\n<tr>\n<td>Callbacks within 30 days<\/td>\n<td>Number and rate<\/td>\n<\/tr>\n<tr>\n<td>Converted to interview<\/td>\n<td>Number<\/td>\n<\/tr>\n<tr>\n<td>Feedback received<\/td>\n<td>Verbatim<\/td>\n<\/tr>\n<tr>\n<td>Decision for the next round<\/td>\n<td>Continue, change or drop<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h2 id=\"what-do-these-findings-not-mean\"><span class=\"ez-toc-section\" id=\"What-do-these-findings-not-mean\"><\/span>What do these findings not mean?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<ul>\n<li>The field experiments were conducted with fictitious candidates in specific countries, sectors and periods; the rates cannot be carried over directly to other markets.<\/li>\n<li>The AI studies are preprints and concern a single freelancing platform; they cannot be generalized to corporate hiring.<\/li>\n<li>The spelling experiments rest on hypothetical scenarios; a CV influences the recipient&rsquo;s decision, it does not measure the candidate&rsquo;s worth.<\/li>\n<li>This article is for general information; it is not personalized career or legal advice.<\/li>\n<\/ul>\n<h2 id=\"why-was-this-article-rewritten\"><span class=\"ez-toc-section\" id=\"Why-was-this-article-rewritten\"><\/span>Why was this article rewritten?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The 2021 version presented its five recommendations without citing sources, addressing the reader informally in the second person and in categorical terms. Five changes were made in this update:<\/p>\n<ol>\n<li>The eight claims were matched to the sources that could be opened and each was given a verdict.<\/li>\n<li>The sentence &ldquo;Research &hellip; only seven seconds&rdquo; was traced to its source and labeled as a commercial, small-sample measurement.<\/li>\n<li>Statements for which no source could be found were removed: that a resume should usually be one page, the 10-year experience threshold, that adding keywords makes a candidate stand out in the ATS, and the promise of a flawless resume.<\/li>\n<li>The old images were removed; the style was made editorial.<\/li>\n<li>The experiment table, the evidence scorecard, the Eurostat table, the 15-minute test and the logging template were added.<\/li>\n<\/ol>\n<h2 id=\"frequently-asked-questions\"><span class=\"ez-toc-section\" id=\"Frequently-asked-questions\"><\/span>Frequently asked questions<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p><strong>Do recruiters really look at a CV for 7 seconds?<\/strong> The number comes from a career site&rsquo;s eye-tracking study: 6 seconds with 30 recruiters in 2012, and 7.4 seconds in the 2018 release. What is measured is the initial screening glance; a peer-reviewed article reports the screening time as about 45 seconds.<\/p>\n<p><strong>Should a CV be one page?<\/strong> The sources opened do not support this as a general rule. The consistent finding is not to exceed two pages; in the only experiment that could be opened, two-page resumes were ranked more favorably.<\/p>\n<p><strong>Does a single spelling error end an application?<\/strong> That is not known; the lowest level measured in the experiment opened was two errors, which lowered the probability of an interview by 7.3 points. Five errors lowered it by 18.5 points.<\/p>\n<p><strong>Is it a problem to write a CV with AI?<\/strong> Writing assistance that flags errors raised hiring by 8% in the experiment with 480,948 people. On the other hand, on the freelancing platform the time spent editing the draft was associated with success.<\/p>\n<p><strong>Is a cover letter still necessary?<\/strong> The older literature supports the expectation of a cover letter. After the AI tool, the link between cover letter alignment and callbacks weakened by 51%.<\/p>\n<h2 id=\"sources\"><span class=\"ez-toc-section\" id=\"Sources\"><\/span>Sources<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<ol>\n<li>TheLadders. Keeping an eye on recruiter behavior: New study clarifies recruiter decision-making. 2012. Company report, commercial; a copy of the document on a university server was opened.<\/li>\n<li>Ladders, Inc. Ladders Updates Popular Recruiter Eye-Tracking Study With New Key Insights on How Job Seekers Can Improve Their Resumes. Press release, PR Newswire, November 6, 2018. Commercial; the 2018 report itself could not be opened.<\/li>\n<li>Risavy SD. The Resume Research Literature: Where Have We Been and Where Should We Go Next? Journal of Educational and Developmental Psychology, 2017; 7(1): 169-187. Full text; review, the older studies it reports were not opened.<\/li>\n<li>Sterkens P, Caers R, De Couck M, Van Driessche V, Geamanu M, Baert S. Costly mistakes: Why and when spelling errors in resumes jeopardise interview chances. PLOS ONE, 2023; 18(4): e0283280. Full text.<\/li>\n<li>Martin-Lacroux C, Lacroux A. Do Employers Forgive Applicants&rsquo; Bad Spelling in R\u00e9sum\u00e9s? Business and Professional Communication Quarterly, 2017; 80(3): 321-335. Abstract only.<\/li>\n<li>Wiles E, Munyikwa ZT, Horton JJ. Algorithmic Writing Assistance on Jobseekers&rsquo; Resumes Increases Hires. NBER Working Paper 30886, 2023; Management Science, 2025; 71(12): 10144-10164. Working paper, full text.<\/li>\n<li>Kristal AS, Nicks L, Gloor JL, Hauser OP. Reducing discrimination against job seekers with and without employment gaps. Nature Human Behaviour, 2023; 7(2): 211-218. Full text.<\/li>\n<li>Kroft K, Lange F, Notowidigdo MJ. Duration Dependence and Labor Market Conditions: Theory and Evidence from a Field Experiment. NBER Working Paper 18387, 2012. Full text.<\/li>\n<li>Eriksson S, Rooth DO. Do Employers Use Unemployment as a Sorting Criterion When Hiring? Evidence from a Field Experiment. American Economic Review, 2014; 104(3): 1014-1039. Abstract only.<\/li>\n<li>Bertrand M, Mullainathan S. Are Emily and Greg More Employable than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination. NBER Working Paper 9873, 2003; American Economic Review, 2004; 94(4): 991-1013. Working paper, full text.<\/li>\n<li>Kline PM, Rose EK, Walters CR. Systemic Discrimination Among Large U.S. Employers. NBER Working Paper 29053, 2022; Quarterly Journal of Economics, 2022; 137(4): 1963-2036. Working paper, full text.<\/li>\n<li>Kessler JB, Low C, Sullivan C. Incentivized Resume Rating: Eliciting Employer Preferences without Deception. NBER Working Paper 25800, 2019; American Economic Review, 2019; 109(11): 3713-3744. Working paper, full text.<\/li>\n<li>Blackburn-Brockman E, Belanger K. One Page or Two? A National Study of CPA Recruiters&rsquo; Preferences for Resume Length. Journal of Business Communication, 2001; 38(1): 29-45. Abstract only.<\/li>\n<li>Arnulf JK, Tegner L, Larssen \u00d8. Impression making by r\u00e9sum\u00e9 layout: Its impact on the probability of being shortlisted. European Journal of Work and Organizational Psychology, 2010; 19(2): 221-230. Abstract only.<\/li>\n<li>Cui J, Dias G, Ye J. Signaling in the Age of AI: Evidence from Cover Letters. arXiv 2509.25054, 2025. Preprint, full text.<\/li>\n<li>Galdin A, Silbert J. Making Talk Cheap: Generative AI and Labor Market Signaling. arXiv 2511.08785, 2025. Preprint, full text.<\/li>\n<li>Fuller J, Raman M, Sage-Gavin E, Hines K et al. Hidden Workers: Untapped Talent. Harvard Business School and Accenture, 2021. Version corrected on October 4, 2021, full text; the co-publisher is a consulting firm.<\/li>\n<li>Eurostat. Artificial intelligence by size class of enterprise (isoc_eb_ai), EU-27, 2021-2025.<\/li>\n<li>Stahle C. Educational Requirements Are Gradually Disappearing From Job Postings. Indeed Hiring Lab, February 27, 2024. Commercial.<\/li>\n<li>Sigelman M, Fuller J, Martin A. Skills-Based Hiring: The Long Road from Pronouncements to Practice. Burning Glass Institute, February 2024. Full text.<\/li>\n<\/ol>\n<p><em>Note: The &ldquo;evidence scorecard&rdquo; and &ldquo;probability of zero callbacks&rdquo; tables are CEOtudent analysis; the decision checklist, the 15-minute test and the logging template are a CEOtudent editorial framework; the loss per error, the multiples, the point difference and the 2,880 total are CEOtudent calculations. The old claims that were removed are listed in the section &ldquo;Why was this article rewritten?&rdquo;. The widely quoted rates claiming that applicant tracking systems (ATS) reject most resumes before a human sees them were not used because their primary source could not be opened. The 1999 study by Thoms and colleagues is not reported because it could not be opened.<\/em><\/p>\n<hr>\n<p><em>This content was compiled with the support of AI following in-depth research, then written and prepared for publication by the CEOtudent editorial team.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>The 2021 list of &#8216;5 secrets of effective CV preparation&#8217; and the &#8216;7 seconds&#8217; claim were tested against primary sources: 1 of 8 claims was verified, 5 are partly supported, and no source was found for 2. In an experiment with 445 recruiters, five spelling errors lowered the probability of an interview by 18.5 points and two errors by 7.3 points; in a field experiment with 480,948 job seekers, writing assistance raised hiring by 8%. The article also offers a 15-minute CV test and a logging template for application experiments.<\/p>\n","protected":false},"author":4640,"featured_media":126759,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[17219],"tags":[],"class_list":["post-156979","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-career"],"_links":{"self":[{"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/posts\/156979","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/users\/4640"}],"replies":[{"embeddable":true,"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/comments?post=156979"}],"version-history":[{"count":0,"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/posts\/156979\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/media\/126759"}],"wp:attachment":[{"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/media?parent=156979"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/categories?post=156979"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ceotudent.com\/en\/wp-json\/wp\/v2\/tags?post=156979"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}