Why your CV gets rejected before a human reads it
The 10 dimensions ATS and recruiters actually score
· 5 min read · ATS
The most repeated statistic in job hunting is that 75% of CVs are rejected by an ATS before a human sees them. It gets quoted in LinkedIn posts, CV-writing ads and career webinars every single day.
Nobody can point to the study. It doesn't exist.
But the frustration behind it is completely real: you apply, you match the role, and you never hear back. So let's replace the myth with the mechanics — what actually happens to your CV in the first few minutes after you hit submit, and the ten things that get scored along the way.
Applicant tracking systems don't reject you. Three other things do.
An ATS is a database, not a judge. It stores applications, moves them through stages and lets a recruiter search. In most configurations it never assigns a pass/fail verdict on its own. What does eliminate you, in this order:
1. Knockout questions — instant, automatic, final
Those short questions in the application form ("Are you legally authorized to work in this country?", "Do you have 5+ years of experience with X?", "What are your salary expectations?") are the only true auto-reject in the pipeline. Answer one outside the accepted range and you are filtered out in milliseconds. Your CV is never opened.
2. The recruiter's search — you're either in the result set or invisible
With 400 applicants, no recruiter reads 400 CVs. They run a search inside the ATS: job title AND skill AND skill, sorted by recency. If your CV doesn't contain the literal terms they searched for, you're not rejected — you simply never appear. This is the single biggest silent killer, and it's why "Growth Lead" won't surface in a search for "Marketing Manager."
3. The scan — a handful of seconds, then a decision
Once you're in the result set, a human gives your CV a fast triage pass. Eye-tracking studies consistently put this between 6 and 30 seconds. They're looking at your most recent title, your current employer, your dates, and whether the top third of page one answers "could this person do this job?"
You are not being evaluated. You are being sorted. Those are very different problems, and they need different fixes.
The 10 dimensions that actually get scored
Whether it's a recruiter's search, a screening checklist or an AI matching layer, the same ten signals decide whether you move forward:
- Title match — does your most recent title resemble the one they're hiring for?
- Exact keywords — the literal skill and tool names from the job description, spelled their way
- Years of experience — extracted from your dates, not from your claims
- Seniority signal — scope, team size, budget, decision rights
- Hard skills coverage — how many of the stated requirements you visibly meet
- File parsability — whether the machine can read your layout at all
- Recency — a skill used in 2016 does not score like one used this year
- Location and work authorization — often a hard gate, not a preference
- Quantified impact — numbers attached to outcomes, not lists of duties
- Tenure pattern — the story your dates tell before you get to explain it
The three that cut most people
File parsability. Two-column layouts, text boxes, tables, icons, headshots and contact details inside the header or footer are the classic offenders. Your PDF looks perfect and parses as fragments — or worse, as nothing. Test it: open your PDF, select all, copy, paste into a plain text file. If it comes out scrambled or half-empty, that is what the recruiter's search is indexing.
Exact keywords. "Led cross-functional delivery" is not the same string as "program management." Human readers understand the synonym. A keyword search does not. Mirror the job description's vocabulary — honestly, for things you've actually done.
Quantified impact. "Responsible for the migration" tells a screener nothing. "Migrated 40 services to AWS in 9 months, cutting infra spend 28%" survives a six-second scan and gives the recruiter something to repeat to the hiring manager. You are writing the argument they'll make on your behalf when you're not in the room.
What to do this week
- Run the copy-paste parse test on your CV. Fix the layout before you touch the wording.
- Put your target job title in your CV — as your headline, or in your summary line.
- Pull the 8–10 recurring terms from three real job postings you want. Work the true ones into your experience.
- Add a number to your top five bullets. Percentage, currency, volume, or time.
- Stop applying to 40 roles with one generic CV. Apply to 8 with a version that matches.
None of this is about gaming a robot. It's about removing the friction between what you've done and the three-second decision someone makes about it. The candidates who get interviews aren't always the strongest — they're the ones whose CV made the match obvious.
Want to see your own scorecard? labor.quest analyzes your CV against a specific job description and shows you exactly where the gaps are — before you apply.
All posts