You have seen the number. Seventy-five percent of CVs are rejected by applicant tracking software before a human ever reads them. It appears in blog posts, LinkedIn advice, and the marketing copy of roughly every résumé tool on the internet.
It is not true, and the story of where it came from is more useful than the number ever was.
Where the number actually comes from
The figure traces back to a 2012 sales pitch from a company called Preptel. Preptel sold résumé optimisation software. Their product existed to get your CV past applicant tracking systems, which gave them a direct commercial interest in job seekers believing that applicant tracking systems were silently binning three quarters of everything they received.
No study was ever published. No sample size, no methodology, no data set. Preptel went out of business in 2013.
The number outlived the company by more than a decade, which is a decent illustration of how this kind of thing works. It also drifts as it travels. You will see it quoted as 70%, as 75%, and as 88%, sometimes on the same site. A real statistic does not wobble by eighteen points depending on who is repeating it. That drift is the tell.
Why it keeps getting repeated
Fear is useful if you are selling the cure. “Your CV is probably being deleted unread” is an extremely effective sentence if the next sentence is “our ATS-optimised templates fix that.”
We build a CV tool, so we are in the same category as most of the sites repeating this, and we are telling you the number is fake. Take that for whatever it is worth. Our view is that a tool worth using should not need you to be frightened first.
What actually happens to your CV
Enhancv interviewed 25 US recruiters about how their systems are configured. Ninety-two percent said they do not use the ATS to auto-reject CVs for formatting or design. Only two out of twenty-five had anything configured to reject on content or match score (the full write-up is here).
An applicant tracking system, in normal use, is a database. It stores applications, parses them into searchable fields, and lets a recruiter search and filter. It is closer to a filing cabinet with a search box than to a bouncer.
There is one real exception worth knowing about, because overcorrecting is its own mistake. Many application forms include knockout questions: are you legally able to work in this country, do you have the required licence, are you willing to relocate. Answers to those genuinely can filter you out automatically. That is real auto-rejection, and it happens on the form, not to your CV.
The thing that actually costs you the job
Here is the more useful way to think about it. Your CV is rarely rejected. It gets buried.
A recruiter with four hundred applicants does not read four hundred CVs. They open the system, search something like “senior marketing manager Boston,” and work through the results. If your CV does not surface in the first fifty, you never hear back. From your side that is indistinguishable from rejection. Mechanically it is not rejection at all, it is a search result you did not appear in.
This reframes what is worth your effort. The question is not “will a machine throw this out.” It is “when a human searches for someone like me, do I come up, and when they open me, do they see it in six seconds.”
What follows from that
Use the words the job posting uses. Not to satisfy a keyword-matching algorithm, but because those are the words the recruiter will type into the search box, and they are working from the same posting you are reading. This is the entire practical core of what people call ATS optimisation, and it is genuinely worth doing. It is also just tailoring your CV to the job, which you should be doing anyway.
Apply earlier. Being in the first fifty applications matters more than almost anything you can do to the document itself.
Make it readable in six seconds by a tired human. Clear role titles, dates that are easy to scan, the relevant thing near the top. This is the same advice as making it parse well, because the things that confuse software (text inside images, critical information in headers and footers, unlabelled sections) are also the things that slow a person down.
What to stop worrying about
- Your font. Nobody is being filtered on Calibri versus Arial. Use something legible and move on.
- Templates marketed as “ATS-friendly.” Almost any clean, single-purpose layout parses fine. This label is mostly a sales feature.
- Keyword density. There is no percentage to hit. Use the relevant words where they are true, and stop.
- Two columns. Modern parsers handle them. Just keep the reading order sane and do not put anything critical in a sidebar image.
- Submitting a PDF. PDFs are parsed fine and have been for years, unless the PDF is a scan of a printed page, which genuinely is unreadable to software. Do not scan your CV.
The short version
The 75% number is a marketing claim from a dead company that never published a study. Applicant tracking systems mostly file and search rather than reject. The real risk is being invisible in a search among hundreds of other applicants, which you address by using the language of the posting, applying early, and being readable fast.
That is a much less frightening problem, and unlike the imaginary one, you can actually do something about it.