ATS Myths, Where They Came From and Which Ones Are Actually True

There is a whole folklore around applicant tracking systems, and it is unusually resistant to correction. Part of the reason is that the people repeating it are rarely lying, they are passing on something they read. Another part is that a good deal of it is sold by companies with a product to move.

The result is that job seekers spend their energy defending against an imaginary machine while the real reasons for rejection sit untouched. This article goes through the myths one at a time, says where each came from where that is knowable, and ends with the warnings that are genuinely worth heeding.

The underlying model, if you want it first, is in what an applicant tracking system actually is.

Myth: 75% of CVs are rejected before a human sees them

This is the single most quoted statistic in the entire field, and you will find it on hundreds of sites, usually with no source at all or with a link to another site that also has no source.

Follow the citations back and they run out. There is no study, no methodology, no sample. The figure circulates because it is repeated, which is a different thing from being true.

What makes it especially misleading is not the number but the framing. "Rejected before a human sees them" implies an automated judgement of quality. What actually happens is that most applications are never opened because a recruiter searched, looked at the first thirty results and stopped. That is a human running out of time, not software passing a verdict.

The practical difference matters. If the machine is rejecting you, you fix the file. If a recruiter never got to you, you fix which words you use and which jobs you apply to.

Myth: there is an ATS score you have to beat

Some systems display a match percentage next to a candidate. It is a sorting aid, calculated by comparing the text of your CV to the text of the job description, and recruiters routinely ignore it because they know how crude it is.

There is no industry standard threshold. No 75%, no 80%, nothing. Different systems calculate different things, many show nothing at all, and no employer has a policy of discarding everyone under a number.

The scores you see from CV checking tools, including ours, are that tool's own estimate. They are useful as a mirror, showing you which requirements from an advert are missing from your document. They are not the number an employer sees, because there is no such number.

Myth: white text keyword stuffing gets you through

The mechanism is real, which is why the myth survives. A parser reads text regardless of its colour, so hidden keywords do get indexed.

Then a human opens the document. They select the text to copy a phrase, or the file gets converted, or they simply scroll and see a strange blank area, and the trick is visible. What follows is not a formatting complaint, it is a judgement about your honesty, applied to everything else on the page.

There is also the small matter that it does not work even on its own terms. Getting into a search result is only useful if the document survives the reading that follows.

Myth: ATS cannot read PDFs

This was worth worrying about roughly fifteen years ago. It has not been true for a long time, and text based PDF is now the safer default for almost every application, because it also preserves your layout for the human reader.

The kernel of truth is that some PDFs contain no text at all: a scan, a photo, or a file exported as flattened artwork from a design tool. Those parse to nothing. The test is whether you can select the text with your cursor.

Send Word when the advert or a recruitment agency asks for it, which is common because agencies often rebrand CVs before passing them on. Otherwise PDF.

Myth: you need an ATS approved template

There is no approval body. No certification exists, so the phrase is marketing language and nothing more.

What those templates usually do have is a single column and plain headings, which genuinely helps. You can get the same result in an empty document for free, and you can get it in something that looks considerably better than most of what is sold under that label.

The irony is that template galleries rank by thumbnail appeal, so the designs that rise to the top are the two column ones with sidebars and skill bars, which are exactly the designs that cause the problems described in the formatting that actually breaks an ATS.

Myth: keyword density matters

You will see advice to repeat your key terms a certain number of times, or to hit a percentage of the document.

Search does not work that way. A search on a database field is a yes or no on presence. Mentioning "Python" four times does not make you more findable than mentioning it once, it just costs you space and makes the writing worse.

What does matter is coverage: whether the essential requirements from the advert appear anywhere in your document, and whether they appear attached to evidence rather than floating in a list. That method is in ATS keywords.

Myth: the software ranks everyone and the top ten go forward

Recruiters build shortlists by searching, filtering and reading, not by taking the top slice of an automated ranking. Where a ranking exists it is one column among several and it is not binding on anyone.

This myth is comforting in a strange way, because it suggests a fair contest with clear rules. The reality is messier and more human: someone with 280 applications and a meeting in forty minutes is looking for eight profiles that obviously fit, and obviousness is doing most of the work. That process is described from the inside in why your CV gets rejected.

Myth: the ATS rejects you for an employment gap

Parsers record dates. They do not evaluate the meaning of a space between two of them, and no system has a rule that discards candidates with a break in their history.

Gaps are a human question, and they are a mild one. A short line of explanation in the document usually settles it entirely, as set out in explaining employment gaps.

What can genuinely hurt you is a parse that loses your dates, so your record shows a gap that does not exist or less experience than you have. That is a formatting problem wearing a gap costume.

Myth: two pages will get you filtered out

The parser reads the whole document. Page count is a convention for human readers, and two pages is standard for experienced candidates across most of Europe.

The related belief that a longer CV improves your keyword coverage is also wrong, for the same reason density does not matter. More pages of weak material do not improve your match, they dilute the reader's attention. The proper treatment is in how long a CV should be.

Myth: applying at the right time beats the system

Advice about applying on Tuesday morning, or within the first hour, treats the pipeline as a queue that gets processed in order.

Most reviewing happens in batches, often after the advert closes, and frequently weeks after you applied. There is a genuine advantage to applying early for roles that fill fast or that close once enough applications arrive, but it is about the role's lifecycle, not about outsmarting software.

Myth: creative section headings show personality

They show personality to a human and nothing at all to a parser, which is looking for recognisable words to decide where your experience section starts.

This is the cheapest possible fix on the entire list. "Experience" costs you nothing and works everywhere. Save the personality for the summary and the bullets, where it can be read by the person deciding.

Myth: the same CV works everywhere if it is optimised enough

This one is rarely stated outright, but it underlies a lot of behaviour. The idea is that there exists a correctly built CV which performs well against any system, and once you have produced it the job is done.

There is no such document, and the reason is that the searches differ. One employer searches for "procurement", another for "purchasing", a third for a specific system you happen to have used. A file optimised for an abstract average of all adverts is optimised for none of them.

The practical consequence is that fifteen minutes of tailoring per application beats any amount of one time optimisation. Rewriting your summary for the role and moving your three most relevant bullets upward does more than every formatting decision on this page combined. The method is in tailoring your CV to a job description.

Myth: recruiters never see the original file

A surprisingly common belief, and it makes the other myths worse, because if you think only the parsed text matters then design and readability seem pointless.

The original file is stored as an attachment and is one click away in every mainstream system. The parsed record is what gets searched and what fills the columns in a list view. The document itself is what gets read once someone decides to look.

That is why the correct approach is not to strip your CV down to plain text. It is to keep a clean single column layout that parses reliably and still reads well to a person, because both audiences exist and they see different things.

Why the myths persist

Worth naming, because it explains why the same claims keep reappearing.

Someone is selling a solution. Template sites, CV writing services and optimisation tools all benefit from a widely believed problem with a purchasable fix. An unsourced statistic about automated rejection is extremely good marketing.

The myths are more comfortable than the truth. A machine rejecting you on a technicality is easier to sit with than a person reading your work and deciding it was not strong enough. The first is fixable by settings, the second by effort.

They are easy to act on. Changing a font takes a minute. Rewriting eighteen bullets to show results takes an afternoon. People do the thing they can finish.

Old advice ages badly. Some of this was accurate in 2010, when parsers were worse and PDF support was genuinely patchy. It was passed on faithfully and never re-checked.

The warnings that are actually true

Fairness requires this section, because dismissing everything is its own kind of misinformation. Four things people call myths are real.

Two column layouts get mangled. A parser reading in visual order can interleave a sidebar with the main content. Some modern systems handle it well and you cannot tell which you are submitting to.

Content in headers and footers is frequently dropped. This one is genuinely dangerous because it usually takes your phone number with it. You can be shortlisted and never contacted.

Text inside images is invisible. Not partly, entirely. A CV exported as an image produces an empty record.

Knockout questions really do reject you automatically. Salary expectations, notice period, years of experience, right to work. This is the automated rejection everyone imagines, and it is the one almost nobody checks. It is also the easiest to fix, since it is a matter of reading the form carefully.

Notice that three of the four are specific, checkable and fixable once. None of them require a special template or a font.

Frequently asked questions

So does the ATS matter at all?

Yes, but as a database rather than a judge. It decides whether you appear in a recruiter's search results and what your record looks like in a list view. It does not decide whether you are good.

Should I ignore CV checking tools then?

No, use them for what they can actually tell you: what a parser extracted from your file, and which requirements from an advert are missing. Ignore any score presented as the number an employer will see.

Is it worth submitting both a PDF and a Word file?

No. Most systems will either overwrite the first with the second or store one and never open it, and small differences between the two look like carelessness. Send the format that was asked for.

Do any systems really use artificial intelligence to score candidates?

Some vendors offer matching features, and their use varies widely by employer and is increasingly constrained by regulation on automated decision making. Even where they exist, they typically surface suggestions to a recruiter rather than reject anyone outright.

What about the advice to mirror the job title exactly?

That one is sound, and it is about search rather than scoring. If the advert says Management Accountant, having that phrase somewhere honest on your CV makes you findable by the most obvious search term there is.

How do I know whether formatting is my actual problem?

Copy your PDF into a plain text editor and read the result. If it reads in the right order and contains your contact details, formatting was never the issue and you can stop thinking about it.

Replace the guesswork with the actual output

Every myth on this page exists because nobody can see what happens to their file after they upload it. CvLaunch removes that blind spot: it reads your CV the way an employer's system does and shows you the extracted result, which roles, dates, sections and skills came through and what was lost. Once you can see it, you can stop arguing with folklore. It takes a few seconds, it is free, and there is no sign up.