What an ATS actually is and does
It is a database with a search box, not a robot that reads resumes and rejects people. Here is what really happens to your application, and where the famous 75% myth comes from.
The one-sentence answer
An applicant tracking system is the software a company uses to collect applications, store them as searchable profiles, and move candidates through hiring steps. Workday, Greenhouse, Lever, Taleo, iCIMS and SAP SuccessFactors are the names you will meet most often.
The part that matters for your resume: an ATS is built for recruiters to organise and find applications, the way an inbox organises mail. It is not built to judge them, and out of the box, none of the major systems reject anybody on their own.
What happens the moment you click Apply
Your file is parsed. The system reads the text of your resume and fills a structured profile from it: name, contact details, work history with employers and dates, education, skills. That profile, not your carefully designed page, is what recruiter searches and filters run against.
Your original file survives. Recruiters can and do open the actual PDF you sent. Parsing decides whether you are findable; the file decides what a human sees once they find you. Both matter, which is why tricks aimed only at the parser fail; more on that in how to pass screening.
Parsing is dumber than you expect. It reads top to bottom, looks for familiar headings like "Work experience" and "Education", and matches date patterns. A two-column layout, a table, or contact details drawn as icons can scramble what lands in the profile. You can see this yourself: the free ATS checker runs a parser on your resume and shows you the extracted profile.
How recruiters actually use it: search
A posting attracts more applications than anyone reads front to back. The recruiter deals with that the way you deal with a full inbox: search and filters. They search for the terms the role needs (a certification, a tool, a job title) and read the profiles that match first.
This is the single most practical fact about the ATS. Nobody scores you out of 100. You appear in a search, or you do not. Which is why mirroring the posting’s own words, where they are true of you, moves your odds more than any formatting trick; the screening guide covers doing that honestly.
Filters and knockout questions: the real rejections
Automatic rejection does exist in one specific, mundane form: knockout questions. "Are you authorised to work in this country?" "Do you hold this licence?" "Can you work this shift?" Answer outside what the recruiter configured, and the system tags or declines the application, exactly as instructed.
Note who made that decision: a person, writing a rule. The ATS executed it. That distinction is not pedantry; it tells you what to fix. You cannot format your way past a work-authorisation requirement, and you do not need to beat a reading robot that does not exist. You need to be findable in search, readable as a profile, and honest in the questions.
The "robots reject 75% of resumes" myth
You have seen the claim: three quarters of resumes are thrown out by software before a human ever looks. It is the founding legend of an entire industry of resume-fixing tools, and it does not hold up.
The number traces back to around 2012 and a company called Preptel, which sold resume optimisation and cited no study, no methodology and no sample for the figure. The company folded soon after; the statistic outlived it because it is excellent marketing: terrifying, precise-sounding, and unfalsifiable. No academic study has ever reproduced it, and every major ATS vendor states their software does not auto-reject.
What is true is less cinematic and better documented. A 2021 Harvard Business School and Accenture study ("Hidden Workers") surveyed more than 2,250 executives: 88% agreed their own screening setup filters out qualified candidates, not because software reads and dislikes resumes but because rigid, human-configured filters exclude people who fail one literal criterion, a gap in employment, or a missing keyword.
So the honest summary: the robot judge is fiction; the badly configured filter is real. The fix is not paying someone to "beat the ATS": it is a resume that parses cleanly, uses the vocabulary of the job, and answers requirements truthfully.
If a tool or coach tells you software rejected your resume for fonts, formatting or a low "ATS score", they are selling the cure to a disease they invented. Rejection by score is not how any of these systems work.
What this means for your resume
Three things follow from how the system actually works, and they are the whole game:
- Be parseable: real text, plain headings, readable dates; the ATS-friendly formatting guide covers exactly what breaks parsing and what does not.
- Be findable: use the words the recruiter will search for, where they are true of you.
- Be honest at the knockouts: those questions are the one place automation really says no, and a false answer surfaces at the background check anyway.
Common questions
Does an ATS automatically reject resumes?
Not by reading them. The only automatic rejections come from knockout questions and filters a human configured: work authorisation, licences, availability. No major system scores your resume and declines it on formatting or keywords.
Do recruiters see my actual resume, or just the parsed version?
Both. Searches and filters run on the parsed profile; once you surface, recruiters open the original file. That is why hidden-keyword tricks fail: the human sees the real page.
Is there a real "ATS score" for my resume?
No system used by employers scores resumes out of 100. Checker scores, including ours, measure how much of a resume a parser can read: useful, but a readability measure, not a hiring prediction.
How do I know which ATS a company uses?
The application URL usually gives it away: greenhouse.io, lever.co, myworkdayjobs.com, icims.com. It rarely changes what you should do: clean parsing and honest keyword use work across all of them.