Do applicant tracking systems reject resumes?
What recruiters say, what parsers need, and what an ATS score measures
Survey of 25 recruiters September and October 2025 No score shown here
Summary
Many resume tools sell an “ATS score” against a robot that, according to the recruiters who run these systems, mostly does not do what the tools say. Here is the survey, its sample size, what an ATS does instead, what a parser needs from your file, and what we show in place of a score.
Make your relevant experience easy for people and software to find, using clear sections and specific statements you can support. Neither matching keywords nor using AI guarantees an interview.
What recruiters said when asked
Enhancv, a resume-builder company, surveyed 25 recruiters in the United States in September and October 2025 about how their applicant tracking systems handle incoming resumes. Three findings matter here.
No automatic rejection for formatting, content or design
Content-based automatic rejection configured
Knockout questions on the application form
Source: Enhancv, “Does ATS reject resumes?”, published on enhancv.com. Twenty-five recruiters is a small sample, chosen by a company that sells resume software, so treat the percentages as indicative rather than precise. It is, as far as we could find, the only recent survey that asked recruiters this question directly and published the counts.
Employers use different systems and settings. Some parse and search resumes; some also use AI to compare qualifications with role-specific criteria. There is no universal ATS score we can reproduce. In the survey above, the rejection that recruiters described as automatic comes from the yes-or-no questions on the application form: do you have the right to work here, do you have the licence, will you relocate. Those cannot be fixed by formatting.
What an ATS actually does with your file
An applicant tracking system is a database for job applications. When you apply, it typically stores your file, runs a parser over it to pull out names, titles, companies, dates and skills into fields, and attaches your answers to the form questions. Recruiters then filter or search the list of applicants, and some systems also compare each application with criteria the employer sets. How a given employer has set this up is not visible from outside.
Two things follow. If the parser reads your file badly, your record has the wrong title in the title field and the wrong dates in the date field, and a search for “account executive” may not return you even though the words are on the page. And if your resume proves a requirement in different words from the posting, a search for the posting's words will not find it, even though a person reading the file would count it. Both are fixable with a clear file and the posting's own words where they are true.
What a parser needs from your file
None of this is about beating a robot. It is about making sure the stored record of you is accurate, so the recruiter's search finds you and the file opens cleanly.
One column, top to bottom
Parsers read in document order, and some read two columns in the wrong order, mixing your dates with your skills.
Standard headings: Experience, Education, Skills
Parsers find sections by their headings. An unusual heading such as "Where I have made a difference" may not be recognised as your experience.
Each role on its own lines: company, title, dates
Those three fields are what gets stored and searched. Keep them in the same order for every role and use one date format.
Contact details in the body, not the header or footer
Some parsers skip header and footer regions, so an email there can be dropped.
No tables, text boxes or images holding facts
Text inside a table cell or a text box is often read out of order or not at all. A logo is fine; a skill list drawn as a chart is not.
A Word file or a text-based PDF
A PDF exported from a design tool can be an image of text. If you cannot select the words in the PDF, most parsers cannot read them either. Selecting the text is a quick check, not proof of how every system reads your file.
Dates with month and year
Parsers compute tenure from dates. "2023 to present" is fine; "a couple of years" is not.
The test takes a minute. Copy the whole resume and paste it into a plain text editor. If the roles come out in order, each with its company, title and dates on readable lines, the parser will see the same thing. That plain-text view is what Eachrole shows as “what a parser sees” beside every tailored resume, and the Word and PDF files it exports follow the rules above.
What an ATS score measures
A score from a resume tool is produced by that tool. It counts how many of the terms it extracted from the posting appear in your resume, sometimes weighted, and adds points or deductions for formatting it can detect. That is a keyword-overlap estimate. It is not the employer's system, which the tool has never seen. It does not know which terms the recruiter will actually search for, and it cannot see the knockout questions that do the rejecting.
The number is useful for one thing: noticing that a posting uses a term for something you did and your resume uses a different term. It is not useful as a prediction of whether you will be read, and a tool that charges to raise it is charging to raise its own number. That is why Eachrole does not show one.
What we show instead
The posting's requirements, as a list. Each one is marked supported, with the sentence from your resume that supports it and where it is; needs your answer, with a plain-words question about what you did; or not shown, meaning your resume says nothing about it and the tailored version will not either. You can read that list and decide for yourself whether to apply. A number cannot tell you that, and it should not tell you to add a line that is not true.
- See how that looks for a real posting in the examples by role.
- Read the method for tailoring a resume by hand.
Questions people ask
Will an ATS throw out my resume for using two columns?
Rarely by itself, according to the recruiters surveyed. It may parse badly, which means the recruiter's search does not find you and the stored record has your dates in the wrong fields. Single column avoids the problem.
Do I need to stuff keywords?
No. Use the posting's terms once, where they are true, in the bullet that proves them. Repeating a term five times does nothing for a search that only needs one match, and it reads badly to the person who opens the file.
Is an ATS score from a resume tool accurate?
It is a measurement of that tool's own keyword list against your text. It is not the employer's system, it does not know the recruiter's search terms, and it cannot see the knockout questions. Treat it as a rough keyword overlap, not a verdict.
What does the 75% rejection figure come from?
It is repeated across the industry, usually without a primary source. We could not find one. The recruiter survey above, small as it is, points the other way.