How ATS Systems Actually Work, and What That Means for Your Resume

An applicant tracking system is a database. It stores applications, lets recruiters search and filter them, tracks candidates through stages, and handles the communication.

For a related Monitask resource on cognitive offloading, see this guide.

That is most of what it does. A large industry of advice has grown up around the idea that it is an automated gatekeeper silently rejecting most resumes, and that idea is mostly wrong — but there are real formatting problems worth avoiding.

What actually happens to your application

You apply, usually through a form that also asks you to attach a resume
The system parses the resume, attempting to extract structured fields — name, contact details, employers, dates, titles, education, skills
The parsed data and the original file are stored
A recruiter searches, filters, or sorts the pool
A human reads what surfaces

The critical point is step five. In most organisations, a person reads the applications. Where automated screening exists, it is usually limited to knockout questions the employer configured — work authorisation, licences, minimum experience — not an algorithm judging your resume's quality.

The myth worth dropping

You will frequently see claims that a specific large percentage of resumes are "rejected by the ATS before a human sees them." That figure circulates without a credible source and does not describe how these systems are typically configured.

What actually happens is more mundane and more fixable: your resume parses badly, so the recruiter's search does not find you, or your record looks incomplete. That is not rejection by an algorithm. It is failure to appear in a search — which has the same practical effect and a completely different fix.

Formatting that genuinely causes problems

These are worth taking seriously.

Text in headers and footers. Many parsers ignore this region. Putting your phone number and email there is the single most common way to end up in a database with no contact details.

Multi-column layouts. Parsers read in a linear order and frequently interleave columns, producing scrambled output. A two-column resume with a sidebar is the most common source of garbled records.

Tables. Same problem. Content inside table cells often parses out of order or not at all.

Text inside images or graphics. Not extracted at all. A skills section rendered as a graphic is invisible.

Unusual fonts and heavy graphic design. Decorative fonts can produce character substitution errors.

Non-standard section headings. "Where I've Been" instead of "Experience" means a parser looking for a work history section may not find one. Be conventional here; save personality for the content.

Dates in unusual formats. Use month and year — "March 2023 – Present." Formats like "'23–now" parse poorly.

File type. A text-based PDF is generally safe. Some older systems handle .docx more reliably. If the application form specifies a format, use it. Never submit a scanned image or a design-tool export that contains no selectable text — if you cannot select the text, neither can the parser.

What to do instead

Single column, top to bottom
Contact details in the body of the document, not the header
Conventional section headings: Summary, Experience, Education, Skills
Standard fonts
Month and year for all dates
Job title, employer, dates, location on separate clear lines
No tables, text boxes, or images containing text
A text-based PDF unless told otherwise

Then check it: open the PDF, select all, copy, and paste into a plain text editor. What you see is roughly what the parser sees. If it is scrambled or missing sections, fix the layout.

Keywords, honestly

Recruiters do search the database, so the words on your resume matter. But the useful version of this advice is narrower than the usual version.

Do: use the same terminology as the job posting where it genuinely describes what you have done. If they say "accounts payable" and you wrote "AP," write both once.

Do: include tools, systems, certifications, and methodologies by name.

Do: spell out acronyms once alongside the abbreviation.

Do not: stuff keywords, hide white text, or paste the job description in a tiny font. Parsers extract it, recruiters see it, and it reads as dishonest.

Do not: claim skills you do not have to match a list. It surfaces in the first interview.

What actually gets you screened out

Far more often than formatting:

Applying without meeting a genuine hard requirement — work authorisation, a required licence, a location the role cannot accommodate
Answering a knockout question in a disqualifying way
No evidence of the core requirement anywhere on the resume
Applying to twenty roles at one company, which is visible in the system and reads as unfocused
A resume that does not make the last two roles obvious in the first ten seconds — because a human is skimming, and that is the real filter

One resume or many

Maintain one strong base resume and adjust it per application — the summary line, the ordering of bullets, the terminology. That takes ten minutes and materially improves relevance.

Rewriting from scratch for every application is not a good use of your time, and applying with an unmodified generic resume to fifty roles is worse than applying carefully to ten.

The proportionate view

Format your resume so it parses cleanly. That is a one-time fix taking about half an hour.

Then stop thinking about the ATS and spend the time on what actually decides outcomes: whether your resume makes a human want to talk to you, and whether you are applying to roles you have a real case for. See resume structure: what to keep and what to cut.

Looking for a role? Browse open positions → For broader reference, consult CareerOneStop.