What an ATS is, and how it actually reads your resume
Applicant tracking systems have collected a lot of myths — from „they reject 75% of resumes” to „you have to stuff in keywords”. Here is what can be said about them without guessing.
Updated
What it actually is
An ATS (applicant tracking system) is recruitment software. It collects applications from various sources, breaks resumes down into data, orders candidates in a list and tracks what stage everyone is at. Most large organisations use one, and plenty of mid-sized ones do too.
What matters most is what an ATS is not. It is not a language model deciding whether you would be good at the job. It is first and foremost a database with a parser — and a parser is an ordinary program guessing which fragment of your file is a company name and which is a date.
How to tell a company uses one
The surest sign: an application form that fills its own fields after you upload a resume — name, job titles, employment periods. That is the parser, working in front of you.
Second sign: the form lives on a different domain than the company site. Third: applying requires an account in a system that looks identical across several different employers.
If you are emailing a resume to a ten-person company, no filter applies to you. Everything about format still helps, because a human is reading it — and clear documents are easier for people too.
Two separate steps that people conflate
This is the most important thing here. Between sending a resume and getting an interview, two separate things happen, and they break for completely different reasons.
- Extraction. The system pulls text out of the file and maps it onto fields. You lose here through format: two columns, a scan with no text layer, icons instead of labels, dates without years, unusual section names. The result is not a low score — it is a candidate with empty fields in the database.
- Search and ranking. A recruiter filters that database by the skills in the job posting. You lose here through content: none of the words the posting uses for what you actually do.
Why keyword stuffing does not work
The „paste the requirements in white text” advice has been circulating for years and is wrong twice over. First, text extraction does not know about colour — white text is as visible to the parser as any other, and to the recruiter who opens the file it looks exactly like what it is.
Second and more important: a list is not evidence. A skill that appears only under „Skills” is a claim. The same skill appearing inside a specific role, next to what you did with it, is a demonstration.
A filter will let both through. The person reading afterwards will let one through. That is why in our report a skill with no backing in your experience counts half as much as one that is evidenced — a binary „present / absent” simply lies.
What these systems do not do
- They do not judge your competence. They judge what can be extracted from a document. That is a different question entirely.
- They do not automatically reject „75% of resumes”. That figure circulates without a source. Systems usually do not delete applications — they order them, and a badly parsed resume lands at the bottom.
- They do not punish a missing exact word if you use a synonym the system knows. Not every system knows every synonym, and that is a real lottery — but it is not the same as „you must write it word for word”.
- They do not decide on the interview. They decide whether a person sees your resume whole or in pieces.
What follows in practice
- Send a PDF with a text layer, or a DOCX. Never a scan, never a file from a design tool where text is outlines.
- Keep one column and standard section names.
- Put contact details in the body, not in a page header or footer.
- Dates with month and year, one format throughout.
- Name your skills the way the posting names them — but only where you have something to back them up.
None of this is a trick. It is a document written so it can be read — by a program and by a person.