ATS myths: what applicant tracking systems actually do with your resume

What does an applicant tracking system actually do?

An ATS stores applications, attaches resumes to candidate records, moves candidates through stages, and sends email. That is the core of it.

Written from the position of building one: in ofper, an application is a row linking a candidate to a job, with a stage. Nothing in the system reads a resume and forms an opinion about it. There is no scoring model, because there is no AI anywhere in the product.

Other systems do more. Some parse resumes into structured fields, some offer AI matching that ranks candidates against a job description, some let employers search a candidate database. These are real features and they vary by product and by which plan the employer is paying for.

Do applicant tracking systems auto-reject resumes?

Some can, but not in the way the advice implies, and the mechanism is usually visible to you.

The common form is a knockout question: the employer asks something like whether you have the right to work in the country, or how many years of a specific experience you have, and sets an answer that disqualifies. You see these as questions in the application form. In ofper they are limited to yes/no and numeric questions, and a candidate who fails one is flagged as screened out rather than deleted, so a person can still look.

What is much rarer than the advice suggests is a system silently scoring your resume text and discarding it without anyone seeing it. That requires the employer to have bought and configured that feature deliberately.

Where does the 75 percent claim come from?

It does not have a traceable source. The figure — sometimes 70, sometimes 75, occasionally 90 percent — appears across career-advice content and resume-checker marketing without a citation to an underlying study, and the products that repeat it most often are the ones selling a fix for it.

That does not mean resumes are not rejected in volume. They are — by people, quickly, often in a few seconds each. The difference matters because it changes what you should do about it.

So what actually helps?

The practical advice is mostly unglamorous and mostly about being readable by a human in a hurry.

• Use a normal document. PDF or DOCX, text-based rather than a scan or an image. If text cannot be selected in your PDF, no system can extract it — ofper, for instance, does no OCR at all.
• Use the words the job description uses for the same things. Not to game a filter, but because the person skimming is looking for them.
• Put the important things in the first third. Whoever reads it is deciding fast.
• Answer the screening questions accurately. These are the only automated filter you are likely to meet, and they are visible.
• Skip heavy multi-column layouts, text inside images, and text in headers or footers, which extraction handles worst.

Are ATS resume scores real?

Tools that give your resume an ATS score are not connecting to any employer's applicant tracking system. They cannot — those are private, per-company systems with no public scoring interface.

What such a tool actually does is compare your resume text against a job description and apply its own heuristic. That can be a useful writing exercise. It is not a measurement of what any employer's system will do, and the number it produces does not exist anywhere in the employer's software.

ofper does not offer a resume score, and will not. There is no AI inference in the product, and a client-side keyword match dressed up as a score would be a worse version of a thing that was already misleading.

Questions, answered straight