What an ATS actually does with your CV

If you've looked for a job in the last few years, you've probably heard that applicant tracking systems reject most CVs before a person ever reads them. The number that gets passed around is 75%.

We spend a lot of time inside these systems. To build IdealJob we read application forms from Greenhouse, Lever, Ashby, Workable, SmartRecruiters and dozens of others every day. What happens to your CV is a lot less dramatic than the internet suggests, and a lot more useful once you understand it.

Where the 75% comes from

As far as anyone can tell, the figure traces back to a company that sold CV optimization services. Nobody has ever published the study behind it. It spread because it's scary, and because it sells CV checkers.

That doesn't mean the ATS is harmless. It means the risks are specific, and you can deal with each of them.

What happens when you click apply

At most companies it goes like this.

1. Your CV is stored and parsed. The system tries to pull your name, contact details, jobs, dates and education into separate fields.
2. Your answers to the application questions are saved next to it.
3. The application joins a list for that role, usually sorted by date.
4. A recruiter or hiring manager works through the list, looking at the parsed profile and often the original PDF.

For most roles, nothing in that sequence reads your CV and decides you aren't good enough. A person does that.

The three things that can actually knock you out

Knockout questions. This is the one real automatic filter. "Do you have the right to work in Germany?" "Do you hold a driving licence?" Answer no to a question the company marked as required and many systems move you straight to rejected. Read those questions slowly and answer them honestly.

Parsing problems. If the system can't read your CV, the recruiter sees an empty or scrambled profile. The usual causes are two column layouts, text inside images, contact details tucked into the page header, and creative section names like "Where I've been" instead of "Experience". A clean single column PDF with standard headings avoids almost all of it.

Search. Recruiters with hundreds of applicants often search the pool for a tool or a title. If the posting says Kubernetes and your CV only says "container orchestration", you won't show up. Name your tools plainly.

What matters less than people say

Hidden keywords in white text. Recruiters look at the parsed text, where hidden keywords sit in plain view and look desperate.

Keyword counts. There's no score that goes up each time you repeat "Python". A person reads the CV, and repetition reads badly.

File names, fonts, one page or two. Use a sensible file name and a readable font. Nobody is being filtered out for using Calibri.

So what should you do?

Use a simple single column CV with standard section headings. Put your contact details in the body of the page, not the header. Name your tools the way the posting does. Answer knockout questions carefully.

Then spend the time you saved applying to roles that actually fit you, because that's what really moves the numbers.

Every template in the IdealJob resume builder is built to parse cleanly. The default one, Modern No Photo, is the one we'd send to an ATS ourselves.