Skip to content
Second·Reader
← Writing

Does an applicant tracking system actually reject your resume?

You've been applying for weeks. Maybe months. You've read that a robot reads your resume, scores it, and throws most applications in the bin before a human.

September 17, 2026

You've been applying for weeks. Maybe months. You've read that a robot reads your resume, scores it, and throws most applications in the bin before a human ever sees them. So you've reformatted, you've stripped out the columns, you've matched the font, and still nothing comes back. Here's the thing worth knowing before you spend another weekend on formatting: the software probably isn't rejecting you the way you've been told it is.

The claim you've heard, and what recruiters actually say

The story goes that an applicant tracking system scans your resume and auto-rejects it on content, layout, or design. Recruiters who use these systems every day mostly say that's not how it works.

In a late-2025 study by Enhancv, 23 of 25 recruiters (92%) said their systems do not auto-reject resumes for formatting, content, or design, and only 2 of 25 (8%) had any content-based auto-rejection turned on at all (Enhancv). Worth the honesty here: that's a small, self-selected sample of 25 people run by a company that sells resume tools, so treat it as directional, not gospel. It lines up with what recruiters say elsewhere, but it isn't a large representative survey.

The one thing that really is automatic

There's a genuine automatic cut, and it isn't the parser. It's the knockout question.

These are the yes/no gates an employer sets on a specific posting: work authorization, a required license, location, a hard floor on years of experience. Answer one the wrong way and your application can route straight to rejected before a recruiter ever opens it. In the same Enhancv data, knockout questions were in use by every recruiter with the feature, while content auto-reject sat at 8% (Enhancv). This has nothing to do with your layout. If you're getting cut this way, the fix is reading the eligibility questions carefully and only applying where you actually clear them.

A failed parse is not a rejection

You may have heard that if the software can't read your resume, it deletes you. It doesn't.

Greenhouse's own support documentation says that when a resume can't be parsed, the system still creates the candidate record and attaches your original file. A recruiter then fills in the missing fields by hand (Greenhouse). You lose clean searchability, not your place in line. The one real risk is a resume saved as an image losing your contact details, so a recruiter can't reach you, which is one of the parse failures Greenhouse documents (Greenhouse on why resumes don't import). That's "unreachable," not "rejected," and it's an easy thing to avoid by keeping your resume as readable text.

What's actually happening to you

Here's the part that stings, because it feels exactly like rejection and it isn't. You're probably buried, not binned.

Recruiters describe the same mechanic over and over. The system parses applications and ranks them. The recruiter reads from the top of that pile, builds a shortlist, and often fills the role before reaching the bottom. Fifteen-year veterans on Reddit put it plainly: the software parses and sorts, a human works down from the top, and in most companies a person still has to click reject (JobSearchMethods thread, ECE thread). If you were applicant #500, or your resume didn't contain the words the recruiter searched for, you never surfaced. Silence follows. It reads identical to a no.

That matters because the fix for "never surfaced" is completely different from the fix for "badly formatted."

The 75% number, and why you should stop trusting it

The figure that started all this panic is that "75% of resumes are rejected by the ATS." When people traced it, it dead-ended at Preptel, a resume-optimization vendor that used a version of the number in marketing around 2012 and shut down in 2013 without ever publishing a dataset or method (UnchartedCareer trace). The number drifts between 70%, 75%, and 88% depending on who's repeating it, which is the tell that there was never one study behind it. It's vendor marketing. There's no primary source, and that's the whole point.

The number people confuse it with is real. In Harvard Business School and Accenture's 2021 "Hidden Workers" report, 88% of employers agreed that qualified, higher-skilled candidates get screened out because they don't match a job description's exact criteria (HBS/Accenture). Read that carefully. It's not a parser deleting resumes. It's rigid filters that humans set, employment-gap cutoffs, hard degree requirements, exact-credential demands. Same report found 90%-plus of employers use a system to make a first cut (HBS/Accenture). The software is in the loop for almost everyone. It's a ranking tool configured by people, not a judge.

The fix that actually addresses the problem

If you're not being rejected on content, and you're not being deleted by a parser, then reformatting won't move the needle. The problem to solve is relevance and searchability.

Recruiters find candidates by running keyword and Boolean searches, both inside the ATS and in LinkedIn Recruiter. LinkedIn's own help docs confirm recruiters filter with AND, OR, and NOT operators across titles, skills, and keywords (LinkedIn). If your resume doesn't contain the real terms someone is searching for, you don't appear in the results. That's it. That's the mechanism.

So the work is matching honest, relevant language to the role. Use the actual titles and skills the posting names, where they're genuinely true of you. Clear the eligibility questions before you spend an hour applying. Keep the file readable so your contact details survive. None of this is a trick, and please don't try one. White-font keyword stuffing doesn't help, because the system strips formatting and a recruiter reading the record just sees the padding. There's no score to game, and Greenhouse states its own matching doesn't auto-advance or auto-reject anyone (applyvita summary).

This is tedious to do well for every application, which is where Second Reader comes in: it writes a tailored resume and cover letter from your real history, checks every claim, and fills the application for you to review and send. Your first package is free. It won't get you a job, and anyone who promises that is selling you the same thing Preptel was.

Sources

If this is where you are

Second Reader writes the application from your real history.

A tailored resume, a cover letter and a requirement-by-requirement matrix for one posting. Anything your history can’t support is removed, and nothing is invented to fill a gap. Your first package is free.