ATS Resume Screening: What Actually Gets You Filtered Out

The most common belief about Applicant Tracking Systems — that they secretly reject well-qualified resumes for arbitrary formatting reasons — is mostly wrong, but it's wrong in a way that hides the real, more boring issues that actually cause problems.

The myth

Modern ATS platforms (Greenhouse, Lever, Workday, and similar) don't run a hidden scoring algorithm that auto-rejects resumes before a human ever sees them, for the vast majority of companies using them. What they actually do is store your resume, extract structured fields from it (name, contact info, work history), and make the whole thing searchable and filterable for recruiters. The "rejection" people attribute to a mysterious algorithm is almost always a recruiter filtering by a keyword or requirement and your resume not matching it.

What actually causes problems

Parsing failure is real. If your resume uses a complex multi-column layout, text inside images, tables, or unusual fonts, the extraction step can genuinely mangle your work history or skip sections entirely — not because of a scoring penalty, but because the parser simply couldn't read it correctly. A single-column, standard-font PDF avoids this entirely.

Missing the actual keywords the job posting uses is the second real issue. If a posting asks for "PostgreSQL" and your resume says "SQL databases," a recruiter searching or filtering by the exact term won't find you, even though you're qualified. This isn't an ATS flaw — it's the same problem as not matching a search query, applied to hiring.

Generic, non-specific bullet points are the third. "Responsible for backend development" tells a recruiter nothing they can filter or judge on. "Built and maintained backend services processing 2M+ transactions/month" gives them something concrete, and happens to also contain more of the specific terms a search would match on.

The practical fix

Keep the format simple — single column, standard fonts, no text embedded in images. Read the actual job posting and use its specific terms where they're honestly true of your experience, rather than a rewording. And favor specific, quantified bullet points over generic responsibility statements. None of this requires guessing at a hidden algorithm — it's just making sure the real, human-facing search and review process actually works in your favor.