How ATS Scoring Actually Works (And What It Doesn't Do)

"The ATS" gets talked about like a single gatekeeper with a secret rulebook. In reality, it's a category of software -- Workday, Greenhouse, Lever, iCIMS, Taleo, and dozens of others -- and they don't all work the same way. Here's what's actually true across most of them, and what's mostly myth.

What an ATS actually does

At its core, an ATS stores resumes, extracts text into searchable fields (name, contact info, work history, education), and lets recruiters search, filter, and sort applicants. Some systems layer on a match score against the job requisition's listed skills and keywords. Very few systems auto-reject a candidate outright -- more often, a low match score just means a resume sorts lower in a recruiter's queue rather than disappearing.

What it does check for

Consistently, across platforms: whether your work history parses into distinct, dated entries; whether contact information is detected correctly; and whether key terms from the job posting appear somewhere in your resume. This is why standard section headings and a single-column layout matter -- not because of some hidden penalty, but because that structure is what parsers are built to read reliably.

What it doesn't do

Most modern ATS platforms don't use OCR tricks that get fooled by "invisible" white-on-white keyword text -- and if a human recruiter does open the file, stuffed or hidden keywords look exactly as bad as they sound. There's also no universal blacklist for employment gaps, career changes, or non-traditional backgrounds; that kind of judgment happens with a human reading your resume, not the software that stores it.

Different companies configure this differently

A company using Greenhouse with no automated scoring enabled behaves nothing like a large enterprise running a fully configured Taleo instance with hard keyword filters. There's no single set of rules that works identically everywhere, which is part of why "ATS-friendly" advice should focus on structure and clarity rather than trying to reverse-engineer one specific system.

What's actually worth doing

Use standard section headings, keep critical information in plain text rather than images or text boxes, and genuinely reflect the job posting's language where it's accurate to your background. Beyond that, write for the human who eventually reads it -- because on the vast majority of applications, one eventually does.