What Your ATS Score Actually Means

A score is a diagnostic, not a verdict. Its only real job is telling you which part to fix first.

From the team behind PassATS. Free to read, no email required.

The bands

Below 45. Something structural is wrong, not cosmetic. Usually a missing Skills section, a layout the parser cannot read in order, or text that is not really text. Fix the structure before touching wording.

45 to 60. Readable, but it is not competing. Typically thin keyword overlap and bullets that describe duties rather than outcomes. This is the band where the most improvement is available per hour spent.

61 to 75. A solid resume that loses to better-matched ones. The work here is targeting: mirroring the specific posting rather than sending one generic document everywhere.

76 to 85. Strong. Remaining gains are narrow and role-specific. Worth tailoring per application rather than rewriting.

Above 85. The resume is not what is holding you back. Look at where you are applying and who is referring you.

How the four parts are weighted

A composite score hides which part is dragging. PassATS weights ours like this, and most tools are directionally similar:

Keywords, 35%. The heaviest single component, and the most fixable in an afternoon.

Formatting, 30%. Whether the structure survives extraction: standard headings, readable order, consistent dates.

Contact information, 20%. Present, parseable, and not stranded in a header region some parsers skip.

Readability, 15%. Sentence length, bullet density, whether a human can scan it in twenty seconds.

The practical consequence: a 68 driven by a 40 on keywords is a different job from a 68 driven by a 40 on formatting. Always look at the components, never just the total.

What a score cannot tell you

It cannot tell you whether you are qualified. It measures how legible your case is, not how strong it is.

It cannot predict a specific employer. Systems differ, configurations differ, and the same resume scores differently across them.

It cannot see your rendered page. Every checker reads extracted text, so none can reliably tell you whether your two columns interleave or whether something is hidden. Open your PDF and try to select a sentence. That ten-second test catches what software cannot.