Exactly How Our Free ATS Checker Scores Your Resume

Most ATS checkers hand you a number — 68, 74, 81 — and a paywall between you and the explanation. We think that's backwards. A score you can't audit is a score you shouldn't trust, so this article walks through our checker's entire scoring logic. Not a simplified version. The actual algorithm.

The headline score is just arithmetic you can check

Your score is a weighted sum of five categories. These are the real weights, straight from the code:

Multiply each bar you see by its weight, add them up, and you get the headline number — always. We enforce this with an automated test in our build: if the score ever disagrees with its own breakdown, our code won't ship. (Earlier versions of our checker had exactly that bug — a gauge that said 59 next to bars averaging 85. We fixed it and then wrote the test so it can't come back.)

Parse safety: the six checks that reflect real failures

The formatting category is driven by the failure modes that actually scramble resumes in applicant tracking systems — not superstitions about fonts:

• Multi-column or table residue — column layouts get linearized left-to-right and interleave your content into nonsense.
• Missing standard headings — parsers file your content under "Experience", "Education", "Skills". Creative heading names break that mapping.
• Contact details outside the body — several parsers skip document headers and footers entirely.
• Exotic bullet glyphs — decorative characters that extraction can garble.
• Garbled encoding — replacement characters that signal text stored as an image or a broken export.
• Inconsistent date formats — mixing "Jan 2022" with "01/2022" degrades how systems parse your work history timeline.

Each finding is shown to you with its severity — nothing is silently folded into the number.

Keyword matching that knows "K8s" means "Kubernetes"

Paste a job description and we compare its meaningful terms against your resume. Three details matter more than they sound:

• Noise words are filtered out. "Responsibilities", "candidate", "excellent" — matching those would inflate everyone's score, so they're ignored.
• Variants count. Our synonym map treats JS/JavaScript, K8s/Kubernetes, SEO/search engine optimization, AWS/Amazon Web Services as the same skill — the way modern recruiting software normalizes them.
• Your title match is scored separately. Whether the job's actual title appears in your resume is one of the strongest ranking signals in real systems, so we surface it on its own instead of burying it.

What runs where — and what we don't do

Everything above runs in your browser. Your resume is not uploaded to our servers for the free analysis, which is also why we can offer unlimited scans without an account: it costs us nothing, so we charge you nothing. If you sign in (free), an optional AI pass adds context-aware suggestions on top — but the deterministic score is the same either way.

Things we deliberately don't do:

• No fake precision. The same resume scores up to 24 points apart across real ATS platforms, so we don't claim to predict your exact score in Workday. We measure the things that are true everywhere: parseability, findability, and writing strength.
• No invented benchmarks. Our comparison chart is labeled "targets for a strong resume" because that's what it is — not fabricated "industry averages."
• No fear pricing. The score, the findings, and the fixes are visible without paying or handing over your email.

Audit us

Run a scan, then check our math: weight the five bars and compare against the headline. If they ever disagree, that's a bug — and we'd genuinely like to hear about it. If you want to fix the file before interpreting the score, start with the ATS-friendly resume format checklist.