ATS Methodology: What Our Resume Checks Measure

Updated 2026-10-07 · Nodesify

What does “ATS-friendly” mean here?

It means a focus on readable document structure and selectable text. It does not mean that an ATS vendor has certified the template, that every field will import correctly or that an employer will advance an application.

Three different questions

• Document extraction: does the final file preserve words, characters and a useful reading order?
• Content completeness: does the pasted resume contain the elements our checklist looks for?
• Employer screening: does the application meet that employer’s criteria? Our tools do not reproduce this decision.

PDF extraction method

Our engineering comparison approach renders known sample content to PDF, extracts text with pdftotext and compares it with the source. Checks cover word integrity, section headings, contact information, key terms and selected Unicode and page-break cases. Visual inspection is still needed for layout.

This page describes the method, and below is the one benchmark result we have published so far. No current vendor pass rate or universal compatibility result is claimed. Extracted text is a proxy for document readability, not a reproduction of a commercial ATS. A sample set cannot cover every input, language or employer configuration.

What our July 2026 benchmark measured

In July 2026 we rendered three synthetic fixtures (plain-text content, Unicode names and symbols, and a multi-page document) through all five templates — 15 document pairs — comparing two PDF engines used during our engine selection, Chromium print-to-PDF (Chrome for Testing 149.0.7827.55) and WeasyPrint 63.0. Text was extracted with pdftotext and scored against the source. Chromium print-to-PDF is the engine our paid exports use today.

Read these numbers with their limits: the fixtures are synthetic, so they exercise known failure modes rather than the full variety of real resumes; the result is dated to specific engine versions and has not been re-run against every later release; and pdftotext is a proxy, not an ATS vendor. These are averages; some individual pairs scored lower. The benchmark compares our own engine options — it is not an ATS acceptance rate, a certification, or a promise about any employer’s software.

These metric averages were corrected on 7 October 2026 against the retained raw report. The earlier narrative mislabeled an overall score as word integrity. Inspect the original results, PDFs, extracted text and scoring source for the calculation and limitations.

How the free checker calculates its score

The readiness score uses three essential checks: full name (weight 10), email format (weight 10), and relevant visible experience (weight 12). Student and early-career users can satisfy the evidence check with a named project that includes a description or bullets. The score is the sum of passed essential weights divided by 32, rounded to a percentage. These weights are product choices, not measured hiring probabilities.

Summary length, phone, location, portfolio links, education, skill count, action verbs and numerical achievements are optional writing advice. They do not lower the readiness score. Hidden sections are excluded from writing advice and matching; hidden content cannot satisfy the relevant-evidence check. A score of 80 or more receives “Excellent”; 55–79 is “Good”; lower scores are “Needs Improvement.” These labels describe this checklist only.

Keyword matching

The free comparison selects up to 30 terms, prioritizing explicit requirements before frequency. It recognizes short technical skills and uses token boundaries to preserve terms such as C++ and C#. Ambiguous names such as Go, R and C require technical context. Each match shows an excerpt from a visible resume section. Missing terms indicate wording we did not find, not a proven lack of ability. Add terms only when your experience supports them.

Smart Match additionally uses phrases, light plural folding and curated synonym groups. Both tools use deterministic rules, not an understanding of eligibility or employer decisions. Requirement labels are heuristic, and languages other than English may produce incomplete results.

What pasted text cannot tell us

Pasting removes the original file’s layout. The checker cannot inspect its fonts, embedded images, columns, page breaks or actual PDF extraction order. Its parser can also assign content to the wrong field. Review the suggestions and the original file; do not add inaccurate claims to improve a score.

Check your own export

1. Open the final PDF and check page breaks and links.
2. Select and copy its text into a plain-text editor.
3. Check names, symbols, section boundaries, dates and reading order.
4. Review imported fields after upload and follow the employer’s instructions.

For an example of vendor-specific limitations, see Greenhouse’s parsing documentation. A failed parse and an employer’s rejection are different events.

Use the free checker, read the formatting guide, or send a correction through our editorial contact.