Back to Articles

Why Resume Parsing Fails (And How to Test Your Resume Before You Apply)

If your resume upload looks fine but the application autofill is wrong, you don’t have a “keyword” problem—you have a parsing problem. Here’s a fast, repeatable QA process to catch it before you apply.

Last updated Apr 2026
Quick answer

Run 2 tests: paste → portal autofill. If either looks wrong, fix formatting before you tailor keywords.

Your goal is simple: make sure your resume becomes clean, correctly ordered text after an ATS strips away layout. If text is missing or sections reorder, your application fields (and recruiter search) can be wrong.

What this catches: missing contact info, scrambled work history, dropped skills, and “resume parse failed” scenarios that happen due to file size, images, and complex layout (tables/columns/headers/footers/text boxes).

What “parsing” actually means

  1. The ATS converts your file into plain text.
  2. Then it tries to detect entities like name, email, job titles, companies, and dates.
  3. Finally, many systems auto-fill your application profile with what they extracted.

That’s why a resume can look perfect to you but still “break” in an ATS: the parser reads your file differently than a human does.

Test #1: The paste test (plain text)

How to do it
  • Copy all resume text (or save as .txt).
  • Paste into Notepad (Windows) or a plain-text editor.
  • Read top-to-bottom like a robot would.
Pass / fail
  • Pass: correct order, headings appear, bullets readable.
  • Fail: missing text, weird symbols, dates jump around, right column appears before left.
If the paste test fails: don’t “keyword optimize” yet. Fix layout first. Keyword work only helps after the parser can read your resume reliably.

Test #2: Autofill sanity check (portal)

This is the most realistic test because it shows what the employer’s system extracted. After you upload your resume, many portals take you to auto-filled fields. Treat those fields as part of your application—not optional.

Check these fields
  • Latest job title + employer
  • Start/end dates for each role
  • Education (degree + school)
  • Location formatting
If it’s wrong
  • Fix the fields if the portal allows edits.
  • Also treat it as a signal: simplify your resume and re-upload.
  • Try the other file format (PDF ↔ DOCX).

If it fails: the fixes that work

  • Remove columns/sidebars: use a single-column layout.
  • Remove tables, text boxes, headers/footers: keep everything in the main document flow.
  • Replace icons/graphics with text: parsers can ignore images.
  • Standard headings: use “Summary”, “Experience”, “Education”, “Skills”.
  • Normalize dates: pick one format (e.g. Jan 2023 – Mar 2025).
  • Keep file size reasonable: some systems can’t parse large files (for example, Greenhouse can’t parse above 2.5MB).

Pre-submit checklist

Copy/paste this

  • Paste test passed (no missing text, no scrambled order).
  • Portal autofill fields look correct (titles, companies, dates, education).
  • One column, no tables/text boxes, no header/footer contact info.
  • Skills and keywords are plain text (no icon-only bullets or skill bars).
  • File format tested (PDF and DOCX both available; upload the one that parses best).
  • File size is reasonable (for Greenhouse parsing, stay under 2.5MB).

Sources

  • Greenhouse Support: “Unsuccessful resume parse” (examples of formatting that breaks parsing; 2.5MB parse limit). View source
  • MIT CAPD: “Make your resume ATS-friendly” (avoid tables/text boxes/graphics; test by saving as plain text and checking missing/out-of-order content). View source