Resume PDF vs DOCX: What Actually Passes the ATS
Send a text-layer PDF exported from a word processor, unless the posting tells you otherwise. That's the answer to the resume PDF or Word question, and the rest of this article is why.
The debate itself is a distraction the ATS benefits from. While candidates argue file extensions in forums, the parser quietly rejects files in both formats for the same underlying reason: it can't read what's inside them. The extension was never the problem. The construction of the file is.
Key Takeaways
- Modern ATS parsers handle both PDF and DOCX reliably when the file comes from a standard word processor.
- PDFs fail when they have no text layer: design-tool exports, scans, and image-based files read as blank pages to the parser.
- DOCX files fail through text boxes, content parked in headers and footers, and layouts that reflow when opened elsewhere.
- When a posting asks for a specific format, that instruction is real. Follow it every time.
- Test your file in 60 seconds: select all, copy, paste into a plain text editor, and read what the machine reads.
What the parser does with your file
An ATS doesn't open your resume the way a human does. It runs the file through a parser that strips away all visual formatting and extracts raw text, then tries to sort that text into structured fields: name, contact, work history, skills, education. Your carefully kerned layout dies in the first step. Only the text survives.
Both PDF and DOCX carry text the parser can extract, when the file is built normally. Export a PDF from Word or Google Docs and the text layer comes along intact. Save a DOCX with a single-column layout and standard headers, and the parser walks through it cleanly. The 75%+ filter-out rate that Harvard Business School's Hidden Workers research documented isn't driven by file extensions. It's driven by everything the parser fails to read or fails to match.
So the real question isn't PDF versus Word. It's whether your specific file survives extraction. Each format has its own way of dying.
When PDF fails
A PDF fails when there's no text layer underneath the pixels. The resume looks perfect on your screen and reads as a blank page to the machine. Three common ways this happens:
Design-tool exports. Canva, Photoshop, Illustrator, and Figma resumes often export text as vector shapes or flattened images. Gorgeous file, zero extractable text. This is the single most common catastrophic format failure, and the people it hits hardest are the ones who invested the most effort in looking good.
Scans and photos. A printed resume scanned back to PDF is an image, full stop. No parser is running reliable OCR on your application. It reads nothing.
Flattened or secured files. Some PDF tools flatten text during compression or "print to PDF" workflows, and password-protected or copy-restricted PDFs can block extraction outright. If the file forbids copying text, it forbids the parser too.
The fix is boring: write the resume in a word processor and use its own export-to-PDF function. That pipeline preserves the text layer every time.
When DOCX fails
Word files fail differently. The text layer is there, but the structure misleads the parser.
Text boxes. Anything living in a floating text box sits outside the document's main text flow. Parsers routinely skip it entirely. If your contact info or a skills sidebar is in a text box, it may not exist as far as the ATS is concerned.
Headers and footers. Same problem. Plenty of templates park your name, email, and phone number in the document header. Plenty of parsers never look there. You get scored as a candidate with no contact information.
Reflow and version drift. A DOCX is a living document. Open it in a different Word version, Google Docs, or LibreOffice and the layout can shift: page breaks move, spacing collapses, fonts substitute. The file you sent is not always the file that gets opened. A PDF is frozen. A DOCX is a suggestion.
These are the same structural traps we covered in what the ATS actually flags, and they matter more than the extension ever will. The resume formatting guide covers the full layout rules if you're rebuilding from scratch.
What "submit as a Word document" instructions mean
Sometimes the posting tells you the format. That instruction is not decoration.
When a company asks for DOCX, it usually means one of two things: their intake system is older and parses Word most reliably, or a human plans to edit the file. Staffing agencies do this constantly, since they reformat resumes and strip contact details before forwarding candidates to clients. HR practice surveys from SHRM have long documented how varied these intake systems are across companies. You don't get to know which system you're facing, so you follow the instruction.
When a posting asks for PDF, same rule. The instruction always beats the general advice, including ours. Ignoring a stated format requirement is a free way to look like someone who doesn't read carefully, and that judgment happens before anyone weighs your qualifications.
The rule, and the 60-second test
The practical rule: text-layer PDF from a word processor, unless the posting says otherwise. It's frozen, it renders identically everywhere, and it parses cleanly. Keep a DOCX version of the same resume on hand for the postings that ask.
Then test the file before you trust it:
- Open your resume file.
- Select all, copy, and paste into a plain text editor like Notepad or TextEdit.
- Read the paste. That's roughly what the parser sees.
If the paste is empty, your PDF has no text layer. If your name and email are missing, they're trapped in a header or text box. If sections arrive scrambled or interleaved, your layout is confusing the read order. Fix the source and test again. Sixty seconds, and you've verified the one thing the format debate was ever about. The rest of passing the ATS is content and keywords, not file extensions.
Stop arguing extensions with a parser
The diagnosis: the ATS was never rejecting your file type. It was rejecting files it couldn't read, in both formats, while candidates argued about the wrong variable. The fix: a text-layer PDF, a DOCX backup, and a 60-second paste test.
Gate Crashers builds your resume parser-clean from the start and hands you three tailored versions in both formats, ready to send. One session, $4.99, no subscription. Send the file the machine can read, and make the argument about your experience instead.