Send a PDF unless the job posting explicitly asks for a Word document. PDF locks your layout on every device, and every modern ATS parses text-based PDFs reliably. Send DOCX when instructions demand it — usually agencies that edit resumes before forwarding. The real danger is not the format; it is exporting a PDF whose text is actually an image.
A DOCX renders differently depending on the fonts installed, the Word version, and whether it is opened in Google Docs, Pages, or LibreOffice. Your carefully aligned dates drift, page breaks move, and the hiring manager sees a layout you never approved. A PDF is a frozen snapshot: what you exported is what every reader sees, on every machine, forever.
The old advice to send Word because ATS cannot read PDF describes software from a previous decade. Workday, Greenhouse, Lever, iCIMS, Ashby, SmartRecruiters — all parse text-based PDFs as a matter of course, and PDF is the majority format they receive. If a parser mangles your PDF today, the cause is almost always layout complexity or image-based text, and the same layout would have failed as DOCX too.
Staffing and recruiting agencies often require DOCX because they reformat resumes — stripping your contact details, adding their branding — before presenting you to clients. Some government and enterprise application portals also hard-require it, and a few old internal systems genuinely parse DOCX more cleanly. When the instructions say Word, send Word; following submission instructions is itself part of the evaluation.
Keep one master copy in an editable format regardless of what you send. The clean workflow is: maintain the source in Word, Google Docs, or a resume builder, export PDF for applications, and export DOCX on demand. Never maintain two diverging masters — that is how the PDF you send tonight ends up missing the bullet you added last week.
The fatal PDF error is text that is not text. Print-to-PDF from a design tool, a scanned page, or a Canva export at the wrong setting can produce a file where your entire resume is one flat image — visually perfect, machine-empty. Test any resume file in seconds: open it and try to select and copy a sentence. If you cannot select the text, a parser reads nothing at all.
For DOCX, the classic failures are content placed in text boxes and headers, which many parsers skip entirely, and fonts the recipient does not have, which reflow the whole document. Keep contact details in the body, not the header, use standard fonts, and your file survives both the parser and the reformat an agency will do to it.
Yes. Every major modern ATS parses text-based PDFs reliably; PDF is the most common format they ingest. Failures come from image-based text or complex layouts, not from the PDF format itself.
Open the file and try to select and copy the text. If you can select sentences and paste them elsewhere intact, a parser can read them. If the page behaves like a photo, your text is an image and an ATS extracts nothing.
Usually because they edit it — agencies strip contact details and apply their own branding before sending candidates to clients. Some legacy portals also require DOCX. When explicitly asked, send it; otherwise default to PDF.
Upload your PDF to the free scorer — it parses the file exactly the way an ATS does and shows you what machines actually extract.
