PDF vs DOCX for an ATS

Summary

Format matters less than most advice suggests and more than most candidates check. The real question is not PDF or DOCX — it is whether the text layer inside either file comes out in a readable order.

The short answer

DOCX is the more predictable file, because it stores text as paragraphs inside a document tree. A parser walks that tree in order. PDF stores text as positioned fragments, so the reading order depends on how the file was generated.

That does not make PDF unsafe. A single-column PDF exported from Word or Google Docs with embedded fonts usually extracts cleanly. A scanned PDF, or a PDF full of text boxes and columns, does not — and no format change will fix a scan.

Why DOCX is usually the safer file

The caveat is that DOCX gives up visual fidelity. If the employer's portal accepts DOCX, that trade is usually worth it: the layout you lose is the layout a parser ignores anyway.

• —Structure survives: headings and bullet lists are still paragraphs, not coordinates on a page.
• —Headers and footers are stored separately, which is why you should never put contact details only there.
• —Tables are still tables. A DOCX built from a two-column layout table breaks the same way a PDF does.
• —Standard fonts are guaranteed to be present, because the document carries the font choice rather than the glyphs.

Where PDF breaks

• —A scanned or photographed resume has no text layer at all. Copy-paste returns nothing, and so does the parser.
• —Two-column templates and sidebars force the extractor to choose between reading across the page and reading down one column. Different tools choose differently.
• —Text boxes, floating shapes, and icon rows are positioned objects. Their text can be written to the file after the body, which puts it at the end of the extracted string.
• —Contact details in a designed header can be dropped before the body is read.
• —Subset or missing fonts can turn text into unreadable glyphs, especially in exports from design tools.

The thirty-second test

Open the file, select everything with the keyboard, copy, and paste into a plain text editor with no formatting. What you get back is close to what a parser gets.

• —If the order is scrambled, the parser sees the same mess — no format change will hide it.
• —If your name and email are missing from the paste, they were in a header, a text box, or an image.
• —If the dates and job titles arrive as separate blocks, the layout is using a table or columns.
• —For DOCX, also open the header and footer separately: if your email exists only there, assume it is lost.

When the portal only accepts one format

Some portals accept DOCX only, some PDF only, and a few accept either. Follow the portal: a perfectly parsed file that the form rejects is worth nothing.

If you must submit PDF, export from a word processor rather than a design tool, keep one column, use standard fonts, keep the contact line in the body of page one, and run the copy-paste test before uploading.