The PDF or Word resume question has outlived several generations of the software that made it a question. The old advice — "always send Word, applicant tracking systems can't read PDFs" — was reasonable in 2010 and is repeated today as if nothing has changed, usually right next to a warning about resume-reading robots. Here is the current, testable answer: both formats parse well in every major modern system, the posting's instructions override everything, and the one file type that genuinely fails is the one this debate never mentions — the image PDF. This guide gives you the default, the exceptions, and the export settings that keep your text layer clean; for the wider picture of what happens after upload, start with the ATS resume guide.
The default answer for 2026 and why it changed over time
Default: send a text-based PDF, unless the posting says otherwise.
Why PDF wins as the default: it freezes your layout. A PDF renders identically on the recruiter's screen, the hiring manager's phone, and the printout in an interview folder. A DOCX reflows — fonts substitute when the reader's machine lacks yours, spacing shifts between Word versions, page breaks wander, and "track changes" or margin comments can resurface embarrassingly. Since a substantial share of resume readings happen in the original file rather than the parsed summary, presentation stability is worth having.
Why this was ever controversial: early ATS parsers genuinely handled PDF badly. Extracting text from PDFs is harder than reading a DOCX — a PDF stores positioned characters, while a DOCX stores structured text — and 2000s-era engines often mangled or skipped PDFs entirely. "Send Word" was good advice then. Parsing engines have since made PDF extraction routine; the mainstream platforms — Workday, Greenhouse, Lever, iCIMS, Taleo — all accept and parse text-based PDFs as a matter of course, and their application forms say so by listing PDF first among accepted types.
The advice fossilized because fear is sticky and testing is rare. It survives in exactly the same ecosystem as the "robots reject 75% of resumes" myth: repeated, never sourced. What actually matters is not the extension but what is inside the file — which is the next section.
How PDFs and DOCX files parse differently, tested
We run both formats through parsing tools continuously while testing Workplacea templates. The honest summary of the differences:
DOCX is the parser's easy case. The file is structured text — headings, paragraphs, lists are labeled in the format itself — so extraction is near-lossless. Its failure modes are structural, not textual: layout tables read in the wrong order, floating text boxes skipped, content in page headers ignored. A plain single-column DOCX parses essentially perfectly everywhere, including older engines.
Text-based PDF parses cleanly on all modern engines we test. The extraction step has one PDF-specific wrinkle: text comes out in the file's internal stream order, which matches visual order for single-column layouts and may not for multi-column ones — the mechanics behind our two-column test results. Occasional PDF-specific artifacts exist (ligatures extracting as odd characters from some generators), all catchable in ten seconds with a paste test.
Image PDF — a scan, a screenshot, or a flattened design export — is the catastrophic case regardless of the debate you thought you were having. There is no text to extract. Some systems attempt OCR; many do not. The parsed record comes out empty: no name, no history, nothing to search. Every hour spent choosing between PDF and Word is wasted if the file is an image of either.
Ranked by parsing outcome: clean DOCX ≈ clean text-PDF, both far above any file with structural traps, all infinitely above an image. Format choice matters less than file hygiene — the same conclusion as the rest of our ATS format testing.
When the job posting tells you the answer — always obey it
Every default in this article yields instantly to one thing: instructions.
- "Submit your resume as a PDF" — send PDF.
- "Word documents only" or a form that accepts
.doc/.docxalone — send DOCX. - A government or staffing-agency portal with format demands that seem archaic — comply anyway. Some agency workflows genuinely edit resumes (removing contact details before client submission is standard practice), which is the most common legitimate reason Word is requested.
Two reasons compliance outranks optimization. Mechanically, some portals reject non-listed file types at upload, making the debate moot. Socially, the instruction is a comprehension test you are being graded on before anyone reads a word of your resume. An applicant who sends a PDF to "Word only" has answered the first screening question incorrectly — a small, silly, human-inflicted wound of exactly the kind that common resume mistakes are made of.
When a form accepts both and the posting is silent, the default stands: text-based PDF. And keep a current DOCX master regardless — you will need it for agencies, for last-minute edits, and as the source from which your PDFs are generated.
Export settings that produce clean, selectable PDF text
A text-based PDF is not a lucky accident; it is the product of exporting correctly. The rules, tool by tool:
- From Word, Google Docs, or LibreOffice: use the built-in export ("Save As PDF" / "Download as PDF"). These generators write proper text layers. Do not "print to PDF" via an image-based printer driver, and never scan a printout.
- From design tools (Canva, Figma, Photoshop): danger zone. Some export paths flatten text into raster images — the exact image-PDF trap. If you must use a design tool, choose the export option that preserves real text, then verify (below). Photoshop resumes are images by construction; avoid.
- Fonts: stick to standard fonts and let the exporter embed them. Exotic display fonts occasionally trigger outlining — conversion of text into vector shapes — which destroys the text layer while looking perfect.
- One page size, no security settings. Password-protected or copy-restricted PDFs can block extraction outright. Ensure no protection is applied.
- Filename before export: set it once, correctly (next section), so the file that leaves your machine is the file you named.
The ten-second verification, every time you export: open the PDF, select all, copy, paste into a plain text editor. Your resume's text, in sane order: clean file. Nothing, or fragments: the export path failed you — fix it before anyone else sees it. Workplacea's exporter produces verified text-layer PDFs by default; the walkthrough is in how to export your resume to PDF.
The image-PDF trap: scans and screenshots that parse as nothing
This failure deserves its own section because it is common, invisible to the person committing it, and total.
How people end up with image resumes, ranked by frequency in files we see: exporting from a design tool with a rasterizing setting; scanning a printed resume (often to capture a signature); screenshotting a resume to "lock" it; downloading a resume image from a portfolio site; and receiving a flattened file from a paid designer who worked in Photoshop.
Why it is invisible: the file looks perfect. It opens, prints, and reads beautifully — to eyes. The absence of a text layer has no visual symptom. The person most likely to be sending an image resume is often the person who invested most in how it looks.
What happens downstream: extraction yields nothing, so the parsed record is blank — the mechanics are in how resume parsing works. In systems that pre-fill application forms from the parse (Workday applicants know this step well), every field arrives empty. In recruiter search, the candidate matches no query, ever. No rejection is issued; the application simply behaves like a ghost.
The paste test catches it every time, in ten seconds, for free. If any single habit from this article survives, make it that one.
File naming conventions recruiters appreciate
Small, free, and noticed more than you would expect — because the recruiter sees the filename in their download bar, their inbox, and the candidate record's attachment list.
The pattern: Firstname-Lastname-Resume.pdf. Optionally add the role for direct-email applications: Firstname-Lastname-Resume-Data-Analyst.pdf.
What to avoid, all real examples from hiring folklore for good reason: resume.pdf (anonymous in a downloads folder full of them), resume_final_v7_REAL.pdf (narrates your editing anxiety), Johns resume (1).pdf (a re-download artifact), a filename containing a different company's name (the tailoring equivalent of calling someone by an ex's name), and dates in filenames (advertises how long you have been searching).
Set the name on the master document once and every export inherits it. Thirty seconds, permanent, done.
Frequently asked questions
Does any major ATS still fail to parse PDFs?
Among mainstream current platforms — Workday, Greenhouse, Lever, iCIMS, Taleo — text-based PDF parsing is routine and PDFs are listed as accepted upload types. Genuinely ancient or niche systems exist in the long tail, which is one more reason the posting's stated preference always wins. Absent instructions, PDF is a well-supported default, not a gamble.
Should I keep both PDF and Word versions of my resume?
Yes, but as source and output rather than as two products: maintain one DOCX (or builder) master, export PDFs from it per application. You will need the DOCX for staffing agencies and format-restricted portals, and single-sourcing prevents the two versions from drifting apart.
Is Google Docs or a shared link an acceptable resume format?
Not as a submission. Upload forms want a file; a link adds friction, can break, and exposes edit history. Export the Doc to PDF (Google's exporter produces clean text layers) and submit the file.
What about plain text (.txt) resumes?
Only when a system explicitly demands pasted plain text — some older portals do. Keep a stripped .txt version for those paste boxes: no bullets beyond hyphens, no tabs, standard headings. Sending .txt as your actual file elsewhere discards all presentation for zero parsing gain.
Export it right, then prove it
The whole format debate collapses into one verifiable property: does your file carry clean, ordered, selectable text? Workplacea's editor exports text-layer PDFs built to pass that test, and the free resume checker will show you the parsed output of whatever file you already have — so you can settle PDF versus Word for your own resume with evidence, in about two minutes.