PDF vs DOCX: The Format Wars of Lost Information
By Umut Barış Terzioğlu · · Updated
Whether to send your CV as a PDF or a Word file is one of the most common questions in a job search, and the answer is less dramatic than the debate suggests. A parser does not care about the file extension as such. It cares whether it can pull clean text out of the file, in the right order. Both formats can deliver that, and both can fail.
The real question: is there a text layer?
A PDF can contain two very different things. A PDF exported from Word, Google Docs or LibreOffice carries the actual characters of your CV, so software can select, copy and search them. A PDF made from a scan, a phone photo or a design file flattened into an image only contains a picture of text. To a parser that second kind looks empty, unless the system runs optical character recognition (OCR) first, and OCR adds errors of its own.
A DOCX file always stores text as text, which is why it has a reputation for being the safe choice. Word files have weak spots too: text boxes, tables used for layout and content in the header or footer are read inconsistently, and some parsers skip headers and footers altogether. If your name and phone number sit in the header, they may never reach the candidate profile.
As the first part of this series explains, what counts is the text a system extracts, not what the page looks like.
PDF or DOCX CV: what each format does well
| Text-based PDF | DOCX | Scanned or image PDF | |
|---|---|---|---|
| Looks the same on every screen | Yes | Not always; fonts and spacing can shift | Yes |
| Text a parser can read | Yes | Yes | No, only after OCR |
| Typical weak spots | Columns and text boxes read out of order, icon fonts, lost ligatures | Headers, footers, text boxes, layout tables | Everything |
| A recruiter can edit it | No | Yes | No |
Which file to send
- Follow the instructions. If the ad or the application portal names a format, use it. Instructions beat any general rule.
- No instructions, both accepted: a text-based PDF exported from a word processor is a sensible default. Your layout stays fixed, and it reads well as long as the layout inside is simple.
- Word requested, or applying through an agency: send DOCX. Some recruiters and agencies want a file they can edit or move into their own template.
- Never send a scan or a photo of a printed CV, and avoid "print as image" options and design-tool exports that turn text into graphics.
- Keep one source document and export both versions from it, so the PDF and the DOCX never drift apart.
The layout inside the file matters more than the format. A two-column layout can scramble the reading order in a PDF and in a Word file alike.
The copy-paste test
Before you upload a PDF, open it, select all the text (Ctrl+A or Cmd+A), copy it and paste it into a plain text editor such as Notepad, or TextEdit in plain text mode. Then read what arrived:
| In the PDF you see | The paste shows | What it means |
|---|---|---|
| "Project Manager, Acme Ltd, 2021–2024" | "Project Manager Python SQL Acme Ltd" | A second column has been woven into the first |
| A phone icon before your number | A box, a stray letter or nothing | The icon is a font glyph, not text; write "Phone:" instead |
| "Certification" | "Certi cation" | The "fi" ligature was lost on export |
| A full page of text | Nothing at all | The page is an image with no text layer |
If you cannot select the text at all, the file is image-based. Export it again from the original document rather than trying to repair the PDF.
Checking both versions here
You can paste the text you copied into the quick test. It reads pasted text only, so it checks what survived extraction rather than the file itself. With a free account, the full analysis reads PDF, DOCX and plain text files, warns when a PDF has pages without a text layer, and the parser view shows which fields were found and which were missed. Upload both versions and compare the results before you decide which one to send.