This question gets answered with more confidence than the evidence supports, usually as a flat "always use PDF" or a flat "always use DOCX". The useful answer is that the extension matters far less than two specific properties of the file, and that you can check both of them yourself in about a minute. We run the extraction step for a free ATS checker, so what follows is grounded in what actually fails when real resumes are parsed.
The short answer
A text-based PDF is a safe default. DOCX is equally safe and occasionally safer. Plain text is the most reliable and the ugliest. If the job posting or the application form names a format, that instruction beats every general rule on this page, including ours.
Failure case one: a PDF with no text in it
This is the big one, and it is invisible unless you look for it. A PDF can contain real text, or it can contain a picture of text. The second kind looks identical on your screen and is completely unreadable to a parser: there are no characters in the file to extract, only pixels. Our own extractor cannot read these files, and neither can anyone else's without optical character recognition.
You get an image-only PDF by scanning a printed resume, by exporting from some design tools, by taking a screenshot and saving it, or occasionally from a phone document scanner. The fix is to export again from the document source rather than from an image.
The test takes five seconds. Open the PDF and try to select a sentence with your cursor. If you can highlight the words, there is text in the file. If your selection draws a box over the whole page and highlights nothing, you have an image, and you should not send it.
Failure case two: reading order
Parsing a document means flattening a two-dimensional layout into one sequence of characters, and the order that falls out is not always the order you read on screen. A two-column resume can extract as the entire left rail followed by the entire right rail, which splits a job title from the bullets underneath it. Text boxes, multi-column tables, and content placed in page headers or footers are the usual culprits, and contact details in a header are the most common single casualty: they are the one thing you most need a recruiter to find.
Single column, standard section headings, real bullet characters, and contact details in the body near the top. That layout is boring and it survives.
Where DOCX wins and where it loses
DOCX stores structure rather than position, so there is no reading-order guesswork and no chance of an image-only file. It also cannot silently become a picture. What it loses is fidelity: fonts, spacing and page breaks shift between versions of Word and between operating systems, so the document a recruiter opens may not be laid out the way you left it. PDF is the opposite trade. It looks the same everywhere and it is the format that can hide a parsing problem.
When plain text is the right answer
When a form gives you a box to paste into rather than a file to upload. Paste the text instead of fighting the box, and check that the bullet characters survived. A plain text version is also the fastest way to see exactly what a parser sees, which makes it a useful diagnostic even when you never send it.
Honest limits
Nobody outside a given company can tell you how that employer's system is configured or which parser it uses, and anyone quoting a precise rejection rate for a file type is quoting something that was never measured. What is checkable is narrower: whether your file contains extractable text, and whether the extracted text comes out in an order that still makes sense. Those two properties are within your control and they are most of the risk.
A sixty-second check before you send
- Try to select text in your PDF. If you cannot, export it again from the original document.
- Copy the whole file and paste it into a plain text editor. Read what comes out.
- Check that your name, email and phone number appear near the top of the pasted text, not missing because they were in a page header.
- Check that each job title is still followed by its own bullets rather than separated from them.
- Check that bullets came through as lines rather than as one run-on paragraph.
- Keep the file under a couple of megabytes. Oversized files are usually a sign that an image crept in.
If the pasted text reads cleanly, your format is fine and you can stop thinking about this. Resumiz exports a single-column PDF or DOCX in standard fonts for the same reason: it is the shape that survives the parse.