6 min read · text layer beats file debates
You're not crazy for asking whether PDFs still die in applicant tracking software. The honest answer's simpler than the forum wars: parsers care about characters they can extract, not the extension on the filename.
Check your resume for free on the exact PDF or DOCX you'll attach. You'll see missing contact lines and column flags while the fix is still on your laptop.
Below: when PDF beats Word, when it doesn't, layout traps that scramble reading order, and a short export routine you can finish tonight. For upload errors on a specific portal, pair this with ADP ATS resume upload problems and beat the bots screening flow .
Quick Wins
- Highlight a bullet in the PDF. No selection means rebuild, not re-upload.
- Put phone and email in the document body, not only in a header shape.
- Match PDF vs DOCX to what the careers page lists; don't rename files.
When can ATS read PDF resumes reliably?
Text PDFs work. The system pulls glyphs into strings, then maps them into name, employers, titles, dates, and skills. That's the same pipeline DOCX uses, just wrapped in a different container.
Image PDFs don't. A phone photo or a flattened Canva export is a picture. Humans read it. Most parsers can't build a searchable profile from it without inconsistent OCR.
Layout still matters on text PDFs. Two columns can feed the left rail before jobs. Tables swap dates and titles. Icons aren't the word Email. The file parses, but the record looks empty or scrambled in recruiter search.
I've screened stacks of these in Workday and Greenhouse, and the uploads that survive are boring: single column, standard headings, contact lines typed in the body, month-year dates on their own line.
PDF vs DOCX is mostly portal policy. When both are allowed, pick whichever export passes the copy test and stays under size limits. When the page says Word only, send DOCX. For a wider format comparison, read PDF vs DOCX for ATS .
| File type | Parser usually… | Typical failure |
|---|---|---|
| Text PDF | Accepts and extracts | Columns reorder jobs |
| Clean DOCX | Accepts and extracts | Text boxes hide contact |
| Scan / image PDF | Stores file, weak fields | Blank experience rows |
Layout habits that make PDFs look readable and parse wrong
Trusting the preview on your screen. Your PDF viewer renders a grid. The importer walks text objects in a different order. That's why a designed file can look perfect and still produce nonsense in the application preview.
Before: Skills sidebar in a left column; jobs start in the right column.
After: Skills section below experience, stacked in one column with plain headings.
Exporting from poster tools for every apply. Design apps love columns, icons, and flatten settings. They're fine after a recruiter asks for a packet. They're risky as your only careers-page file.
Before: Phone number inside a header text box on a designed PDF.
After: Name, phone, email, and city on the first lines of the body; header cleared.
Skipping the portal preview on Workday-style forms. A green upload bar doesn't mean titles populated. Read auto-filled experience before you submit, especially when you're wondering can ATS read PDF content you just attached.
Tonight's export check (PDF or DOCX)
Open the source in Word, Google Docs, or a simple builder. Remove tables used only for alignment, text boxes, and decorative images that inflate file size.
Export PDF when allowed, DOCX when required. Try to select a full paragraph on page one. Copy everything into a plain notes app. If job order or dates look scrambled, fix structure before keywords.
Copy-paste header block for the document body
Morgan Ellis
(555) 013-8840 | [email protected] | linkedin.com/in/morganellis | Denver, CO
Experience
Operations Coordinator, Northline Logistics
March 2021 – Present
- Rebuilt dock scheduling in Excel, cutting late trailers 14% in Q2
Upload the same file to HireFlow's checker, clear critical formatting flags, re-export once, and attach that version. Name it clearly: FirstLast_Role_Resume.pdf. Don't send a renamed PDF that is still an image inside.
If the portal hard-rejects the file, swap format once (PDF to DOCX or the reverse), stay lean on images, and retry on desktop Chrome. One clean swap beats ten identical uploads.
Tools after the file is text-safe
Structure first, then match language to the posting. Score your job match so required nouns land in Skills and one dated bullet, not in a footer keyword wall.
Rebuilding from a broken template? Build your resume in a parser-friendly layout, then export PDF or DOCX per the posting.
When the application asks for a short note with the upload, generate a cover letter that stays focused on scope, not file-format excuses.
Stop debating extensions; test extraction
Can ATS read PDF resumes? Yes for real text in a linear layout. No reliable yes for scans or designed grids that hide contact lines. DOCX when the portal demands it. Always read the preview when the site offers one.
Keep one master, export on demand, and refuse to scale applications on a file you haven't copied into plain text. That's the whole fight in one habit.
Run a free ATS check on the attachment after you rebuild. Readable structure beats another hopeful upload of the same broken export.
Read more
Frequently asked questions
Yes. Modern Workday, Greenhouse, Lever, and iCIMS stacks routinely parse text-based PDFs. The risk is not the .pdf extension. The risk is image-only scans, multi-column layouts, text boxes, and contact details trapped in headers. A clean single-column text PDF usually works. If you cannot select and copy text in Preview or Acrobat, rebuild before you apply at scale.
DOCX is often slightly easier for older parsers, but both are widely accepted in 2026. Pick the format the portal asks for. If either is allowed, use a simple single-column file with a real text layer. Fancy designed PDFs fail more often than plain DOCX or plain PDFs. Keep one master and export to whichever format the posting requires.
What you see is visual layout. What the ATS reads is text order and extractable fields. Two columns can reorder experience. Tables can scramble dates. Headers may be skipped. Export a text PDF from Word or Google Docs, not a flat image from a design tool, then verify with a checker. The Notepad paste test is a fast way to simulate a bad parse before you upload.
Usually poorly. A scan is a picture of text. Many ATS parsers need a text layer. OCR inside the employer stack is inconsistent across vendors and tenants. Retype or rebuild in a word processor and export a text PDF or DOCX before applying at scale. Huge photo PDFs also hit upload limits on some portals.
Open the PDF, try to select and copy a paragraph on page one and page two. If you cannot select text, it is likely image-only. Also upload to a free ATS checker that flags parse issues. If copy works but the checker still flags columns, fix layout next. That short QA routine prevents a week of silent applications.