PDF or Word? The difference that decides whether a system can read your CV

What is actually inside a PDF

A PDF does not store sentences. It stores glyph shapes and their positions on the page — where each character is drawn, and at what size. When an applicant tracking system extracts text from it, it reassembles sentences out of those positions.

For languages written left to right this works well. Arabic, though, is stored in visual order rather than logical order: the order the glyphs are painted on the page, not the order they were typed. The result is that each line comes out with its words reversed.

Typed: محمد العتيبي — extracted: العتيبي محمد

Why you have never noticed

Because when you open the file and copy from it, your browser or PDF reader reorders the text for you before it reaches your eye. It looks correct, so you assume it is.

An applicant tracking system does not open the file and does not copy from it. It reads the text layer directly, with nothing in between to fix the order.

The damage varies. Keyword matching usually survives — "accountant" is still "accountant" whatever surrounds it. Anything that reads a phrase does not: "senior engineer" becomes "engineer senior", and "project management" becomes "management project".

Check your own file, right now — it takes thirty seconds

Open your PDF in a browser, select a line of your CV, copy it, and paste it into any plain text field — a search box, a simple text editor. Then read the order of the words.

If it comes out correct, do not relax: that means the reader you used repaired the order for you, not that your file is sound. Copy the same line from a different reader and you will often get a different answer.

That disagreement is the evidence. If the text were stored in an unambiguous order, two readers could not disagree about it. An applicant tracking system's parser has none of these repairs — and nothing tells you what it extracted.

The rule: if two readers disagree about the extracted order, a third system is guessing.

The worse case: a PDF with no text in it at all

Some PDFs contain no text layer whatsoever. The common ones are scans, photographs of a printed page, and design-tool exports that wrap an image in a PDF shell. You see a complete CV; the system sees a blank page.

There is a nastier variant: a file that does contain text, but whose embedded font carries no mapping table back from glyph shapes to Unicode characters. Extraction returns meaningless codes instead of words.

Arabic is more exposed to this than most scripts, because its letters join and change shape by position, and many composed forms — لا, for instance — are a single glyph in the font but not a single character in Unicode. Without that mapping table they are lost.

Same test, simpler: if you cannot select and copy the text, there is no text — and the system will find nothing.

What Word does instead

A .docx stores logical Unicode text. The bytes in the file are the characters you typed, in the order you typed them, with no relationship to where they are drawn on the page.

Any parser that opens it gets them back exactly. Nothing is reassembled and no order is guessed.

This is why the ATS checker on this site asks you to upload the Word version rather than the PDF.

And what still breaks in Word

Word is not a cure. It solves the ordering problem and nothing else — every structural problem a PDF has, a .docx has too.

Tables first. Plenty of people build a CV as a two-column table to get the alignment right, and a parser may read a table column by column rather than row by row — so every date collapses into one block, detached from the job titles they belong to.

The same goes for text boxes and for anything placed in a page header or footer. What you put in the header — usually your name and phone number — can be dropped entirely. Worse still is putting your details inside an image.

• No tables — keep it one column
• No text boxes and no columns
• Nothing in the page header or footer
• No images containing text
• Plain section headings: "Work experience", not "My journey"

So when should you use PDF?

When a person reads it. A PDF preserves the layout exactly as you designed it: fonts, spacing and line breaks do not move, whatever opens it. Email it to a recruiter, print it, attach it to a direct message.

When a system reads it — a job portal, an application platform, a file uploaded into a web form — Word is the safer answer.

The practical rule: if the posting names a format, use it. If it does not, and you are uploading into a form on a website, choose Word.

• Direct email to a recruiter → PDF
• Printing or handing it over → PDF
• Job portal or upload form → Word
• The posting names a format → use that

What about an English CV?

The problem is much smaller: English is written left to right, so visual and logical order agree and extraction is usually clean.

The other PDF problems remain in both languages, though. Tables, columns and text boxes, and anything in a page header or footer, can all be dropped by a parser or read in the wrong order. That is why all four templates on this site are a single column with no tables.

Myths that keep coming back

"PDF is more professional." Professionalism is in what you wrote, not in a file extension. Nobody at the other end judges you by the format — it is either a person reading the content or a system reading the text.

"Systems reject Word files." They do not. Most job portals accept both, and some ask for .docx specifically because it is easier to parse. What actually gets rejected is everything else: .odt, .pages, .rtf, and archives.

"Export it as an image so nobody can edit it." That is the worst option available. It prevents editing and prevents reading with it — and being read is the entire point.

".doc is safer because it is older and more widely supported." The opposite. .docx is an open format any parser can read; .doc is a legacy binary format many systems stumble over. Always save as .docx.

Small details that get sound files rejected

Once the format and the structure are right, there is a layer of purely administrative things that can drop an application before anyone reads it.

The file name matters most. Name it after yourself and the role, in Latin characters — some upload systems mangle or reject non-Latin file names. And "CV final 3 edited.pdf" arrives at the other end exactly as written.

• File name: Mohammed-Alotaibi-Accountant.docx
• One file only — never upload an archive
• Save from Word or an editor that exports real .docx, not .odt or .pages
• Keep it under 2 MB — some forms reject larger files with no clear message
• No password protection and no editing restrictions

The practical answer: keep both

Do not choose. Build the CV once, download it in both formats, and use whichever fits where it is going.

That is what the osaCV editor does: the same document exports as PDF and as Word, from the same template, so the two never disagree about the content or the order of the sections.