ATS Cannot See Design: Anatomy of the Real Text Output

A CV built in a design tool can look polished on screen and still reach an applicant tracking system as a short, jumbled list of words. The reason is simple: before anything is matched or scored, the system runs ATS text extraction. It pulls the characters out of your file and leaves almost everything else behind. Colours, icons, bars, photos and the position of boxes on the page do not travel with the text.

This article opens a short series on how parsers read CV files. It explains what the extracted text is, what disappears on the way, and how to look at the text of your own CV.

What the parser actually receives

A PDF does not store sentences. It stores small runs of characters, each with a font and a position on the page. The extractor collects those runs and rebuilds lines from their positions, usually top to bottom and left to right. A DOCX file keeps its text in document order, but headers and footers are stored separately, and text boxes are anchored objects outside the normal flow of paragraphs. Their content can come out in an unexpected place, or not at all.

Whatever the format, the result is plain text. Only that text is passed on to the step that looks for your name, job titles, dates and skills. If a detail is not in the extracted text, it does not exist for the rest of the process.

What gets lost, and what to write instead

The last row is common enough to have its own article: two-column CVs and ATS.

A worked example

Picture the top of a designed CV: name and title, a phone and an envelope icon in front of the contact details, and a skills panel with bars. A plausible extraction of that area looks like this:

JANA WEBER
Data Analyst
□ +49 151 2345 6789 □ jana.weber@example.com
Python
SQL
Tableau

The bars left nothing behind, so "advanced" and "basic" now look identical. The icons turned into unknown characters (shown here as □) sitting right against the phone number. Written with extraction in mind, the same block reads:

Jana Weber
Data Analyst
Phone: +49 151 2345 6789
Email: jana.weber@example.com
Skills: Python (advanced), SQL (advanced), Tableau (working knowledge)

Nothing is lost in this version, because no information depends on the design. Why icons turn into stray characters is covered in icons on a CV.

How to see your own extracted text

You do not need special software for a first look:

1. Open the PDF in a browser or PDF viewer, select all text and copy it.
2. Paste it into a plain text editor, such as Notepad or TextEdit in plain-text mode.
3. Read it top to bottom. Is every section there? Does each job title sit next to its own employer and dates? Are the words spelled correctly?
4. Look for empty boxes, question marks and words with letters missing, such as "Pro le" for "Profile".

If text is missing or broken in the copy, it is missing or broken in the file itself, and no parser can recover it without OCR. Reading order, on the other hand, can differ between tools, so treat a clean copy as a good sign rather than a guarantee.

For a structured second look, paste the same text into the quick test. It runs in your browser and reports parseability and structure findings, such as signs of side-by-side columns, icon characters and missing section headings.

What to take from this

Design is for the person who opens the file later. Every automated step works from the extracted text, so everything that matters (titles, employers, dates, skills, contact details) has to be there as ordinary text, in reading order. Headings belong on that list too, because they tell the parser where each section begins. That is the subject of the last part of this series, CV section headings.