When you apply through an online portal, your CV rarely reaches a recruiter first. An applicant tracking system (ATS) reads it before any human does. What the ATS extracts from your file determines whether your experience, skills, and qualifications make it into the system at all.
Disclosure: I build Parviso, which is mentioned below.
This article covers what a parser actually does, which layout choices cause it to fail, how to check what yours extracts, and where a parse check runs out of answers.
What a Parser Does
A parser does one thing: it converts your CV file into structured data. It pulls out text, then tries to sort that text into fields — name, contact details, work history, education, skills, languages.
It does not read your CV the way a person does. Visual design, column position, and font size mean nothing to it. It reads a stream of characters and looks for patterns it recognises: a date range following a job title, or the word "Education" as a section heading.
If text extraction fails, everything downstream fails with it. A recruiter searching the system for "C1 French" or "EU work permit" will not find your profile — even if both appear clearly on your printed CV.
Six Layout Choices That Break Parsing
1. Two-column layouts
Some CV templates, including some Europass-style ones, may use columns. On a two-column layout, the text can come out in an unexpected order — which can place your contact details between two unrelated jobs. The result is garbled data.
Even when a parser can technically handle both columns, many application portals cannot. The layout carries risk regardless of how clean it looks on screen.
2. Tables
Tables are a common way to align dates with job titles. Parsers frequently misread them. A cell boundary can appear as a word break, or cell content can be extracted in the wrong order. A job title and employer that look paired on screen may arrive in the parser as separate, unconnected fragments.
3. Text boxes
Text boxes in Word or design tools sit outside the main text flow of the document. Many parsers skip them entirely. If your name, contact details, or a key skill sits inside a text box, the parser may not see it at all.
4. Image-only text
A CV scanned to PDF, or one where text has been converted to an image, contains no selectable characters. The parser receives a picture and extracts nothing. This applies to logos, icons used as bullet points, and any text embedded in a graphic.
5. PDFs with no real spaces
Some PDF export tools omit the space character between words. The file looks correct on screen, but the underlying text reads as one long string: "SeniorEngineerBerlin2022". A parser that relies on word boundaries to identify terms finds nothing it recognises. A search for "Berlin" or "Senior Engineer" returns no match.
6. Non-standard section headings
A parser looks for headings it knows: "Work Experience", "Education", "Skills", "Languages". If your CV uses "What I've Done" or "My Background", those sections may not be classified correctly. This matters when a recruiter filters by years of experience or highest qualification — both depend on the parser having read those sections accurately.
One more thing worth checking: accented characters in names and headings, such as Müller, Pérez or Søren. Make sure they survive in the plain-text copy.
How to Test Parsing Yourself
You do not need a tool to run a basic check. Open your CV PDF, select all text (Ctrl+A on Windows, Cmd+A on Mac), and paste it into a plain-text editor. What you see there is roughly what a parser receives.
Read through it in order and ask:
- Does the text flow logically, or do column fragments interrupt each other?
- Are job titles and employers paired correctly, or separated by unrelated content?
- Do dates appear next to the right roles?
- Do section headings survive as recognisable words?
- Are there runs of text with no spaces between words?
If the plain-text version is hard to follow, a parser will find it harder still.
Check your language levels too. "French: C1" written as plain text survives extraction cleanly. The same level conveyed by a row of filled circles or a progress bar does not — that information exists only as a visual element.
The same applies to right-to-work status. If your EU work permit or right to work in a given country appears only inside a text box or graphic, the parser will not find it.
Using Parviso's Plain-Text Preview
Parviso runs a parse-integrity check as part of its free CV scan. You upload your CV and paste the job posting; the scan takes about a minute and requires no account. Your CV is read in memory and never stored.
One of the outputs is a plain-text preview labelled "What the ATS sees". It shows the extracted text in the order a parser would read it, with no formatting. You can see immediately whether a two-column layout has scrambled your work history, whether your contact details survived, and whether your section headings are recognisable.
The check covers columns, tables, text boxes, image-only text, and words without spaces. According to Parviso's public methodology, a two-column layout is flagged even when Parviso itself can read both columns, because many application portals cannot.
The methodology page also notes that Parviso ran the same CV files through one real ATS — Zoho Recruit, on a trial account, on 29 September 2026. In that test, a PDF with no real spaces in its text came back with no jobs and no programming skills extracted. Parviso is clear that this is one ATS, not the whole market, and the result does not represent every system.
Section headings are recognised in eight languages: English, French, Spanish, Portuguese, German, Italian, Dutch, and Polish. If your CV is in German or French, headings like "Berufserfahrung" or "Expérience professionnelle" will be read correctly.
Beyond the parse preview, the scan returns a fit verdict — Strong fit, Possible fit, Long shot, or Blocked — and a ranked list of requirement gaps, with quotes from both your CV and the job posting as evidence. Run the check at parviso.app/check at no cost.
What a Parse Check Cannot Tell You
A parse check tells you whether your text is extractable. It does not tell you what any specific ATS will do with it.
Every ATS product — Workday, Taleo, Greenhouse, SAP SuccessFactors, Zoho Recruit, and others — has its own parser with its own rules. Parviso does not simulate any of them and does not claim to. Its methodology page states explicitly that parser-risk signals do not guarantee a specific ATS will fail on your CV.
Recruiters also differ. Some use automated keyword filters; others read CVs manually. Some roles carry strict qualification requirements — in regulated professions such as medicine, law, or engineering, a licence or formal credential is a legal requirement, not a preference. Others list a degree as a preference that direct experience can offset.
No tool can promise you an interview or predict exactly how a specific employer's system will behave. What a parse check gives you is a clearer picture of what you are actually sending, so you can fix obvious problems before they become invisible ones.
FAQs
Does a PDF always parse worse than a Word file?
Not always. A well-exported PDF with real selectable text parses reliably. The risk comes from specific problems: no real spaces between words, text in image form, or content inside text boxes. A Word file with a complex table layout can parse just as badly. The format matters less than the quality of the underlying text.
My CV is in German. Will an ATS read it correctly?
The bigger risk is section headings. Non-standard phrasing can prevent the parser from classifying your sections correctly. Parviso recognises German headings such as "Berufserfahrung" and "Ausbildung" as part of its eight-language support. Check that umlauts survive in the plain-text copy before you submit.
I used a Europass template. Is that a problem?
Some CV templates, including some Europass-style ones, may use columns — which is a known parsing risk. The visual result is clean and familiar to European recruiters, but the text extraction may arrive scrambled inside an ATS. Check the template you use by running a plain-text copy before you apply.
Can Parviso tell me whether a specific ATS will reject my CV?
No. Parviso identifies parse-integrity risks and keyword gaps, but it does not simulate Workday, Taleo, Greenhouse, or any other vendor's system. Its methodology is public and states clearly that a flagged risk does not guarantee a specific system will fail. What it gives you is the extracted text and a structured gap analysis — enough to make an informed decision before you apply.