How an applicant tracking system reads your CV

An applicant tracking system reads your CV in three steps: it extracts the text from the file, tries to sort that text into fields such as name, job titles, dates and education, and stores the result so a recruiter can view or search it. Clear structure makes each step more reliable.

Key points

• The software extracts text first, then guesses which text belongs in which field.
• Reading order matters: simple, single-column layouts are read most predictably.
• Headings, dates and job titles are what the software looks for to build your work history.
• Behaviour varies between systems, so aim for clarity rather than tricks.
• You can test your own CV by copying its text into a plain file.

What actually happens when I upload my CV?

When you attach a file or paste your details, the system receives a document. It cannot "see" the page the way you do. It has to pull out the characters and put them in a sequence. Think of it as the document being read aloud in a line, from the first character to the last, with the layout stripped away.

That sequence then goes through a parser, which is a part of the software that guesses what each piece of text means. It looks for patterns: something that looks like an email address, a block of text that follows the word "Education", a phrase that looks like a date range. The output is a set of fields that a recruiter can read in a standard layout, search or filter.

Most recruiters can also open your original file. The extracted fields are a convenience, and what they show depends on how well the parsing went.

Step one: how does it extract the text?

For a Word document or a PDF made from a word processor, the file already contains text, so extraction is usually straightforward. The order in which the text is extracted follows how the file is built, which is not always the order you see on the page.

Problems appear when:

• The file is an image or a scan, so there is no text to extract.
• Text sits in a text box, a table or a graphic.
• The page uses columns or a sidebar, and the extractor reads across them.
• Special characters or unusual fonts are replaced with odd symbols.

A quick test: open the file, select all, copy, and paste into a plain text editor. What you see is close to what many parsers receive.

Step two: how does it sort text into fields?

The parser looks for clues. Headings such as "Work experience" or "Education" tell it where a section starts. Lines that contain a title and a company name, followed by a date range, suggest a job entry. A set of words after "Skills" may be treated as a skills list.

This is why consistent patterns help. If every job entry follows the same order, the parser is more likely to read them correctly. Compare these two:

Consistent	Inconsistent
Stock Controller, Hillside Hardware, Pretoria, March to August	Hillside Hardware (Pretoria). From March. I was the stock controller until August
Cashier, Fresh Mart, Durban, September to February	Worked at Fresh Mart as a cashier, left in February

Both mean the same to a person. The first is easier for software to separate into title, employer, place and dates.

Names of sections matter as well. "Work experience", "Employment history" and "Experience" are all easy to recognise. A heading such as "My journey so far" is not.

Step three: how does it store and search?

Once the fields exist, the employer's system can list your application with a name, a most recent job title and a date, show a profile page, and make the text searchable. A recruiter can search for a word, such as a job title or a tool, and see which applications contain it. Some systems also compare the words in the advert with the words in your CV and suggest a match.

How much of this is used depends on the employer. A small business may only read the original file. A large one may filter by answers to screening questions, search by terms and use tags. Because you cannot see their setup, the safest approach is a CV that reads well in both forms.

Why do the reading order and layout matter so much?

A person looks at a page and groups things by position. A parser reads text in sequence. If your page has a left sidebar with skills and a right column with experience, the parser may read the whole sidebar first, then the main column, or mix lines from both. Your most recent job might end up attached to the wrong dates.

A single column means the sequence on the page matches the sequence in the file. It also helps on phones, where columns are often squeezed. If you want some visual separation, use headings and white space rather than boxes.

What does it do with dates?

Dates help the system work out your timeline. Use a clear, consistent format for all of them, such as the month and year at the start and end of each job, written the same way every time. Avoid mixing styles. If you are still in a job, write "to present" or "to date". A mix such as "since the start of last year" cannot be read as a date.

What happens to bullets, symbols and special characters?

Plain bullets are generally fine. Unusual symbols, arrows, emoji and decorative characters may turn into question marks or boxes, or be dropped. Use the standard bullet your word processor provides, and keep punctuation simple. Avoid putting important content inside a symbol or an icon.

Does it read my skills the way I expect?

A skills list is easy to read when it is plain words separated by commas or lines. A graphic of stars or bars may be invisible to the system and says little to people anyway. If you list a tool or software, write its name in full, such as "Microsoft Excel", then use the common short form elsewhere. This helps both software and a person from outside your field.

What about headers, footers and the file name?

Some systems do not read the header and footer areas of a document, so a phone number stored there may not appear in the extracted fields. Put your contact details in the main body of the page. The file name may be shown to a recruiter, so use a clear one with your name, such as "Thandi-Mokoena-CV".

Can I know exactly how a given system reads my CV?

Not reliably. Systems differ, and the same system may be set up differently by each employer. Updates also change behaviour. This is why guidance about "how an ATS works" should be treated as general, and why no one can honestly promise that a particular format will pass. What you can do is remove common causes of error and then test.

How can I test my own CV?

Try this quick routine:

1. Copy all the text from the file and paste it into a plain text editor.
2. Check that your name and contact details are at the top and complete.
3. Check that each job has its title, employer and dates in the same order as on the page.
4. Check that no section is missing or duplicated.
5. Look for strange symbols.
6. Open the file on a phone and see whether it is readable.

If something looks wrong in the text, fix the layout that caused it, rather than adding words.

What should I do if the application form asks me to re-type my details?

Many systems ask you to upload a CV and then show fields to correct. Take your time to check them. Make sure employer names, job titles and dates are right. Fields are often what the recruiter sees first, so it is worth the few minutes. Paste text from your CV rather than typing from memory, to keep wording consistent.

Frequently asked questions

How does an ATS read a CV?

It extracts the text from your file, tries to sort the text into fields such as name, job titles, dates and education, and then stores it so recruiters can view or search it. Simple, single-column layouts with standard headings are read most reliably.

What is CV parsing?

Parsing is the step in which software turns the text of your CV into organised fields. It looks for patterns such as headings, dates and job titles. If the layout is complex, the parser may put information in the wrong place.

Does an ATS read tables and columns?

It may, but the text order can become mixed up, so details can end up in the wrong fields. Behaviour varies between systems. A single-column layout without tables is the safest choice.

Is there a way to see how my CV is parsed?

You cannot see inside the employer's system, but you can copy the text from your file into a plain text editor and see how it comes out in order. If it looks jumbled there, a parser may struggle too.

Next steps

Run the six-step test above on your current CV, fix the first thing that looks wrong, and test again. If you want a guided way to put this together, you can build your CV step by step with CV Sorted: it is free to build and preview.