# What an ATS does with your resume

> An applicant tracking system stores your application, tries to read your resume into fields, and lets a recruiter search and sort what it stored. Plain, single-column files with standard headings read best. What any one employer's system rejects automatically is set by that employer, and you cannot see it.

Source: https://zorlya.com/blog/what-an-ats-does-with-your-resume

By Zorlya. Published 2026-09-12. Resumes and applications. 5 min read.

Few topics in job searching carry as many confident claims as the applicant tracking system, or ATS. You will read that it rejects most resumes before a human sees them, that you need to hide white text full of keywords, that a two-column template is fatal. Some of that is partly true, some is false, and much of it cannot be known from outside.

This post separates what the software does from what people say it does, and says plainly where the public record runs out.

## What an ATS is for

An ATS is the system an employer uses to post openings and collect applications. Greenhouse, Lever, Ashby, Workable, Workday and SmartRecruiters are examples. Its main job is bookkeeping for the hiring team: one place where every applicant for a role lives, with a stage (new, screening, interview, offer) and notes from the people who met them.

It is a database with a workflow on top. It is not a single product with one behaviour, and each employer configures its own.

## What happens to your file

When you submit, three things usually happen.

**1. The file is stored.** The original PDF or Word document is kept, and a recruiter can open it as you made it.

**2. The file is parsed.** The system tries to read the text and sort it into fields: name, email, phone, employers, titles, dates, schools, skills. Some fill in the application form for you from this. Parsing is software guessing from layout and wording, so it is the stage where odd formatting does damage.

**3. The record is made searchable.** A recruiter can search applicants by words, such as a skill or a job title, and filter by fields. Some systems also show a ranking or match indicator. Whether yours was ranked, and how, depends on that system and that employer.

## Myth and mechanism

| Claim | What is likely true |
|---|---|
| "The ATS rejects most applicants automatically." | Rejections can be automated, usually through knockout questions the employer wrote, such as work authorisation or a required licence. Whether a system also discards resumes on text matching is a setting the employer controls, and applicants cannot see it. |
| "You need to match the exact keywords." | Search by words is real, so using the posting's terms for work you did is sensible. Stuffing terms you do not have helps no one, because a person reads the resume next. |
| "White text full of keywords beats it." | Text a person cannot see is easy for a system to extract and for a recruiter to notice when they paste the document. It looks like an attempt to deceive, and it can cost you the application. |
| "Tables and columns break parsing." | They can. Parsers read text in an order, and a layout that reads left to right across two columns can interleave the content. A single column removes the question. |
| "PDF is unsafe, use Word." | Either generally works if the text is real text. A scan or an image of a resume has no text to read. |

The right-hand column is cautious on purpose. Employers do not publish their configuration, and vendors describe their products in marketing language more than in detail.

## What actually breaks parsing

These problems have a mechanism behind them, so avoid them.

- **Text inside images.** A logo, an icon set or a skills graphic with rating bars is not read. A skill shown only as five dots is invisible.
- **Headers and footers.** Some parsers skip them. If your name and email live only in the page header, they may not be found.
- **Multiple columns and text boxes.** Reading order can scramble, so a job title ends up next to the wrong employer.
- **Unusual headings.** "Where I've been" instead of "Experience" gives a parser nothing to recognise. Standard names (Experience, Education, Skills) are safe.
- **Dates in odd forms.** "Spring 2021 to now" is harder to read than "Mar 2021 to Present". Pick one format and keep it everywhere.
- **Fancy bullet glyphs and special characters.** They can turn into stray symbols. Plain bullets are fine.

A quick test costs nothing: open your PDF, select all, copy, and paste into a plain text editor. If the order, the dates and your contact details come through cleanly, a parser has a fair chance. If the paste is jumbled, fix the layout.

## Try the paste test now

Run the copy and paste test before reading further. It takes under a minute, and it tells you more about your own file than any generic advice here.

## What the public record does not tell you

Be wary of anyone who quotes a rejection percentage, a time a recruiter spends per resume, or the exact algorithm of a named system. Those are rarely published in a form you can check, and they differ by employer.

What you can say with confidence is limited: the software stores, parses and indexes applications, a person usually makes the decision to move someone forward, and the formatting choices above affect how well the first two steps go. Everything past that is a guess dressed up as a fact.

## A practical checklist for this week

1. One column, standard headings, a plain font, real text.
2. Name, email and phone in the body of the page, not only in a header.
3. Job titles, employer names and dates in a consistent pattern.
4. A skills section that spells out tools by their common name, such as "PostgreSQL" as well as the abbreviation if you use both.
5. Where a posting uses a term for something you really did, use their term once, in a bullet that describes the work.
6. Save as PDF with text, and run the paste test.

Then stop. Over-optimising for a parser you cannot see costs time that a better bullet or a better-chosen job would repay.

## What Zorlya's ATS check does and does not do

Zorlya has a check called ATS check, and its limit is worth stating. It compares the skills a job posting requires against the skills on your profile, using a canonical form so that "React.js" and "React" count as the same skill, and it shows which are present and which are missing, plus a score for the share present. It is deliberately not a simulation of any real ATS. It tells you about a keyword gap against a specific posting, and nothing about how an employer's own system will parse or rank you.

Use it as a prompt to ask: do I have this skill and not say so, or do I lack it? Only the first is fixable with editing.