I have hired for roles at three companies, and at each one every resume I ever looked at first passed through an applicant tracking system. Candidates tend to imagine the ATS as a gatekeeper robot that scores their resume and throws most of them in the trash before a human sees them. That picture is mostly wrong, and believing it leads people to do strange things to their resumes that hurt more than they help. Here is what actually happens.
What an ATS Actually Is
An applicant tracking system is a database with a workflow attached to it. Its main job is to help a company organize applicants, not to judge them. When you apply, the system stores your resume, records which job you applied to, tracks what stage you are at, and lets recruiters search and filter the pool. Common systems you will run into include Workday, Greenhouse, Lever, iCIMS, Taleo, and SmartRecruiters. They differ in polish, but the core purpose is the same: keep hundreds or thousands of applicants from turning into an unmanageable pile of email attachments.
The reason it matters to you is a smaller feature buried inside that database: parsing. Before the system can store your information in a searchable way, it has to read your resume file and pull the text apart into fields. That parsing step is where good resumes quietly get mangled.
The Parsing Step: How the Machine Reads You
When you upload a PDF or Word file, the ATS does not "see" your resume the way you do. It extracts the raw text and then tries to guess which chunk of text is your name, which is your email, which lines are job titles, which are employers, and which are dates. It is doing pattern recognition on a document that was designed for human eyes, and it gets things wrong.
A few concrete examples of what parsing does with real resumes:
- It reads text in the order the file stores it, which is not always the order you see on screen. Two-column layouts often get read left-to-right across both columns, scrambling your work history into nonsense.
- It frequently cannot read text that lives inside an image, a text box, or a graphic. If your name or skills are in a decorative header image, the system may register them as blank.
- It maps your content into standardized fields such as "Company," "Title," "Start Date," and "End Date." If it misreads a date because you wrote it in an unusual format, your six years of experience can show up as a few months.
- It stores your skills and keywords so recruiters can later search for them. Skills buried in dense paragraphs are harder for both the parser and the recruiter to surface.
The output of parsing is a structured profile. When I open a candidate in the system, I often see the parsed version first, sometimes alongside the original file. If the parsed version is garbage, my first impression of an otherwise strong candidate is confusion.
The Myth of Automatic Rejection
Here is the part people get most wrong. The overwhelming majority of applicant tracking systems do not automatically reject resumes based on a hidden score. There is no universal algorithm that decides you are a 62 out of 100 and deletes you. What exists is far more mundane, and in some ways more important to understand.
What Really Filters People Out
Two things do most of the filtering, and neither is a mysterious AI judgment:
- Knockout questions. Many applications ask direct screening questions: Are you legally authorized to work in this country? Do you have the required license? Are you willing to work on-site? A "no" to a hard requirement can genuinely route your application out of the active pool automatically. This is a rule the employer wrote, not an opinion the software formed.
- Recruiter search and filtering. When I have 400 applicants for one opening, I search the pool for the terms that match the role, such as a specific certification, tool, or job title. If your resume never contains the words that describe what you do, you simply do not appear in my results. You were not rejected. You were never found.
That second point is the real lesson. Getting "screened out" by an ATS usually means your resume was searchable but never surfaced, or parsed so poorly that the searchable version did not reflect your actual experience.
Why This Changes How You Should Write
Once you understand that the machine is a parser and a search index rather than a judge, the right strategy becomes obvious and honest. You are not trying to trick an algorithm. You are trying to make sure a human recruiter can find you and read an accurate version of your background.
That means writing so the parser succeeds and the recruiter's search finds you:
- Use the same words the job description uses for skills and titles, because those are the words recruiters search for. If the posting says "accounts payable" and you wrote "AP," include both.
- Put your real job titles in plain text, not creative ones. "Growth Ninja" does not match a search for "Marketing Manager."
- Keep the structure simple enough that the parser maps your history correctly. Standard section headings like "Experience," "Education," and "Skills" are recognized reliably.
- Never stuff hidden keywords in white text or tiny fonts. Recruiters and modern systems catch this, and it reads as dishonest the moment a human sees the file.
A Practical Takeaway
Before you apply anywhere, do one test. Copy all the text from your finished resume and paste it into a plain text editor such as Notepad or TextEdit. Read what comes out. If your name, contact details, job titles, employers, and dates all appear in a sensible order and nothing is missing, the parser will very likely read you correctly too. If the text is scrambled, out of order, or dropping whole sections, fix the layout before you send it anywhere. That five-minute check does more for your callback rate than any keyword trick, because it makes sure the accurate version of you is the one that reaches a human.