An applicant tracking system is not a robot deciding whether you're good enough. It's a database with a search box. Once you understand that, most "ATS tricks" stop being mysterious.
Almost every mid-size and large employer runs applications through an applicant tracking system (ATS) — Greenhouse, Lever, Workday, Taleo, and so on. A lot of advice online treats these systems as an AI gatekeeper that scores you out of 100 and bins you if you fall short. That is not really how they work, and believing it leads people to do strange things, like stuffing white keyword text into the margins.
What an ATS actually does is more boring and more fixable: it parses your file into structured fields — name, employer, title, dates, skills — and stores them so a recruiter can search and filter. When your resume gets "filtered out," it usually means one of two things happened. Either the parse went wrong and your experience never made it into the database correctly, or it parsed fine but you didn't use the words the recruiter searched for.
1. Your layout confuses the parser
This is the single most common technical failure, and it is invisible to you. Your resume looks perfect in a PDF viewer. The parser sees something else entirely.
The usual culprits:
- Text inside images. A resume exported as an image, or with your name in a graphic header, parses as an empty document.
- Tables used for layout. Some parsers read tables cell by cell in the wrong order, interleaving your job titles with your dates.
- Text boxes and multi-column layouts. A two-column design can parse as one scrambled column, mixing your skills list into the middle of a job description.
- Headers and footers. Contact details placed in the actual header/footer region of a document are sometimes skipped entirely — which is how people end up with no phone number in the system.
- Uncommon fonts or heavy glyph substitution, which can turn accented characters and bullets into question marks.
How to check this yourself in 30 seconds
Open your resume PDF, select all the text, copy it, and paste it into a plain text editor. What you see is roughly what the parser sees. If sections are out of order, words are glued together, or your contact line has vanished, that is your problem — and no amount of keyword tuning will fix it.
2. Your section headings are creative
Parsers look for conventional headings to decide where one section ends and the next begins. "Where I've Worked" and "My Journey" read nicely to a human and mean nothing to a parser expecting Experience, Education, or Skills.
This one costs you a lot for very little creative upside. Use standard headings. Spend your originality on the content underneath them.
3. Your dates and titles are formatted inconsistently
Employment dates are one of the fields recruiters filter on most often — years of experience, current employment, gaps. If your date formats vary between entries ("2021-2023" in one, "March 2021 – Present" in another, "'21–'23" in a third), some entries will parse and others won't.
Pick one format and use it everywhere. Mar 2021 – Present is unambiguous and parses reliably. Same principle for job titles: put the title and the employer on their own clearly separated lines, not merged into a single sentence.
4. You're missing the words the recruiter is searching for
This is the real "keyword" issue, and it's less sinister than it sounds. A recruiter filling a Kubernetes role searches for "Kubernetes." If your resume says "container orchestration" but never the product name, you don't appear in that search. You weren't rejected — you were never retrieved.
The fix is not to paste the job description into your resume. It's to check whether the specific nouns in the posting — tools, platforms, certifications, methodologies — appear anywhere in your resume in a way that's actually true of you.
Same experience, different retrievability
❌ Managed our container infrastructure and deployment pipeline, improving release reliability.
✅ Managed Kubernetes infrastructure and the CI/CD pipeline (Jenkins, Docker), cutting failed releases by [X%].
Note what changed and what didn't. The work is identical. The second version just names the tools, which makes it findable — and adds a metric slot, which makes it persuasive to the human who eventually reads it.
Don't fake it
Only add a keyword if you can defend it in an interview. Listing a technology you've never touched gets you retrieved into a conversation you will immediately lose. Keyword coverage is about not under-selling what you did do, not about claiming what you didn't.
5. You're applying with one resume for everything
A generic resume is optimized for no particular search. If you're applying to backend roles, data engineering roles, and platform roles with the same document, it is probably a weak match for all three, because each of those recruiters is filtering on a different vocabulary.
You don't need to rewrite from scratch each time. In practice, tailoring means: adjust the summary line to name the target role, reorder your skills so the relevant ones lead, and make sure the two or three central nouns from the posting appear in your experience bullets where they're genuinely true.
Paste a job description and your resume to see which of its keywords you're actually covering — and which ones you're missing.
What actually matters, in order
If you only fix a few things, fix them in this order:
- Make sure it parses. Do the copy-paste test. A beautiful resume that parses as gibberish is worth less than a plain one that parses cleanly.
- Use standard section headings and consistent date formats. Cheap to fix, disproportionately costly to get wrong.
- Name the specific tools and terms from the posting that are genuinely true of you.
- Then, and only then, worry about design. Once it parses correctly, a clean single-column layout is enough. Nobody has ever been hired because of a resume's colour scheme.
One thing ATS advice usually gets wrong
You'll see claims that ATS systems auto-reject the overwhelming majority of resumes with a specific score. Treat those numbers with suspicion — they're widely repeated and rarely sourced, and different systems behave very differently. Some knock out applications on hard requirements like work authorisation. Many do no automatic rejection at all and simply rank and filter for a human.
The practical takeaway doesn't change either way: your resume needs to survive parsing and show up in the right searches. Both are within your control, and neither requires gaming anything.
