An applicant tracking system is database software. It stores applications, parses them into structured fields, and lets recruiters search and filter. It is not an artificial intelligence that judges your worth, and it does not automatically bin most resumes, whatever you may have read. Understanding what these systems actually do removes most of the anxiety and points at the handful of changes that genuinely matter.
What an applicant tracking system actually does
When you submit a resume, the system attempts to read your document and split it into fields: name, contact details, employers, job titles, dates, education, skills. That parsed record is what the recruiter sees in their pipeline view, and it is what gets searched.
Three things follow from that.
- Parsing accuracy is the real risk. If your job titles land in the wrong field or your dates are dropped, you look like a worse candidate than you are, and you never find out.
- Search is how recruiters find you. For a role with 400 applicants, a recruiter will often search the pool for two or three required terms rather than reading everything. If your document does not contain those terms, you are invisible even though you were never rejected.
- Knockout questions do the actual filtering. Most automatic rejections come from the application form, not the resume: work authorisation, minimum years, required licence, salary expectation. Answer those carefully.
Some systems do offer scoring or ranking features, and some employers use them. But a score is a sorting aid for a human, not a gate. The persistent myth of a robot silently rejecting 75 per cent of applicants is not something anyone has ever demonstrated with evidence.
Formatting for clean parsing
This is the highest-value work, and it is mostly about removing things.
Layout
- Single column. Two-column layouts are the most common cause of scrambled parsing. A parser reading left to right across the page can interleave your sidebar skills with your job descriptions into nonsense.
- No text boxes or shapes. Content inside a drawing object frequently does not extract at all.
- Nothing important in headers or footers. Putting your phone number and email in the header is a classic way to submit a resume with no contact details attached.
- No tables for layout. A simple table of skills usually survives; a resume built as a nested table grid often does not.
- No text as images. Skill icons, logo graphics and headshots contribute nothing readable.
- No columns of dates aligned with tabs into a separate visual track. Keep the date on the same line as the job title and employer.
Fonts and characters
Use a standard font: Arial, Calibri, Helvetica, Georgia, Garamond. Use the normal bullet character rather than dingbats or emoji, which can extract as junk or vanish. Avoid ligature-heavy display fonts.
Headings
Parsers look for conventional section names. Use them exactly:
| Use this | Not this |
|---|---|
| Professional Experience / Work Experience | Where I've Made an Impact |
| Education | Academic Journey |
| Skills | My Toolkit |
| Certifications | Badges and Accolades |
Dates and job entries
Use one consistent format throughout, ideally month and year: "March 2022 - Present". Spelling out months, or at least using standard three-letter abbreviations, is safer than numeric formats, which are ambiguous across regions. Put employer, job title, location and dates on adjacent lines in the same order for every role.
File type
Submit a PDF unless the posting specifies otherwise, and make sure it is a text PDF rather than a scan or an exported image. If you can select and copy text from it in a PDF reader, a parser can read it. If the posting asks for .doc or .docx, comply; some older systems genuinely handle Word better.
All of our 28 resume templates are built as single-column, text-based layouts for exactly these reasons, so you can look designed without gambling on the parser.
Keywords without keyword stuffing
Since recruiters search the pool, your document needs to contain the words they will search for. That is the whole idea. It does not require anything dishonest.
Finding the right terms
Take the job posting and pull out every noun phrase that describes a skill, tool, methodology, certification or domain. Typically you will get fifteen to twenty five. Sort them into three groups: things you genuinely do well, things you have touched, and things you cannot claim. Work with the first two groups only.
Placing them
The strongest placement is inside an experience bullet, because it comes with evidence. "Rebuilt the quarterly forecast in Excel and Power BI" is worth more than a Skills line reading "Excel, Power BI", both to a human and to a keyword search.
Use a Skills section for tools that do not fit naturally into a bullet, and keep it honest and short.
Match the exact phrasing
A search for "project management" does not necessarily return a document that only says "managed projects", depending on how the search is configured. Where the posting uses a specific term, use their term at least once. Where a term has an acronym and a full form, include both on first use: "search engine optimisation (SEO)", "Chartered Institute of Personnel and Development (CIPD)". This costs nothing and covers whichever version the recruiter types.
What not to do
Do not paste the job description in white text at the bottom of the page. It is visible in the parsed text, recruiters have seen it many times, and it reads as an attempt to deceive. Do not build a wall of eighty comma-separated keywords; it dilutes the document and a human reader discounts all of it. Do not claim skills you cannot demonstrate, because the interview is where that unravels.
You can check what your document actually looks like after parsing, and which posting terms are missing, with our free ATS resume checker.
Job titles that do not match the market
Internal titles are a real problem for searchability. If your employer calls you a "Customer Happiness Ninja" and the market calls the job "Customer Success Manager", a search will not find you.
The fix is to write the recognised title and keep the official one alongside it:
Customer Success Manager (internal title: Customer Happiness Ninja)
This is accurate, searchable and does not misrepresent anything. The same approach works for level mismatches, where "Analyst II" means little outside the company.
The application form matters more than people expect
Where a system asks you to complete fields as well as upload a document, fill them in properly even when it feels duplicative. Recruiters filter on those structured fields, and an incomplete profile is easy to skip. Specifically:
- Answer knockout questions truthfully but do not disqualify yourself unnecessarily. If a salary field is mandatory and you would rather discuss it later, give a range rather than a single figure where the form allows text.
- Do not leave the "years of experience" field blank.
- If the system offers to autofill from your resume, check the result. Autofill mistakes are common and you are the only person who will see them before submission.
Tailoring per application, efficiently
Sending one document to fifty employers is the main reason people conclude the systems are rigged. A thirty-minute tailoring pass on ten roles beats an untargeted blast at fifty.
A practical routine:
- Keep a master document with everything you have done.
- Copy it, then cut it to the roles and bullets relevant to this posting.
- Reorder bullets so the ones matching the posting appear first in each job.
- Rewrite the summary line to name the target title and the two specialisations the posting emphasises.
- Update the Skills section to lead with the tools named in the posting that you genuinely use.
- Run the checker, fix gaps, submit.
Our free resume builder keeps stored versions, so step two is a duplicate-and-trim rather than a rebuild.
Questions people ask about these systems
Does a PDF really get rejected?
No. Every mainstream system in current use reads text-based PDFs without difficulty. The advice to always submit Word documents dates from an era of much weaker parsers and persists because it is repeated rather than tested. The genuine failure case is a PDF that contains an image of a document, which happens when people scan a printed copy or export from a design tool with text converted to outlines. Open your file and try to select a line of text. If you can, you are fine.
Should I put keywords in white text?
No. The text is fully visible in the parsed record, several systems flag it, and recruiters who spot it treat it as dishonesty rather than cleverness. It also does nothing useful, because a recruiter who finds you through a term you cannot support will end the conversation in the first two minutes.
Do these systems reject people for formatting alone?
Not as a rule. What happens is subtler and worse: your record is created with missing or misplaced information, so you look weaker than the person next to you in the pipeline. Nobody rejects you; you simply never surface in the search. That is why parsing accuracy, rather than any notion of an optimisation score, is the thing to spend your attention on.
Does applying early make a difference?
Often, yes, though not because of anything the software does. Recruiters commonly review in batches and start interviewing before the posting closes, so a strong candidate who applies in week one can effectively fill the shortlist. If you see a role you want, apply within the first few days rather than waiting until you have perfected the document.
Should I apply for the same role twice?
Reapplying with an improved document to the same requisition usually produces two records and looks disorganised. Applying to a different role at the same company later is normal and expected. If you genuinely submitted the wrong file, contact the recruiter and say so rather than resubmitting silently.
What the system cannot do for you
Two limits worth being realistic about.
First, a perfectly optimised document does not compensate for a genuine mismatch. If a role requires five years in a regulated environment and you have none, formatting will not close that. Spend the effort on roles where you are a credible candidate.
Second, plenty of hiring never touches the pipeline at all. Referred candidates commonly enter partway through the process or bypass the queue. Applying through the system and separately contacting someone at the organisation is not redundant; it is the version of the process that works.
A short checklist
- Single-column layout, standard font, standard bullets
- Contact details in the body, never the header
- Conventional section headings
- Consistent month-and-year dates on every role
- Recognised job titles, internal titles in brackets
- Acronyms and full forms both present
- Posting terms placed inside evidence-bearing bullets
- Text-based PDF, sensibly named
- Application form fields completed, autofill checked
- Parsed and verified before submitting
Work through that list once and the systems stop being an obstacle. What remains is the ordinary difficulty of persuading a person that you are worth an hour of their time, which is where your energy should have been going all along.
