9 min read · updated 10 September 2026
Almost every mid-size and large employer receives applications through an applicant tracking system, or ATS. The system is not an artificial gatekeeper that rejects people for sport — it is a database. Understanding what that database does with your file explains nearly every piece of resume advice you have ever been given, and lets you ignore the advice that is simply folklore.
This guide walks through the three stages a resume passes through inside a typical ATS: parsing, indexing and ranking. At each stage I describe what actually breaks, why it breaks, and the specific change that fixes it.
Stage one: parsing turns your document into fields
When you upload a file, the ATS does not store it as a picture of a page. It runs a parser that reads the text layer and tries to map fragments of that text onto structured fields: name, email, phone, employer, job title, start date, end date, school, degree, skills.
Parsers are pattern matchers. They look for a line that resembles a heading ("Experience", "Education"), then treat everything until the next heading as belonging to that block. Inside a block they look for date ranges to split one role from the next. This is why unusual section names such as "Where I've Made An Impact" cause damage that has nothing to do with keywords — the parser cannot find the boundary, so three jobs collapse into one blob of text with no employer attached.
- Use conventional headings: Summary, Skills, Experience, Education, Certifications.
- Keep one role per block with employer, title, location and a date range on the same or adjacent line.
- Write dates in a consistent, unambiguous format such as "Mar 2023 – Present".
- Export a PDF with a real text layer, never a scan or an exported image.
Why tables, columns and text boxes cause silent damage
A two-column layout looks orderly to a human because the eye follows visual grouping. A parser follows the reading order stored in the file. Depending on how the document was produced, that order can zig-zag between columns, interleaving your skills list with your job bullets so that neither reads as a coherent field.
Text boxes and headers/footers are worse: in some export paths their content sits outside the main content stream, so a phone number placed in a header can vanish entirely. Graphical icons instead of labels have the same effect — a small envelope glyph next to your email carries no text, so a parser that expects the word "Email" nearby loses its anchor.
None of this means a designed resume is useless. It means the version you upload to a job portal should be the plain single-column one, and the designed version belongs in an email attachment or a portfolio link, where a person opens it directly.
Stage two: indexing decides which words are searchable
Once fields are extracted, the text is indexed so recruiters can search it. Most systems index the literal words you wrote, sometimes with light stemming — "forecasting" matching "forecast" — but rarely with real synonym intelligence. "Management reporting" and "MIS reporting" are two different search terms to a database even though they describe the same work.
The practical consequence: write the phrase the employer used, and where two names for the same thing are both common, include both once. Spell out an acronym on first use with the short form in brackets — "Financial Planning & Analysis (FP&A)" — so either search term finds you.
What does not work is a hidden block of white keyword text at the bottom of the page. Recruiters see it the moment they open the parsed text view, and it reads as an attempt to game the process.
Stage three: ranking is mostly recruiter search, not AI judgement
Some systems produce a match score against the job description. Many recruiters barely look at it. What they actually do is search the pipeline for two or three must-have terms, then read the shortlist in date order.
That changes your priorities. The top third of the first page carries almost all the weight, because it is what appears in the preview pane. A four-line summary naming your function, years of experience, core systems and one measurable result does more for you than a beautifully written bullet on page two.
- Put your target job title near the top, matching the language of the posting.
- Name the systems you actually operate — SAP, Power BI, SQL, Excel — as concrete nouns.
- Lead each experience block with the outcome, then the mechanism: "Cut monthly close reporting from 5 days to 2 by rebuilding the extract in Power Query".
- Quantify with whatever number you can defend: value, volume, headcount, days saved, error rate.
Writing bullets a human wants to read
A parser tolerates weak writing; a hiring manager does not. The most common failure is describing duties rather than results. "Responsible for monthly reporting" tells a reader only that the job existed. "Owned the monthly OpEx pack for 14 cost centres, flagging variances above 5% with drivers before the review meeting" tells them the scale, the standard and the judgement involved.
Aim for three to five bullets per recent role and one or two for older ones. Cut adjectives, keep numbers, and make sure every bullet could survive the question "how would you prove that in an interview?"
A pre-submission checklist
Before you upload anything, run these checks. They take five minutes and catch most of what causes an application to disappear.
- Open the PDF and try to select your name with the cursor. If you cannot, it is an image.
- Copy the whole document into a plain text editor. Read what comes out — that is roughly what the ATS sees.
- Confirm every role has employer, title and dates in the extracted text.
- Check that the job title in the posting appears somewhere in your top section.
- Keep the filename human: Firstname-Lastname-Role.pdf.
Frequently asked questions
- Is a one-page resume required?
- No. One page suits early-career applicants; two pages are normal after about five years. What matters is that the first third of page one carries your strongest, most relevant material.
- PDF or Word?
- PDF with a selectable text layer is safest and preserves layout. Use Word only when the portal explicitly asks for .doc or .docx.
- Do photos hurt an application?
- In most of the US, UK and India-based corporate hiring, a photo adds no value and takes space that a summary could use. In markets where photos are conventional, keep it small and outside a text box.
- Should I tailor my resume for every job?
- Tailor the summary line, the job title and the skills list — that is ten minutes of work and covers most of the benefit. Rewriting every bullet for every application is rarely worth the time.
Put it into practice
Builds a single-column PDF with a real text layer and scores your draft against a job description.
Open the free ATS Resume BuilderWritten by Syed Danish Arafat, a financial analyst working in FP&A, reporting and finance automation. Shared for educational purposes — review anything here against your own circumstances before acting on it.