ATS-Friendly Resume Format: Rules That Pass

An applicant tracking system is screening middleware: it ingests applications, extracts text, maps that text against the job description, and ranks candidates on match. Human recruiters then review the ranked shortlist — which means a resume that parses badly never reaches the human.

The extraction step is where most of the damage happens. Text inside tables, images, text boxes and multi-column layouts is either dropped or read out of order, turning five years of solid experience into a word salad the scorer cannot match.

PDF and DOCX are the safe formats. A PDF must be a real text PDF — generated from a word processor or a builder — not a scan or an image export. Scanned PDFs and image-only files hand the system an unreadable picture.

If a posting explicitly asks for a particular format, send that format; otherwise PDF is the standard choice because it preserves layout across devices and emails.

Parsers match section headings against known vocabulary. 'Experience', 'Education', 'Skills', 'Projects' and 'Certifications' are recognised; 'My Journey So Far' or 'Career Highlights' may fail to map, dumping those sections into an unknown bucket.

There is no creative upside in custom headings. Keep the standard labels, and use one clear hierarchy: section headings bold and larger, role titles below them, bullets beneath.

Extraction runs left to right, top to bottom. A two-column layout — skills on the left, experience on the right — gets re-assembled into interleaved lines: a skill name, then a fragment of a bullet from the other column, then the next skill.

Single-column restores the order the parser expects, and it happens to match the layout recruiters find easiest to scan. The design question disappears once you realise both audiences want the same thing.

Text placed in a table is extracted row by row, so a skills table reads as a scrambled run-on sentence in the parser's output. Charts, logos, icons and decorative graphics carry no readable text at all — they are just noise the parser skips.

Some sophisticated systems attempt to read tables, but none guarantee it. The safe position is zero tables and zero images in the body of the document.

Standard fonts — Arial, Calibri, Georgia, Times New Roman — embed cleanly into PDFs. Exotic display fonts, symbol glyphs, emojis and decorative bullets like arrows or stars can drop or garble during extraction, taking part of a sentence with them.

Keep bullets as simple hyphens or dots, avoid emojis entirely, and skip underlines and highlights that the parser may misread or ignore.

The system scores occurrences of job-description terms in your text, but not all positions weigh equally: phrases in headings, the summary and role bullets outscore a dense skills dump. Mirror the exact wording of the JD — 'agile development' stays 'agile development', not 'agile methodology'.

Repeat a critical keyword two or three times in context — summary and bullets — and it reads naturally while scoring well. Fifteen scattered buzzwords without context score lower than three exact-phrase matches with evidence.

Some systems ignore content in headers and footers, or only read the first page's header. If your name, phone or email live there, a candidate can lose their identity and every follow-up channel at once.

Place the contact block in the main body at the very top — name as a plain heading, then phone, email, city and LinkedIn as a text line. The human sees the same information, and the parser captures it.

Copy the entire resume text into a plain-text editor and read it top to bottom. If the order survives, every line reads sensibly, and each JD keyword appears in the text, most parsers will handle the file correctly.

Arjun Malhotra, a network engineer in Bangalore applying to a large IT services firm, ran this test and found his two-column skills table re-assembled into nonsense. He rebuilt the resume single-column, and the firm's portal moved his application to 'under review' within a week — a complete switch from silence.

Run through this list in under two minutes before every submission. The checklist catches the mechanical failures that quietly kill otherwise qualified applications.

The Crafex resume builder generates a single-column, text-based PDF with standard headings from the start — the parser-clean baseline described above — leaving you to focus on keywords and content.

Build once, test with the plain-text method, and keep the master file; tailoring for each application then touches only the summary and skills blocks.