Almost every rejection story we hear in coaching sessions begins the same way: "I applied to sixty roles and heard nothing." Then we open the resume and find a two-column layout, a photo, a skills chart with dots, and the work history inside a text box.
An applicant tracking system is not judging that design. It is trying to convert your file into fields — name, email, employer, title, dates, skills — and a layout like that makes half of those fields unreadable. What the recruiter then searches, your resume cannot answer.
How parsing actually works
Most ATS platforms read a document top to bottom, left to right, looking for recognisable section headings and date patterns. Then they store the result as structured data. Recruiters search that data, not your PDF.
Three consequences follow:
- Anything inside an image, chart, icon or text box may not exist. Skills expressed as five filled dots are, to the system, nothing at all.
- Multi-column layouts get flattened. Your right-hand skills column can end up interleaved with job descriptions, producing a garbled record.
- Non-standard headings get skipped. "My Journey" is not "Experience". "What I'm Great At" is not "Skills".
The structure that parses cleanly
Use a single column, standard headings, and a plain .docx or text-based PDF exported from a word processor.
- Header — name, city, phone with country code, email, LinkedIn URL as text. No photo, no logos, no header/footer bars (many parsers ignore true headers and footers entirely).
- Summary — three lines. Role, years, domain, and the two or three capabilities you want to be searched for.
- Skills — a plain comma-separated list, grouped by type. Write the words a recruiter would type: both "PowerBI" and "Power BI", both "HRMS" and "HRIS" if both are true.
- Experience — reverse chronological. One line per role:
Title | Company | City | Mon YYYY – Mon YYYY. Then three to five bullets. - Education and certifications — degree, institution, year. Certifications with issuing body.
Bullets that survive both the machine and the human
The parser wants keywords; the recruiter wants evidence. One bullet can serve both if you write action, object, method, result:
Reduced offer-drop rate from 22% to 9% across 40 tech roles by restructuring the offer-stage follow-up and introducing a 48-hour manager touchpoint.
Compare it with what we usually receive:
Responsible for end-to-end recruitment activities and coordination.
The second sentence contains no searchable specificity and no proof. If you have no numbers, use scale instead: team size, requisition count, geographies, tools, cycle times.
Keyword matching without keyword stuffing
Open the job description, list the nouns that repeat, and check honestly whether each one appears in your resume where it is genuinely true. That is tailoring. Pasting a white-text keyword block at the bottom is not — recruiters see it the moment they open the parsed record, and it reads as dishonesty.
Two words on titles: if your internal designation is unusual ("Talent Wizard"), write the market-standard title and keep the internal one in brackets. Systems and humans both search the standard.
Formatting details people get wrong
- One file, two pages maximum for under fifteen years of experience.
- Name the file
Firstname-Lastname-Role.pdf. Recruiters download hundreds of files a week. - Dates in a consistent
Mon YYYYformat. Ranges without a year confuse both parsers and interviewers. - No tables for work history. Tables are the most common cause of scrambled parsing.
- Employment gaps: state them in one neutral line rather than leaving an unexplained hole.
A 20-minute self-audit
Save your resume as plain text and reopen it. Everything that vanished or scrambled is exactly what an ATS may lose. Then search that plain text for the five keywords from your target role. If you cannot find them, the system will not either.
The goal is not to defeat the software. It is to be legible to it — so that the human on the other side gets to read the version of you that actually deserves the interview.