You see a designed page. The system sees a text file it has to rebuild. When a resume arrives, an applicant tracking system runs an extraction step first: it pulls the characters off the page, drops them into fixed fields like name, employer, title, dates, and skills, and only then shows a recruiter something readable. Every judgement about your format is made in that first step, before a person has seen your work.
That is why an ATS friendly resume format is not about looks. A two-column layout can look sharp and still come back with a phone number stranded in the middle of a sentence. A single-column file with a text box behind it can lose an entire section in silence — the recruiter sees a short resume, not a broken one. Nothing warns you. The extraction simply returns less than you wrote.
Text extraction reads the page in the order the text objects were drawn, and it reads left to right across the page. So position matters more than style. A file built in one column, top to bottom, with normal headings and no containers, gives the parser the same order you read. Anything else asks that parser to guess, and a guess costs you a field.
There is a second, smaller trap before format even starts: the file must carry a text layer at all. A scanned PDF is a picture of a resume. It has no characters to pull, so extraction returns an empty string. Export a text-based PDF from Word or Google Docs, upload the DOCX itself, or paste the text. If you are not sure which one you have, open the file and try to select a single word with the mouse. If you cannot select it, no parser can either.