A safe ATS resume format is one column, standard section names, nothing in the page header or footer, no tables or text boxes, a common font at 10 to 12 point, saved as .docx or a text-based PDF. That is the whole spec. Everything below it is the list of ways people break it without knowing.
The parser is not judging your design. It converts your file to plain text, then hunts for blocks it recognizes: contact details, experience, education, skills. Anything that survives as text in the right order gets through. Anything that doesn't gets dropped, jumbled or attached to the wrong job.
How a parser reads your file
Advertisement
Two facts explain almost every failure. The software reads text in the order it finds it in the file, not the order your eye follows. And it decides where a section starts by looking for a heading it recognizes, followed by a gap.
Resume-parsing vendor RChilli spells out the first one for two-column layouts: the parser reads all of the left-hand content, then all of the right-hand content, so a job title on the left and its description on the right come apart and the description no longer matches any employer. The same page notes that a section heading without a tab, colon or new line after it means "the parser will not pick up the whole block."
Workday, one of the largest ATS platforms, keeps its own advice to one line: results "can vary based on resume format and order of words," and for best results use resumes that don't have images or image-based styles. Order of words. That is the tell.
The 12 mistakes, and what the parser sees instead
1. Two columns. You see a tidy sidebar with skills and a main column with jobs. The parser reads the left column top to bottom, then the right, so your skills list lands between your name and your first employer, and every bullet is separated from the job it belonged to.
2. Contact details in the page header. The University of Georgia career center puts it plainly: most systems "are unable to read headers and footers," so a phone number or email placed there likely will not be scanned. RChilli says the same for .doc and .docx files. The parser sees a resume with no owner.
3. Tables for layout. A table with a job title, dates and location in three cells looks aligned on the page. RChilli's example shows what comes out: all the column headers first, then all the values, as unrelated lists. UGA's version is shorter: "Many systems are unable to pull information from tables."
4. Text boxes. A text box is a floating object, not part of the main text flow. Depending on the converter, its contents come out at the end of the document, at the start, or not at all.
5. Icons for phone, email and location. The little envelope glyph is either an image, which becomes nothing, or a symbol from an icon font, which becomes a random character. Type the word "Email:" if you want a label.
6. A photo. Both Workday and UGA say images confuse or are ignored by parsers. In the US, UK, Canada and Australia recruiters would rather not have one on file anyway.
7. Skill bars and star ratings. A five-star graphic for "Excel" parses as the word Excel with no level, or as nothing. Write the skill and, if you must, the level in words.
8. Creative section names. "Where I've Made an Impact" is not a heading a parser knows. UGA recommends sticking with "traditional headings such as education, work experience, certifications, campus involvement, and skills." Without a recognized heading the block gets filed under whatever section came before it.
9. Dates in odd formats. RChilli's parser will not treat a single year like "1998" as a job period, and a range split across two lines, 1998 on one and 1999 on the next, fails too. Use "Jan 2023 to Mar 2025" or "01/2023 to 03/2025" on one line, the same way every time.
10. Fancy or decorative fonts. Script fonts, condensed display fonts and anything downloaded from a template site can map characters unpredictably when the file is converted. UGA's list of safe choices is Calibri, Arial, Verdana or Tahoma.
11. Unusual file types or a scanned PDF. Greenhouse accepts .doc, .docx, .pdf, .rtf and .txt; a Pages file, a PNG or a "PDF" that is really a photo of a page is not text at all. If you cannot select the words in your PDF with a cursor, a parser cannot either.
12. A careless file name. "Resume_FINAL_v3 (2).pdf" parses fine, but it is the first thing the recruiter sees in the download list, and some systems display it as the document title. Use Firstname-Lastname-Resume.pdf.
The safe format spec
Copy this into the top of your working file and delete it before you save.
| Element | Safe choice |
|---|---|
| Layout | One column, top to bottom, no sidebars |
| Font | Calibri, Arial, Verdana, Tahoma or Georgia |
| Size | 10.5 to 12 point body, 14 to 16 point name |
| Margins | 0.7 to 1 inch (1.8 to 2.5 cm) on all sides |
| Section names | Summary, Experience, Education, Skills, Certifications |
| Contact block | In the body of the page, plain text, first four lines |
| Dates | Month and year, one line, same format throughout |
| Emphasis | Bold only; no boxes, shading, lines drawn as shapes |
| Bullets | Standard round or square bullets, no icons |
| File type | .docx, or PDF exported from Word or Google Docs |
| File name | Firstname-Lastname-Resume.pdf |
The 60-second plain-text self-test
Open your resume, select all, copy, and paste into a plain text editor: Notepad on Windows, TextEdit in plain-text mode on a Mac, or the character counter, which keeps only the raw text. Then read it top to bottom as if you were the parser and ask four things.
Is your name on the first line and your phone and email within the next three? If they are missing, they were in a header.
Do the jobs appear in order, with the title, employer and dates together, followed by that job's bullets? If a block of skills or a list of years appears somewhere strange, you had columns or a table.
Are there stray characters, empty squares or odd symbols where your icons and bullets were?
Do the section headings appear as words on their own lines? If "Experience" is glued to the end of the previous sentence, add a blank line after each heading.
Fix whatever failed, save, and run the test again. When the plain text reads like a sensible document, you're done with format. A word count at this stage also tells you whether you have room left for content; a one-page resume in 11 point usually holds 400 to 600 words.
PDF or Word
Use whichever the application form asks for. When it does not say, .docx is the safer default; RChilli goes as far as advising against PDF because of conversion errors with columns and images. A clean single-column PDF exported directly from Word or Google Docs parses well in modern systems and protects your layout from font substitution.
Never "print to PDF" or scan a printout. Export or save as, then run the plain-text test on the PDF itself, because that is the file the employer gets.
If you'd rather see the result than infer it, upload the final file to the ATS resume checker and read the extracted text panel. That is exactly what the parser sees.
Advertisement
FAQ
Is a two-column resume ever OK for ATS?
Only when a human is guaranteed to read it first, such as a referral emailed straight to a hiring manager. For anything uploaded to an application form, use one column. The reading-order problem RChilli describes is not fixable from your side, because you cannot control which converter the employer's system uses.
Can I keep my name in the header if the rest is in the body?
Move it. Both a major university career center and a parsing vendor say header content is skipped or treated differently, and a resume whose owner cannot be identified is the worst possible parsing failure. Put your name as the first line of the body at 14 to 16 point and it will still look like a header.
Does an ATS reject resumes automatically for bad formatting?
Usually not in the sense of a bounce message. The file is accepted, the parse is poor, and your record ends up with a blank or scrambled work history that scores badly against the posting or never surfaces in a recruiter's search. Silent failure is the reason to test before you upload.
