ATS Resume Format: The Dos and Don'ts That Actually Affect Parsing
FitWard AI Team
You spend two hours building a sharp-looking resume in Canva. Custom fonts, two clean columns, a skill bar showing "Python — 85%." You send it out. Nothing happens. Not a single callback.
The problem almost certainly isn't your experience — it's the ATS resume format. Most ATS systems never parsed that file at all — they stored a blank profile, or worse, scrambled text that made no sense. Harvard Business School's "Hidden Workers" report found that 88% of employers believe qualified candidates are filtered out by automated systems before a human reviewer sees them — and formatting failures are where that filtering begins. To understand the full mechanism, read how ATS actually parses your file — this article focuses on the specific formatting choices that cause those failures and exactly what to use instead.
What Is an ATS Resume Format?
An ATS-friendly resume format is one that a parser can extract as clean, ordered text. It has three requirements: a text-based file (DOCX or selectable PDF), a single-column layout with no tables or text boxes, and standard section headings like "Work Experience" and "Education." Content outside these constraints may be scrambled or lost entirely.
Modern ATS platforms — Workday, Greenhouse, Lever, iCIMS — don't read your resume the way a human does. They pull the raw text layer from your file, map it to structured fields (job title, employer, dates, skills), and store that structured version in a candidate database. Modern AI-enhanced versions of these platforms add semantic matching and context analysis on top of the basic parse — but they cannot match what they cannot first extract. What you see on screen is irrelevant. What the parser extracts is everything.
This distinction matters because an ATS-friendly resume format isn't about looking plain — it's about keeping text extraction reliable. A well-formatted resume can still have clean typography and a professional appearance. The rules below target the specific choices that break extraction, not aesthetics in general.
File Format: DOCX, PDF, and What Goes Wrong
The file type you submit determines whether the ATS can extract text at all. Three formats behave very differently.
DOCX is the safest choice. Every major ATS parses it natively, and Word's XML structure makes text extraction straightforward. If a job posting doesn't specify a format, submit DOCX.
Text-based PDF works reliably when exported correctly. Open the PDF and try to highlight a sentence with your cursor. If you can select text, the file contains a readable text layer and most modern ATS platforms will parse it without trouble. Export from Word, Google Docs, or LibreOffice using "Save as PDF" — not "Print to PDF," which compresses layout into less parseable output.
Image-based PDF is invisible to ATS parsers. This is the silent failure mode. A scanned resume, a screenshot saved as PDF, or any file where text cannot be highlighted — these appear as blank pages to the parser. The system stores an empty profile.
Canva is a particularly common source of this problem. When you export a Canva design as PDF, the output is often a flattened image file, not a text document. The resume looks perfect. The ATS sees nothing. If you use a Canva template, select "PDF (Print)" or check that the text layer is selectable before submitting. When in doubt, rebuild the content in Word or Google Docs and use your Canva design as a visual reference only.
Layout and Structure Don'ts (and What to Use Instead)
Layout is where most ATS-format failures happen. The issues below are specific to how ATS parsers extract text from document structure — not style preferences.
ATS parsers extract text in document order, not visual order. A two-column layout looks like parallel columns to your eyes, but the parser reads across the full page width left-to-right, then moves down. The result: job titles from the left column appear next to bullet points from the right column, producing scrambled output. Single-column layouts eliminate this entirely.
The specific elements to avoid:
Two-column or multi-column layouts. The parser merges both columns into a single text stream. A skills section in the right column lands next to a job description in the left, creating unreadable output. Multi-column layouts are among the most consistently documented ATS parse failures across all major platforms. Use a single column throughout.
Tables used for layout. ATS parsers read table cells in document order, which means cell 1 (top-left), then cell 2 (top-right), then cell 3 (bottom-left). A two-column skills table reads as two unrelated words per line. Use a comma-separated list or a vertical bar method (Python | SQL | Tableau) instead.
Content in Word headers or footers. Many parsers skip the header/footer region entirely. Contact information placed there — your name, phone number, LinkedIn — disappears from the parsed output. Put all contact details in the document body before your first heading.
Text boxes and floating objects. Word text boxes are treated as external objects by most parsers. Content inside them is skipped. If you used a text box to frame your summary or pull out a highlight, that content likely won't be parsed. Move it into the document body as regular paragraph text.
One nuance worth knowing: most ATS guides say "never use tables." A simple two-row header table containing your name and contact info is lower risk than a Word header/footer — because the document body table will be parsed even when the header region is ignored. The rule to follow is more specific: don't use tables for multi-column layout, skills grids, or any content where text order matters. A single-row contact table in the document body is acceptable.
Fonts, Bullets, and Visual Elements
These choices have a smaller blast radius than layout issues — but they still produce parse errors in specific systems.
Fonts. Standard system fonts parse reliably: Arial, Calibri, Georgia, Verdana, Garamond, Times New Roman, Helvetica. Decorative or downloaded fonts (Montserrat, Proxima Nova, Avenir, anything from DaFont) may not be embedded in the final PDF. When the parser encounters a character it can't map, it outputs [NULL] or garbled text. The word "Profile" becomes "Pro?le." The word "office" becomes "o?ce." These aren't display artifacts — they're what's stored in the candidate database. Use boring fonts. Boring is reliable.
Skill bars and progress indicators. A graphic showing "Python — 85%" contains no text. Parsers extract text from the document layer; anything rendered as a graphic simply doesn't exist in the output. If your skills section uses rating bars, replace them with plain text: "Python (5 years)" or "Python — advanced." The information survives. The visual doesn't need to.
Before (invisible to ATS):
[graphic: skill bar labeled "Python" with 85% fill]
After (parseable):
Python — 5 years | SQL — 3 years | Tableau — intermediate
Custom bullet symbols. Standard bullets — •, -, * — parse universally. Unicode symbols like ✔, ►, ★, or → often render as boxes (□) or question marks in the database. One garbled character isn't catastrophic, but a full bullet list of "□ Managed teams" signals a formatting problem to any recruiter who reviews the parsed output.
Icons next to contact details. A phone icon before your number, an envelope icon before your email — these render as □ characters or are dropped entirely. Use plain text labels if you want labels at all: "LinkedIn:" or just present the URL directly.
Section Headings and Dates
After parsing text from your file, the ATS maps content to database fields. It looks for recognizable section labels and date patterns. Non-standard labels get filed in an "other" bucket — or dropped.
Section headings. ATS systems look for standard labels. Variations that work: "Work Experience," "Professional Experience," "Experience," "Employment History." Variations that fail: "Career History," "My Journey," "Where I've Been," "The Work." Creative naming signals nothing to the parser and may cause the entire section to be miscategorized.
The same applies to these common creative substitutions:
- "My Toolbox" → use "Skills" or "Technical Skills"
- "About Me" or "My Story" → use "Summary" or "Professional Summary"
- "How to Find Me" → use "Contact" or include contact details without a heading
Date formats. ATS parsers use employment dates to calculate total years of experience and determine whether a role is current. Ambiguous formats cause calculation errors — or the parser skips the date entirely. Use one consistent format throughout your resume: "January 2022 – March 2024," "Jan 2022 – Mar 2024," or "01/2022 – 03/2024." For current roles, write "Present."
Formats that break parsers: "Summer 2023" (no month), "Jan '22" (apostrophe confuses tokenizers), "'22 – Present," "2022 – now" (use "Present," not "now" or "current"). Inconsistency within the same resume also creates issues — if you use "01/2022" for one job and "March 2021" for another, some parsers treat the difference as a data anomaly and skip one of the dates.
How to Test Your Resume Before You Apply
Two tests take under two minutes and catch the majority of ATS-format failures.
To test whether your resume is ATS-compatible, use the plain-text copy-paste test: select all text in your resume, paste it into Notepad, and read through it. If the text appears in logical order with no missing sections, your formatting is likely safe. If text is scrambled, merged, or sections are missing, the ATS will encounter the same problem.
The copy-paste test — step by step:
- Open your resume file (Word or PDF).
- Select all text (Ctrl+A on Windows, Cmd+A on Mac).
- Copy to clipboard (Ctrl+C).
- Open Notepad (or any plain-text editor) and paste (Ctrl+V).
- Read through it: are all job titles, dates, and bullet points present? Is the order logical — name and contact first, then experience, then education? If sections are jumbled or content is missing, you have a formatting issue.
The PDF highlight test: Open your PDF in any PDF viewer. Try to highlight a single word by clicking and dragging. If you can select text, the file has a readable text layer. If clicking selects a rectangular region regardless of where you click — or you can't select at all — your PDF is image-based and won't parse.
If you want a faster alternative, FitWard AI's resume scoring checks formatting as one of its four dimensions — flagging inconsistent dates, punctuation, and bullet length — alongside completeness, impact, and clarity. You can check your ATS score free at Fitward — upload takes under a minute.
Frequently Asked Questions
Is PDF or Word better for ATS?
DOCX is the safest default because it parses natively in every major ATS. A text-based PDF — exported from Word or Google Docs — works well in modern systems like Workday and Greenhouse. The risk with PDF is accidental image export: always do the cursor-highlight test to confirm your PDF is text-based before submitting.
Can I use a two-column resume for ATS?
No. Two-column layouts cause ATS parsers to merge both columns into a single text stream, which scrambles job titles, descriptions, and dates into unreadable output. Submit a single-column version to any ATS application. Keep your visual two-column version for emailing directly to a contact who will read it as a human.
What fonts are safe for ATS?
Arial, Calibri, Georgia, Verdana, Garamond, Times New Roman, and Helvetica parse reliably across all major systems. Decorative fonts — Montserrat, Avenir, Proxima Nova — may not embed correctly in PDFs, causing characters to render as garbled text. Body text at 10–12pt, headings at 14–16pt.
Does ATS auto-reject resumes for formatting problems?
ATS doesn't auto-reject — it parses and stores what it finds. SHRM's talent acquisition research confirms ATS functions as a ranking and database tool, not a gatekeeper: applications aren't deleted, they're ranked too low to surface. The damage from a formatting failure is that your profile gets stored with scrambled or missing fields, dropping your keyword match score until you're below the threshold a recruiter will ever reach.
Can I use tables in an ATS resume?
Avoid tables for any content where reading order matters — skills grids, two-column layouts, work history. Table cells are read sequentially in document order, which scrambles multi-column content. A simple single-row table in the document body for contact information is lower risk than a Word header/footer, since headers are often skipped by parsers. The practical rule: no tables for layout.
Fitward is free to start — upload your resume and get your score in under a minute, including a formatting check that flags the specific issues covered in this guide.