Here's what nobody tells you: you can write a genuinely great resume with real keywords, real accomplishments, and real experience that matches the job, and still end up buried in the candidate pile because of your font. Not because recruiters think it looks bad. Because the ATS parser literally could not read it. The word 'profile' became 'pro□le'. The word 'office' vanished. 'PowerPoint' got garbled into 'PowemPoint' because two letters blurred together. And your keyword match score dropped without a single error message, warning, or clue that anything had gone wrong.
This is what researchers call silent degradation, and it's one of the most frustrating problems in the modern job search. You did everything right, and a typographic quirk you never even thought about quietly tanked your chances. Let's fix that.
99.7%
Of Recruiters Use ATS
From a 2024 survey of 384 recruiters
97.8%
Of Fortune 500 Companies
Use ATS software as of 2025
23%
Of ATS Failures Are Formatting-Related
Including font issues, per Enhancv 2025 study
What's Actually Happening Inside the Parser
ATS software doesn't 'see' your resume the way you do. It extracts raw text, maps it to fields, and generates a score based on keyword matches, completeness, and structure. The parser works mechanically, often at scale across thousands of applications. When it hits a font it doesn't recognize, it doesn't pause and ask for help. It substitutes, guesses, and moves on.
The Technical Terms Worth Knowing
- Ligature
- A single merged glyph representing two characters, like 'fi' or 'fl'. In typesetting software (LaTeX, InDesign, Acrobat), these are encoded as one character. When an ATS parser extracts text, it maps ligature glyph IDs back to characters using a lookup table called a ToUnicode cmap. If that mapping is missing or wrong, the ligature gets extracted as '\x00' (an empty value) or dropped entirely.
- Glyph Substitution
- When a font isn't installed on the parser server, the ATS replaces it with a generic fallback. The spacing, kerning, and character encoding all shift. Words can merge, split, or render as □□□ (tofu boxes), which are unreadable placeholder characters.
- Font Fallback
- The ATS parser's default behavior when it encounters an unknown font: substitute a generic serif or sans-serif and hope for the best. The results are often not great.
The Three Ways Fonts Break Your Parse
Character Spacing Breaks
Kerning data is font-specific. When the parser substitutes a fallback font, the spacing between characters shifts. 'Senior Engineer' can become 'SeniorEngineer', a single merged word that matches nothing in the recruiter's keyword filter.
Ligatures Map to Empty Characters
PDF text extraction tools map 'fi', 'ff', and 'fl' ligatures to '\x00' (an empty Unicode value) when the font's ToUnicode cmap is incorrect. Words like 'office', 'profile', 'efficient', and 'conflict' are common resume keywords that get silently dropped or corrupted. This is confirmed at the developer level in PDF parsing libraries.
Glyph Substitution Produces Nonsense
Decorative or display fonts that aren't installed on the parser server trigger full glyph substitution. Characters become tofu boxes (□□□), random replacements, or disappear. In one documented case, a script font used for headings caused the ATS to parse the heading as a blank field, so the candidate's job title simply didn't exist in the system.
The Specific Fonts That Wreck Your Score
Some of these fonts look genuinely great. Pacifico is friendly and modern. Lobster has personality. Montserrat is clean and contemporary. But looking great on screen and parsing correctly inside an ATS are two entirely different standards, and on corporate hiring software, the parser wins every time.
High-Risk Fonts and Their Failure Modes
| Font | Risk Level | What Breaks |
|---|---|---|
| Brush Script | Critical | Cursive characters merge during parsing; entire words become unreadable |
| Pacifico | Critical | Not installed on parser servers; font substitution scrambles the document |
| Lobster | Critical | Decorative ligatures fail to map; characters drop or duplicate |
| Impact | High | Compressed letterforms cause adjacent-word merging on 30%+ of parses |
| Papyrus | High | Rough textures sometimes export as graphic shapes, not selectable text |
| Montserrat | Moderate | SAP SuccessFactors and older Taleo fall back to plain substitution, breaking layout |
What Ligature Failure Actually Looks Like
Skills: Proficient in PowerPoint, efficient with Excel, experienced managing conflict resolution across offices [What you typed: looks perfect in your PDF viewer]
Skills: Pro□cient in PowemPoint, e□□icient with Excel, experienced managing con□ict resolution across o□□ices [What the ATS parser may actually extract. Critical keywords gone.]
Even the 'Safe' Classics Aren't Always Safe
Times New Roman gets recommended everywhere. It feels like the responsible choice: traditional, professional, universally known. And for most things, it's fine. But here's what the research actually shows: Times New Roman has a 5-8% parse error rate on older Taleo and iCIMS configurations. Its thin serifs occasionally mis-parse on these systems, producing 'chum' instead of 'churn' or '$4,2M' instead of '$4.2M'. That numerical comma alone is enough to drop a quantified achievement from the ATS's numerical value extractor. The achievement you worked hardest to write, gone.
The Canva Problem (And the LaTeX Trap)
Two Risky Resume-Building Approaches
Canva Templates
Independent testing of 50 popular Canva resume templates found that 72% failed basic ATS parsing, not because of content, but because of how Canva builds the document. Canva's rendering engine sometimes embeds text as graphic elements within layered structures. Even when text appears selectable in your PDF viewer, the underlying structure may not map to the logical reading order that an ATS expects. Enhancv's 2026 benchmark found Canva templates scoring anywhere from 52% to 92%, with variance driven almost entirely by sidebar use and embedded graphics.
LaTeX-Generated PDFs
LaTeX is popular among technical candidates and academics who genuinely know their way around document formatting. This makes the problem particularly insidious: LaTeX-generated PDFs are specifically vulnerable to ligature and encoding issues. A LaTeX PDF can parse perfectly if it's single-column with clean text extraction, but it can fail badly due to ligature handling, font encoding, or layout complexity. Technically sophisticated candidates are the least likely to question their document quality, and the most likely to be blindsided.
Special Characters: The Bonus Risk You're Probably Ignoring
Fonts aren't the only encoding problem. Special characters carry the same risk. Underlines can cause lowercase p's, q's, and g's to parse as different characters on some ATS configurations. Accented letters like é and ñ, non-standard bullet points like ★ or ◆, and decorative symbols can be misread or cause entire sections to drop. Worth noting: the word résumé itself is not ATS-friendly because of the accent marks. Use 'resume' in the plain text of your document.
ATS-Friendly vs. ATS-Risky Choices
Do This
Avoid This
Calibri, Arial, Georgia: clean parse across all major ATS platforms
Script or display fonts (Brush Script, Pacifico, Lobster): critical parsing failures
Standard bullet points (•): universally recognized
Decorative symbols (★, ◆) as bullet points: misread or dropped
Plain 'resume' without accent marks in body text
Accented characters (é, ñ): may cause omission of entire sections
Single-column layout with standard section headers
Canva templates with sidebars or layered graphics: 72% fail basic ATS parsing
PDF exported from Microsoft Word or Google Docs
Google Fonts or DaFont downloads: may not be installed on parser servers
Do This
Calibri, Arial, Georgia: clean parse across all major ATS platforms
Avoid This
Script or display fonts (Brush Script, Pacifico, Lobster): critical parsing failures
Do This
Standard bullet points (•): universally recognized
Avoid This
Decorative symbols (★, ◆) as bullet points: misread or dropped
Do This
Plain 'resume' without accent marks in body text
Avoid This
Accented characters (é, ñ): may cause omission of entire sections
Do This
Single-column layout with standard section headers
Avoid This
Canva templates with sidebars or layered graphics: 72% fail basic ATS parsing
Do This
PDF exported from Microsoft Word or Google Docs
Avoid This
Google Fonts or DaFont downloads: may not be installed on parser servers
The 60-Second DIY Test You Can Run Right Now
You don't need any special software to check if your resume has font-related parsing issues. This test replicates what an ATS parser does when it extracts text from your PDF, and it takes less than a minute.
The Copy-Paste Parse Test
Open your resume PDF
Open the PDF in any standard viewer: Adobe Acrobat, Preview, Chrome, wherever you normally view it.
Select all text
Press Ctrl+A (Windows) or Cmd+A (Mac) to select all the text in the document.
Copy and paste into a plain text editor
Open Notepad (Windows) or TextEdit in plain text mode (Mac) and paste. Strip all formatting.
Read what you actually see
If words are merged, characters are missing, sections are out of order, or you see □□□ placeholder boxes, that is exactly what the ATS parser is extracting from your resume. Any keyword that looks garbled here is a keyword the system cannot match against the job description.
Fix it, then test again
Switch to a clean font like Calibri or Arial, rebuild in Microsoft Word or Google Docs, export a fresh PDF, and run the test again. Clean paste = clean parse.
A Quick Note on the '75% Rejection' Stat
You've probably seen the claim that '75% of resumes are rejected by ATS before a human sees them.' That number is almost certainly fabricated. Multiple sources trace it to a 2012 sales pitch by a defunct startup called Preptel, with no published methodology or sample size. A 2025 study of 25 US recruiters across 10+ ATS platforms found that 92% do not configure auto-rejection rules based on resume content. ATS doesn't reject your resume. It scores it. Recruiters then decide which scores are worth reviewing. The real problem isn't automatic rejection. It's that a font-related parsing failure drops your score, and a lower score means a recruiter is less likely to ever open your file. The outcome can feel similar. The mechanism is very different, and understanding the difference is what lets you actually fix it.
Key Takeaways
- Fonts cause silent ATS scoring failures with no error message, no notification, and no clue it happened.
- Ligatures in PDF-exported files (fi, fl, ff) can map to empty characters, silently dropping keywords like 'office', 'profile', and 'efficient'.
- Calibri and Arial have under 2% parse error rates across all five major ATS platforms. They're the safest choices.
- Canva templates fail basic ATS parsing 72% of the time; LaTeX PDFs carry encoding risks even for technically sophisticated users.
- The copy-paste test is free and shows you exactly what an ATS parser extracts from your document.
- ATS scores resumes. Recruiters do the filtering. Fixing your font improves your score and your visibility, not just your design.
- Times New Roman has a 5-8% parse error rate on older Taleo and iCIMS configurations. It's not as safe as its reputation suggests.
Frequently Asked Questions
Does ATS automatically reject resumes with bad fonts?
No. This is a common misconception. ATS software scores resumes and helps recruiters prioritize candidates. A font-related parsing failure lowers your score and pushes you down the candidate pile. Recruiters then decide who to review based on those scores. The effect can feel like rejection, but the mechanism is about visibility, not automatic filtering.
Is Times New Roman safe for ATS?
Mostly, but not universally. Times New Roman has a 5-8% parse error rate on older Taleo and iCIMS configurations. Its thin serifs can cause mis-parses that corrupt quantified achievements. Calibri or Arial are safer choices with documented sub-2% error rates across all five major platforms.
Are Google Fonts safe to use on a resume?
Risky. Google Fonts may not be installed on parser servers. If they're not embedded correctly in your PDF, the parser falls back to a generic font, which can break character spacing, ligatures, and encoding. Stick to system fonts with broad compatibility.
What's the safest way to create an ATS-friendly resume?
Build in Microsoft Word or Google Docs, use Calibri or Arial at 10-12pt, export as PDF, and run the copy-paste test to verify clean text extraction. Avoid tables, columns, graphics, and any font downloaded from third-party sites.
Why do ligatures cause keyword failures?
In PDF files, ligatures like 'fi' and 'fl' are stored as single glyphs with a lookup table (ToUnicode cmap) that maps them back to their component characters. When that mapping is missing or incorrect, common with decorative and typesetting fonts, the extraction tool returns an empty value instead of the characters. Words like 'profile', 'efficient', and 'conflict' disappear from the parsed text.