ATS

How to Format a CV for ATS in 2026: Complete Recruiter Guide

99% of Fortune 500 employers and most recruitment agencies use Applicant Tracking Systems. This guide covers the 12 rules of ATS-safe CV formatting in 2026 — what's changed, what hasn't, and what to fix when a CV fails.
FormaCV Editorial

How to format a CV for ATS in 2026: a recruiter's complete guide

99% of Fortune 500 employers and the vast majority of recruitment agencies use Applicant Tracking Systems (ATS) to ingest candidate CVs. When the ATS can't parse a CV correctly, the candidate doesn't show up in the right searches — and even an excellent candidate gets dropped before a human ever reads them.

For recruitment agencies, this is doubly painful: a poorly-formatted candidate CV submitted to your client's ATS reflects on your agency, not the candidate.

This guide covers exactly what ATS-safe means in 2026 — the rules that have changed, the rules that haven't, and what to fix when a CV is failing.

What "ATS-safe" actually means

An ATS is essentially a parser plus a database. When you submit a CV:

  1. The ATS reads the file (PDF, DOCX, etc.)
  2. Extracts structured data: name, contact, work experience, education, skills, etc.
  3. Maps that data into the system's database
  4. Indexes it for keyword search

ATS-safe = a CV the parser reads correctly. ATS-unsafe = a CV where the parser misses key fields, gets the structure wrong, or fails to ingest at all.

This is different from "ATS-optimized," which is a candidate-facing concept (matching keywords from a job description). For recruiters submitting candidates to clients, ATS-safe is the priority — the candidate's content is set, you just need it to parse cleanly into the client's system.

The 12 rules of ATS-safe CV formatting in 2026

1. Single-column layout, always

Multi-column layouts are the single biggest cause of ATS parse failure. Parsers read left-to-right, top-to-bottom. A two-column CV with sidebar (skills, contact) on the left and content on the right gets read as one mashed-up paragraph in 60% of ATS systems we've tested.

Rule: Single column for everything except optional contact info at the very top.

2. Standard system fonts only

Safe fonts in 2026: Arial, Calibri, Helvetica, Times New Roman, Georgia. Body text 10–11pt. Headers 12–14pt.

Avoid: Custom fonts (won't render on the parser server), decorative fonts, anything below 9pt or above 14pt for body.

3. PDF or DOCX — preferably DOCX

DOCX parses cleanest in 2026 — modern parsers handle it natively. PDF works but adds an OCR step in some ATS, which introduces noise.

Avoid: Image-based PDFs (scans), PDF forms, .pages files, .odt files, RTF.

4. Standard section headers

ATS parsers look for specific section header keywords:

  • "Work Experience" / "Professional Experience" / "Experience" — all parse
  • "Career History" / "Employment" — parse in most modern ATS
  • "What I've Done" / "My Journey" — fail in most parsers
  • "Education" / "Academic Background" — both parse
  • "Skills" / "Core Competencies" / "Technical Skills" — all parse

Rule: Stick to recognizable English section names.

5. Reverse-chronological work experience

Most-recent role first. ATS parsers expect this order and use date math to compute "years of experience" from the first listed role to the last.

Functional / skills-based CVs that omit dates entirely break experience-year calculations in most ATS.

6. Dates in standard format

Best: "May 2021 – Present" or "May 2021 – August 2023" Acceptable: "May '21 – Present" Avoid: "May 21 – present", "5/21–present", "Now", or just years ("2021–2023")

7. No icons as section labels

Some modern templates use small icons (an envelope for contact, a graduation cap for education) instead of text labels. Parsers can't read icons. The whole section gets misclassified.

Rule: Always have a text label in addition to any icon.

8. Skills as a clean comma-separated list (not infographics)

Skills sections that use bar charts, percentage rings, or "5/5 stars" graphics don't parse. Plain comma-separated lists or bullet points parse cleanly.

Best: A "Technical Skills:" section with comma-separated lists grouped by category (Languages, Frameworks, Cloud, Tools, etc.)

9. Contact info as text (not in a header/footer)

Many parsers ignore content in Word document headers and footers. Contact info there often gets dropped.

Rule: Contact info as the first line of body text.

10. No tables for layout

Tables are fine for actual tabular data (project lists, certifications). Tables used to create multi-column layouts break parsers.

11. No text boxes

Text boxes get dropped or rendered out of order in many parsers.

12. File name with candidate name

Some ATS use filename as a fallback when name parsing fails. "John_Smith_CV.docx" parses as name "John Smith" if the parser can't find name in the body.

Common ATS systems and their quirks

Bullhorn

  • Generally robust DOCX parser
  • PDF parsing weaker — converts to plain text via OCR which loses structure
  • Skill section parsing benefits from explicit "Skills:" header
  • LinkedIn export parsing is hit-and-miss; reformat first

JobAdder

  • Modern parser, handles both DOCX and PDF well
  • Date parsing is strict — non-standard formats fail
  • Multiple email addresses in one CV: only the first parses; put the primary email at the top

Vincere

  • Good DOCX support, weaker PDF
  • Skills section needs explicit header
  • "Years of experience" calculated from first role; missing-dates kills the calculation

Workday (client-side ATS your agency submits to)

  • Most strict parser of the major ATS
  • Multi-column layouts almost always fail
  • Tables for layout always fail
  • Section header phrasing is whitelisted — use standard names

Greenhouse (client-side)

  • Modern, robust parser
  • Handles DOCX and PDF reliably
  • More forgiving on creative formatting than Workday

iCIMS (client-side)

  • Similar to Bullhorn — solid DOCX, weaker PDF
  • Sensitive to non-standard date formats

SmartRecruiters (client-side)

  • Handles most modern formats well
  • Sensitive to image-based content (treats as missing data)

How to test if your CV is ATS-safe

Three approaches:

  1. Copy and paste test: Open the CV in Word/Reader → Select All → Copy → Paste into a plain text editor. If the result is in logical reading order with all sections clearly labeled, the CV will parse. If it's mangled or sections are out of order, the CV will fail in many parsers.

  2. Parse via your own ATS: Upload the CV to your ATS as a test candidate. Check whether name, email, phone, work history dates, and skills all populated correctly.

  3. Use a parsing-test tool: Tools like Jobscan or Resume Worded run a CV through a parser and report what got extracted. Free tier usually sufficient for spot checks.

What to fix when a CV fails

If you're submitting a candidate and the client's ATS isn't pulling the right data:

  1. Re-export the source. If the original is a PDF, ask for the DOCX version.
  2. Check for multi-column layout. Convert to single column.
  3. Check for tables. Replace tables-as-layout with plain text.
  4. Check for icons-only section labels. Add text labels.
  5. Standardize date format. Use "Month YYYY – Month YYYY".
  6. Reformat into your agency template. Most agency templates are already ATS-safe by design.

For agencies handling 50+ CVs/week, manual reformatting doesn't scale. Tools like FormaCV, HireAra, Allsorter, or Daxtra Styler reformat any input into an ATS-safe branded template in 60 seconds. See our comparison of CV formatting tools for recruitment agencies.

Special cases

LinkedIn export CVs

LinkedIn's PDF export uses a layout that breaks most ATS parsers (multi-column, decorative section labels). Treat LinkedIn exports as a starting point — reformat into your agency template before submission.

Scanned / photographed CVs

Scanned CVs require OCR. OCR introduces error rates of 2-5% on standard scans, much higher on photographed CVs from a phone. After OCR, the text needs structural reformatting because the document structure is lost in the scan.

Multi-language CVs

ATS parsers are usually English-trained. A French or German CV often parses with mangled section labels in English-language ATS. For multi-language pipelines, format two versions: original language for record, English for client submission.

CVs from screening calls

If your "candidate CV" is a screening call transcript, no ATS parser will accept it. Restructure into a proper CV first. Most CV formatters can't do this — FormaCV is built specifically for transcript-to-CV conversion. See our call transcript to CV guide.

What's changed since 2020

Three things to update from older guidance:

  1. The "one page" rule is dead. That was always candidate-facing advice. For recruiter submissions, the right length is whatever the role requires (often 2–3 pages for senior, 4+ for executive search). Modern ATS parse multi-page CVs without issue.

  2. Tables of contents and decorative timelines are penalized harder than ever. Modern ATS parsers are more lenient than 2020 parsers in some ways but stricter on visual decoration. Plain text wins.

  3. Photo handling has changed. EU practice still includes photos; UK/US practice rarely does. Modern parsers ignore photos cleanly when they're standalone images. Photos embedded inside layout tables sometimes corrupt the table parse.

FAQ

What's the safest file format for ATS submission?

DOCX is most universally parsed cleanly. PDF works but adds risk in legacy ATS. Avoid all other formats.

Should CVs include a photo?

Depends on geography. EU continental practice often includes photo. UK and US practice rarely does. The photo itself doesn't affect ATS parsing if it's a standalone image; what causes issues is photos embedded in layout tables or backgrounds.

Do I need to use specific keywords from the job description?

For recruiter submissions, no — the candidate's content is what it is. For candidate-side optimization, yes (different conversation). What matters for recruiters is that the CV's actual content parses correctly into the client's ATS.

How long should a CV be in 2026?

Whatever the role requires. The "one-page rule" was always candidate-facing. For executive search, 4-8 pages is normal. For mid-level, 2-3 pages. Length doesn't break ATS parsing.

Do AI-generated CVs parse correctly?

If they're well-formatted (single column, standard sections, no decoration), yes. AI generation doesn't inherently break parsing. Decorative templates break parsing whether human or AI-authored.

How can I batch-fix 100 CVs that all fail my client's ATS?

Manual reformatting takes 25-45 minutes per CV. AI CV formatters (FormaCV, HireAra, Allsorter, Daxtra Styler) reformat into ATS-safe templates in under 60 seconds per CV. For batch work, dedicated tools save 95%+ of the time.

My candidate insists on a creative two-column "designer" CV. What do I do?

Submit the formatted single-column version to the ATS for parsing. Optionally include the candidate's design version as a separate attachment for the human reviewer. Two-column CVs work for human eyes; they break ATS parsers.

Are PDF text and PDF image the same to an ATS?

No. Text-based PDFs (where you can copy-paste text from the document) parse with most ATS. Image-based PDFs (scans, photos) require OCR which introduces errors. Always provide text-based PDFs when possible.

Try FormaCV

If you're spending more than 5 hours/week reformatting CVs into ATS-safe templates, start a 7-day FormaCV free trial. No credit card. Format any input — PDF, DOCX, LinkedIn export, scanned doc, screening call transcript — into your branded, ATS-safe template in 60 seconds.

ATSCV formattingBullhornJobAdderVincereWorkday

Try FormaCV free for 7 days

Format any input — PDF, DOCX, LinkedIn export, or screening call transcript — into your branded, ATS-safe template in 60 seconds. No credit card required.