Blog

Why PDFs Break Accessibility (And What to Do Instead)

TestParty
TestParty
February 1, 2026

PDFs optimize for print fidelity, not adaptive accessibility. The format was designed to preserve exact visual layout across devices—ensuring a document looks the same whether viewed on a MacBook or printed on a laser jet. But accessibility requires the opposite: content that reflows, adapts, and exposes semantic structure to assistive technology. These goals are fundamentally in tension.

A PDF can be tagged and still fail real assistive technology usage. Tagging adds semantic information—headings, lists, reading order—but the underlying document remains a collection of positioned text boxes. Screen readers attempt to linearize this visual arrangement into a logical reading sequence, often with poor results. Two-column layouts become scrambled. Tables lose their header associations. Forms lack proper labels. The "tagged but unusable" PDF is depressingly common.

If information must be accessible, HTML is usually the safest default. HTML was built semantics-first: headings, landmarks, form labels, and document structure are native to the format. Content reflows automatically. Testing and remediation are straightforward. When organizations publish critical information exclusively as PDFs—policies, applications, reports—they create barriers for the 70+ million Americans with disabilities who rely on assistive technology.


Key Takeaways

PDFs create accessibility barriers that are difficult and sometimes impossible to fully remediate.

  • Print fidelity vs. accessibility – PDFs preserve visual layout; accessibility requires semantic structure and content reflow—opposing goals
  • Semantics are optional in PDFs – Unlike HTML where headings and structure are native, PDFs require explicit tagging that's often missing or incorrect
  • Reading order breaks constantly – Multi-column layouts, sidebars, and footnotes create scrambled sequences for screen reader users
  • "Tagged PDF" isn't a guarantee – Tagging is necessary but not sufficient; poor tags produce experiences as bad as no tags
  • HTML-first publishing is the solution – Default to web pages; offer PDFs only as optional print alternatives, not primary access points

What PDFs Were Built For

Understanding PDF's design goals explains why accessibility is so difficult to achieve.

The Original Purpose

PDF stands for Portable Document Format. Adobe created it in 1993 to solve a specific problem: ensuring documents look identical regardless of what software or device displays them. A legal contract should render the same whether opened on Windows, Mac, or printed on paper.

This required:

  • Fixed positioning – Every element placed at exact coordinates
  • Embedded fonts – Typography preserved exactly
  • Print-ready output – What you see is what you print
  • Cross-platform consistency – No layout surprises

These goals served their purpose well. PDFs excel at preserving complex visual layouts, legal documents, and print-ready materials.

The Accessibility Conflict

Accessibility requires nearly the opposite:

PDF PriorityAccessibility Need
Fixed layoutResponsive reflow (content adapts to screen size, zoom, user preferences)
Visual positioningSemantic structure (headings, lists, landmarks that AT can navigate)
Print fidelityAdaptability (content can be reformatted for different needs)
Page-basedContinuous flow (no artificial page breaks interrupting content)

When you optimize for visual consistency, you sacrifice the flexibility that makes content accessible.


Why Semantics Are Optional (and Why That's Fatal)

In HTML, semantic structure is built into the format. In PDFs, it's an afterthought.

HTML: Semantics Are Native

When you write HTML, semantics come automatically:

<h1>Annual Report 2024</h1>
<nav aria-label="Table of contents">
  <ul>
    <li><a href="#section1">Financial Summary</a></li>
  </ul>
</nav>
<main>
  <section id="section1">
    <h2>Financial Summary</h2>
    <p>Revenue increased by 15%...</p>
  </section>
</main>

A screen reader immediately knows: this is a heading level 1, here's navigation, here's main content, here's a subsection. The structure is intrinsic.

PDF: Positioned Text Boxes

A PDF is fundamentally different. At its core, a PDF says:

  • Put the text "Annual Report 2024" at position (50, 700) in 24pt Times Bold
  • Put the text "Financial Summary" at position (50, 650) in 18pt Times Bold
  • Put the text "Revenue increased by 15%..." at position (50, 600) in 12pt Times

The PDF doesn't inherently know that "Annual Report 2024" is a heading or that "Financial Summary" is a subheading. It only knows where text boxes go.

Tagging Attempts to Add Semantics

PDF tagging adds a parallel structure declaring that certain elements are headings, paragraphs, lists, etc. But this structure is:

  • Often missing – Many PDFs are never tagged
  • Frequently incorrect – Automated tagging tools guess wrong
  • Hard to verify – Requires specialized tools and expertise
  • Easily broken – Edits to the PDF may not update tags

The fundamental problem: PDF starts with visual layout and tries to retrofit semantics. HTML starts with semantics and adds visual styling.


The Reading Order Problem

Multi-column layouts are standard in print design. They're catastrophic for PDF accessibility.

How Screen Readers Navigate PDFs

Screen readers attempt to read PDF content in a linear sequence. Without proper tagging, they follow the order in which content was placed in the file—which may have nothing to do with logical reading order.

Consider a two-column newsletter:

Visual layout:

| Column A (continued) | Column B          |
| More text from       | New article:      |
| previous article     | Headline here     |
| ...                  | Article text...   |

What a screen reader might announce: "Column A continued... More text from... New article... Headline here... previous article..."

The reader jumps between columns unpredictably because the underlying PDF structure doesn't encode "read column A top to bottom, then column B."

Common Layout Failures

Layout PatternAccessibility Failure
Two-column textContent scrambles between columns
SidebarsSidebar intrudes mid-sentence into main content
Pull quotesQuote text repeats or appears out of context
FootnotesFootnote content mixes with body text
Tables spanning pagesTable structure breaks across page boundaries
Headers/footersPage numbers and headers inject into content flow

Every sophisticated print layout creates reading order challenges in PDFs.


Images of Text: The Scanned PDF Trap

Scanned documents represent the most inaccessible PDFs possible.

The Scanning Problem

Many PDFs are created by scanning paper documents. The result is a PDF containing images—photographs of pages—not text. A screen reader encountering this sees... nothing. Or rather, it sees "image" with no content.

Even if OCR (Optical Character Recognition) is applied:

  • OCR accuracy varies – Handwriting, unusual fonts, poor scan quality all reduce accuracy
  • "Searchable" isn't "accessible" – OCR text layers often have errors invisible to sighted users but announced by screen readers
  • Structure isn't captured – OCR extracts text but not headings, lists, or tables
  • Images within scans disappear – Charts, diagrams, and photos in scanned documents have no descriptions

Common Scanned Document Sources

Organizations frequently scan:

  • Historical documents
  • Signed contracts
  • Paper forms
  • Old reports and policies
  • External documents received on paper

Each becomes an accessibility barrier when published as the PDF of record.


Forms and Tables: Where PDFs Fail Most

Interactive content in PDFs frequently lacks basic accessibility.

PDF Form Failures

PDF forms require:

  • Labeled fields – Each input needs an associated label
  • Tab order – Fields should be navigable in logical sequence
  • Error handling – Validation messages need to be accessible
  • Instructions – Required fields and format hints must be clear

Most PDF forms fail multiple requirements:

RequirementCommon PDF Failure
Field labelsVisually adjacent text not programmatically associated
Tab orderJumps randomly based on when fields were added
Required fieldsVisual asterisks without programmatic indication
Error messagesPop-ups that screen readers may miss entirely

Users with disabilities often cannot complete PDF forms independently—a serious barrier when the form is a job application, benefits enrollment, or legal document.

PDF Table Failures

Data tables require header cell associations so screen readers can announce "Revenue: $1.2M" rather than just "$1.2M" in isolation.

PDF table issues:

  • Headers not marked – First row looks like headers visually but isn't tagged
  • Complex tables impossible – Merged cells, nested headers, spanning cells rarely tag correctly
  • Multi-page tables break – Tables continuing across pages lose header associations
  • Layout tables misidentified – Visual layout using table structure confuses AT

For data-heavy documents like financial reports, these table failures make content incomprehensible.


Even "Tagged PDF" Isn't Enough

Organizations sometimes believe "tagged PDF" equals "accessible PDF." This is wrong.

Tagging Is Necessary but Not Sufficient

Proper tags are required—but quality matters enormously:

  • Incorrect heading levels – H1 used for everything, or headings skipped
  • Missing alt text – Images tagged as figures but with no descriptions
  • Wrong language metadata – Mispronunciation throughout
  • Poor reading order – Tags exist but sequence is wrong
  • Artifacts not marked – Decorative elements announced as content

A PDF with bad tags can be worse than one with no tags—screen readers follow the tags even when they're wrong, creating a confusing experience that contradicts visual layout.

What "Accessible PDF" Actually Requires

Truly accessible PDFs need:

RequirementImplementation
Proper tag structureHeadings, lists, paragraphs correctly identified and nested
Reading orderLogical sequence matching visual flow
Alternative textMeaningful descriptions for all informative images
Language metadataCorrect language specified (including spans of foreign text)
Table structureHeaders associated with data cells
Form field labelsProgrammatic association between labels and inputs
Link textMeaningful link descriptions
Document titleMeaningful title in document properties

Creating and maintaining all of these in a complex document requires significant expertise and effort.


The Solution: HTML-First Publishing

For most content, HTML provides superior accessibility with less effort.

Default to HTML

Publish as web pages first:

  • Policies and procedures – Web pages, not PDF downloads
  • Reports and data – Interactive HTML with PDF as optional download
  • Help documentation – Web content that can be searched and bookmarked
  • Forms and applications – Web forms that work with assistive technology
  • Product information – Web pages that update easily

HTML provides:

BenefitWhy It Matters
Native semanticsHeadings, lists, forms are intrinsic
Responsive reflowContent adapts to screen size and zoom
Easy testingStandard accessibility tools work
Simple remediationFix code, deploy—done
SEO benefitsSearch engines prefer HTML
Easy updatesNo file regeneration needed
Universal accessWorks on any device with a browser

When PDFs Are Necessary

Some situations genuinely require PDFs:

  • Legal documents requiring exact reproduction – Contracts with specific formatting
  • Regulatory submissions – Forms that must match official templates
  • Archival records – Historical documents preserved in original form
  • Print-intended materials – Brochures where print layout matters

For these cases, provide:

  1. HTML equivalent whenever possible
  2. Accessible PDF following PDF/UA standards
  3. Alternative formats on request
  4. Contact information for accessibility assistance

If You Must Use PDFs: Minimum Standards

When PDFs are unavoidable, apply rigorous accessibility standards.

PDF Accessibility Checklist

Before publishing any PDF:

  • [ ] Document is tagged (not just "searchable")
  • [ ] Reading order is logical (test by tabbing through)
  • [ ] Heading hierarchy is correct (H1, H2, H3 in sequence)
  • [ ] All images have alt text (or are marked decorative)
  • [ ] Tables have header associations
  • [ ] Forms have labeled fields with logical tab order
  • [ ] Language is correctly specified
  • [ ] Document title is meaningful
  • [ ] Links have descriptive text
  • [ ] Color is not sole means of conveying information

Validation Approaches

Test PDFs with:

  • Automated checkers – Adobe Acrobat's accessibility checker, PAC (PDF Accessibility Checker)
  • Screen reader testing – Actual navigation with NVDA, JAWS, or VoiceOver
  • Keyboard testing – Tab through forms and links
  • PDF/UA validation – Conformance to the PDF accessibility standard

When Converting to HTML

If you convert PDF content to HTML for accessibility:

  • Ensure the HTML is itself accessible
  • Verify headings, lists, and structure transferred correctly
  • Add alt text to images
  • Test forms and interactive elements
  • Apply the same CI/CD and monitoring practices as other web content

Source code remediation applies here too: if you're generating HTML from templates, those templates need to produce accessible output.


FAQ

Can AI make PDFs accessible automatically?

AI can assist—generating alt text suggestions, identifying potential heading structure—but it cannot guarantee accessible PDFs. AI-based remediation tools face the same fundamental challenge: PDF structure is visual, not semantic. AI can improve tagging but cannot overcome inherent format limitations. For critical documents, expert review and assistive technology testing remain necessary.

Is PDF/UA compliance sufficient for accessibility?

PDF/UA (Universal Accessibility) is the ISO standard for accessible PDFs, and conformance is meaningful. However, PDF/UA compliance doesn't guarantee good usability—a document can be technically conformant while still having poor reading order or unhelpful alt text. Think of PDF/UA as a minimum baseline, not a guarantee of quality user experience.

What about PDF forms that legally require signatures?

For documents requiring signatures, consider hybrid approaches: web form for data entry with a PDF generated for signature, or electronic signature solutions that work with accessible web forms. Many organizations have moved to e-signature platforms that provide better accessibility than PDF-based signing workflows.

Why do governments still publish PDFs?

Legacy systems, existing workflows, and legal requirements for document preservation contribute to PDF prevalence in government. However, federal accessibility requirements (Section 508) apply to PDFs, and agencies are increasingly required to provide accessible alternatives. Best practice is HTML primary, PDF as downloadable alternative.

Should we remediate existing PDF archives?

Prioritize by usage and criticality. High-traffic documents (current policies, active forms, frequently accessed reports) deserve immediate attention. Historical archives with low access can be remediated on-request. Consider bulk HTML conversion for content that will remain active, with legacy PDFs available for archival purposes only.

How do we handle PDFs from external sources?

You're responsible for accessibility of content you publish, even if you didn't create it. Options include: requesting accessible versions from the source, converting to HTML, providing alternative formats, or including an accessibility statement explaining limitations and offering assistance. The key is ensuring users can access the information somehow.


Internal Links

External Sources


This article was written by TestParty's editorial team with AI assistance. All statistics and claims have been verified against primary sources. Last updated: January 2026.

Stay informed

Accessibility insights delivered
straight to your inbox.

Contact Us

Automate the software work for accessibility compliance, end-to-end.

Empowering businesses with seamless digital accessibility solutions—simple, inclusive, effective.

Book a Demo