Accessible PDF Tagging: Tag Tree Step by Step

ENGLISH VERSION

What is PDF tagging and why should you care?

Imagine you receive a book with no table of contents, no chapter headings, and no page numbers. You could read it, sure, but finding anything specific would be a nightmare. Now imagine that same experience, but without being able to see the document at all. That is exactly what a blind person experiences when they open an untagged PDF with a screen reader like JAWS or NVDA.

Accessible PDF tagging means adding an invisible semantic structure to the document: headings, paragraphs, lists, tables, figures with alternative text, and a logical reading order. This structure is organized in what we call the tag tree, which is essentially the skeleton that allows assistive technologies to correctly interpret the content.

Personally, I believe PDF tagging is one of the most undervalued skills in digital accessibility. We all talk about WCAG and accessible websites, but how many organizations publish hundreds of untagged PDFs every year? According to a study by the PDF/UA Foundation, over 95% of PDF documents on the internet do not comply with the PDF/UA standard (ISO 14289). That is a staggering figure.

The PDF/UA standard: the foundation of everything

Have you heard of PDF/UA? The acronym stands for “PDF/Universal Accessibility” and refers to the international standard ISO 14289, which defines the technical requirements for making a PDF accessible. It is to PDF what WCAG is to the web.

PDF/UA-1 vs PDF/UA-2: which one to apply?

There are currently two versions of the standard:

If you are just starting out, my recommendation is to work with PDF/UA-1. The tools are more mature, documentation is more abundant, and most validators fully support it. PDF/UA-2 is the future, but we are still in the transition period.

Tools for accessible PDF tagging

What tools can you use to create and remediate a tag tree? Here are the main ones:

Adobe Acrobat Pro: the industry standard

Adobe Acrobat Pro remains the most widely used tool for PDF tagging. Its Tags panel allows you to create, edit, and reorganize the tag tree visually. It includes the “Make Accessible” accessibility wizard that automates part of the process, although personally I believe you should never blindly trust automation—you always need manual review.

• Tags panel: direct access to the complete tag tree.
• Reading Order panel: lets you define and reorder document zones.
• Accessibility Check: runs an automated check with a detailed report.
• Preflight with PDF/UA profile: advanced technical validation against the standard.

PAC 2024: the free validator

PAC (PDF Accessibility Checker) is a free tool developed by the PDF/UA Foundation that validates documents against the PDF/UA standard. The 2024 version includes significant improvements in structure error detection and initial support for PDF/UA-2. If you can only use one validation tool, make it PAC.

CommonLook: for professional environments

CommonLook PDF offers a complete remediation environment that integrates as a plugin in Adobe Acrobat. It is particularly powerful for complex documents with nested tables and multi-page layouts. CommonLook Global Access also enables remediation directly from Microsoft Office.

The tag tree step by step: practical guide

Let us get to the point. How do you build a correct tag tree? Think of the tree as a structural table of contents: every element in the document has its hierarchical place.

Step 1: Basic document structure

Every tag tree begins with the root tag <Document>. Inside it, you organize content into semantic sections:

• <Document> → root tag of the tree
• <Part> or <Sect> → sections or parts of the document
• <H1>, <H2>, <H3>… → hierarchical headings
• <P> → text paragraphs
• <L> and <LI> → lists and list items
• <Table>, <TR>, <TH>, <TD> → table structure
• <Figure> → images and graphics with alt text

Step 2: Logical reading order

Reading order is critical. A PDF can look perfect visually but have a chaotic reading order. Have you ever copied text from a PDF and it came out all jumbled? That is exactly the problem a screen reader encounters.

In Adobe Acrobat Pro, use the Reading Order panel to define zones and their sequence. Golden rule: the order in the tag tree must match the logical reading order of the document, from top to bottom, left to right.

Step 3: Tagging tables correctly

Tables are the most complex element to tag. A poorly tagged table is completely unreadable with a screen reader. Here is the correct structure:

Key tip: always assign the Scope attribute to header cells (<TH>). This indicates whether the header applies to the column (Column) or the row (Row). Without this attribute, screen readers cannot associate data with their headers.

Step 4: Images and figures

Every image must be wrapped in a <Figure> tag with a descriptive alt text. If the image is decorative, mark the artifact as “Background” so the screen reader ignores it. The rule? If the image conveys information, it needs alt text. If it is pure decoration, it is an artifact.

Step 5: Artifacts

Artifacts are elements that are not part of the actual document content: page numbers, repeating headers, decorative lines, backgrounds. In the tag tree, these elements must NOT have tags; they must be marked as artifacts so assistive technology ignores them.

Common PDF tagging mistakes

What are the errors I see time and time again in my professional experience? Here are the 5 most common:

• 1. Headings that skip levels: jumping from H1 to H3 without H2. The hierarchy must be consecutive.
• 2. Tables without header tags: using <TD> instead of <TH> for header cells.
• 3. Images without alt text: the <Figure> tag exists but the /Alt attribute is missing.
• 4. Decorative content not marked as artifact: page numbers, repeated logos the screen reader reads unnecessarily.
• 5. Incorrect reading order: the tag tree does not reflect the visual order of the document.

Validation: how do I know if my PDF is accessible?

Tagging alone is not enough. You need to validate. Here is the validation workflow I recommend:

• Step 1: Run the Adobe Acrobat Pro accessibility check. Fix all errors it detects.
• Step 2: Validate with PAC 2024 against PDF/UA-1. PAC is stricter than Acrobat in many areas.
• Step 3: Manual testing with a screen reader (NVDA is free). Navigate the entire document and verify the order and structure make sense.
• Step 4: Human review by an accessibility specialist. Automation does not catch everything.