AccessScanRun a free scan

Guide

PDF Accessibility: A Practical Guide for Organisations Publishing Documents

A PDF that looks perfect on screen can be completely unusable with a screen reader. If the file has no tag structure, assistive technology sees a flat run of characters with no headings, no reading order, and no idea which image is decorative and which carries meaning. PDF accessibility is the work of adding that hidden structure so the document is perceivable and operable for everyone, not just sighted mouse users.

This matters legally as well as ethically. The European Accessibility Act treats documents as part of a product or service, so a non-accessible policy PDF, invoice template or product manual can put an otherwise-compliant service out of conformance. This guide covers the six things that actually move the needle: tags, reading order, alt text, headings, tables and the standards that bind them together.

What "tagged PDF" actually means

Under the hood, an accessible PDF carries a tag tree — a hidden, hierarchical structure that maps every visible element to a semantic role, much like HTML. A heading becomes an <H1> tag, a paragraph a <P>, a list an <L> with <LI> children, and a data table a <Table> with <TR>, <TH> and <TD>. Screen readers ignore the visual layout entirely and read this tree instead.

A PDF exported with no tagging is called an *untagged* PDF. To a screen reader it is often just a single run of text with no structure, or — worse — a scanned image with no real text at all. The fastest test: open the file, press Ctrl/Cmd+A to select all, and try to copy the text into a plain document. If nothing comes across, or it arrives scrambled, the document needs OCR and re-tagging before anything else.

The formal target here is PDF/UA (ISO 14289), the standard for universal accessibility in PDF. It is the document-world equivalent of meeting WCAG success criteria, and it is what auditors and procurement teams increasingly ask for by name.

Reading order: the most common silent failure

Tags give elements meaning; reading order decides the sequence in which they are announced. These are two separate things, and they disagree more often than people expect. A two-column newsletter can look right but read as the full left column, then the full right column, jumping mid-sentence across the page. Pull quotes, sidebars, headers and footers frequently land in the wrong place or get read twice.

Reading order is defined by the order of elements in the tag tree, not by their position on the page. In Adobe Acrobat Pro you inspect and fix it through the Tags panel and the Reading Order tool; in other editors it is the "content order" or "structure" view. The rule of thumb: read the document as a person would naturally read it aloud, and make the tag sequence match that.

  • Move decorative page furniture (running headers, footers, page numbers) out of the reading flow by marking it as an artifact.
  • Check that captions read immediately after the figure they describe, not three paragraphs later.
  • Verify multi-column and table-heavy pages explicitly — these are where reading order breaks most often.

Alt text, headings and tables done right

Alt text for images

Every image that conveys information needs a text alternative on its <Figure> tag. Describe the *meaning*, not the appearance: "Bar chart showing 2024 revenue up 18% on 2023" beats "chart." Purely decorative images — dividers, background flourishes, a logo already named in adjacent text — should be marked as artifacts so screen readers skip them entirely.

Headings that nest

Visual styling is not structure. Text that merely looks big and bold is invisible to assistive technology unless it carries an <H1><H6> tag. Use one <H1> for the document title, then nest logically — don't skip from <H1> to <H3>. Screen reader users navigate long PDFs by jumping between headings, exactly as a sighted reader skims, so a correct heading outline is often the single biggest usability win in a long report.

Tables with real header cells

A data table is only accessible if its header cells are tagged as <TH> (not <TD>) and given a scope of Row or Column. Scope tells the screen reader that "€4,200" belongs to the "Q3" row and the "Net revenue" column, so the value is announced with its context instead of as a bare number. Never use tables purely for visual layout — and never fake a table with tab stops or spaces, which produces meaningless output.

Why EN 301 549 and the EAA cover documents too

A frequent and costly misconception is that accessibility law stops at the web page. It does not. The European Accessibility Act (Directive (EU) 2019/882), with requirements applying from 28 June 2025, covers a range of products and services, and those requirements reach the documents those services rely on — think bank statements, e-book files, ticket confirmations, contracts and product manuals delivered as PDFs.

The technical baseline is set by the European standard EN 301 549, which references WCAG 2.2 Level A and AA. WCAG's four principles — Perceivable, Operable, Understandable and Robust — apply to non-web documents as well as web pages, which is precisely what a PDF is. Colour contrast rules carry over too: body text needs a contrast ratio of at least 4.5:1, and large text (18pt, or 14pt bold) at least 3:1 — the same thresholds as on the web (WCAG 1.4.3).

Enforcement is handled by national authorities, so fines vary by member state and can reach tens of thousands of euros, but the practical risk is simpler: a single inaccessible PDF buried in an otherwise-compliant checkout or onboarding flow can break conformance for the whole service. Microenterprises providing services (fewer than 10 staff and under €2,000,000 annual turnover) are largely exempt for those services, but most organisations publishing customer-facing documents are not.

A practical workflow for accessible PDFs

Accessibility is far cheaper to build in at the source than to retrofit. Most remediation pain comes from fixing exported files one by one when the source document was never structured properly.

  • Start in the source. Use real heading styles, list styles and table tools in Word, InDesign or Google Docs — structure exports into tags automatically when you do.
  • Set the document language and title in the file properties so screen readers announce the right pronunciation and show a meaningful title in the tab.
  • Tag a logical reading order and confirm it with the structure view, not just the visual preview.
  • Add alt text and mark artifacts before export, not after.
  • Run a checker. Acrobat's built-in Accessibility Check (and the free PAC checker for PDF/UA) catches missing tags, language and alt text. Automated tools find only a portion of issues — always finish with a manual screen-reader pass.
  • Prefer HTML where you can. If a document is really just web content in disguise, publishing it as an accessible web page is almost always easier to maintain than a perfectly tagged PDF.

If you are auditing a whole site and unsure where documents sit on your priority list, our accessibility checklist puts PDFs in the context of the wider WCAG and EAA picture, and you can run a free accessibility scan of your site's pages to see where the gaps are before you commit budget.

Check your site against AccessScan

See your issues ranked by impact in seconds — free.

Run a free accessibility scan

FAQ

Are PDFs really covered by the European Accessibility Act?

Yes. The EAA (Directive (EU) 2019/882), with requirements applying from 28 June 2025, covers products and services and the documents they depend on. A PDF delivered as part of a covered service — a statement, contract, ticket or manual — is expected to meet the WCAG 2.2 Level A and AA baseline referenced by EN 301 549, just like a web page.

How can I quickly tell if a PDF is accessible?

Two fast checks: try selecting and copying the text (if nothing copies, it is likely a scanned image with no real text), and run Acrobat's Accessibility Check or the free PAC checker. These flag missing tags, reading order, language and alt text. Then do a short manual pass with a screen reader, since automated tools catch only a portion of issues.

Is it better to make an accessible PDF or publish HTML instead?

If the content is essentially web content, accessible HTML is usually easier to build, maintain and keep compliant than a perfectly tagged PDF. Reserve PDFs for documents that genuinely need a fixed, printable layout — and tag those properly. Many organisations offer an HTML version alongside the PDF.

Do I need to fix old PDFs already on my site?

Prioritise by use. Documents inside a covered service flow — onboarding, billing, contracts — and high-traffic files should be remediated or replaced first. Truly archival documents may be lower risk, but check your member state's rules. Rebuilding from an accessible source is usually faster than remediating a bad export by hand.

More guides