The Complete Guide to WCAG 2.1 Compliance for Publishers

Why Accessibility Matters for Publishers Now

The European Accessibility Act (EAA) came into full effect in June 2025, requiring all digital products and services sold in the EU to meet accessibility standards. For publishers, this means every ePub, PDF, and digital publication must comply with WCAG 2.1 AA criteria.

The scope is broader than many realize. It covers not just the content itself, but the platforms used to deliver it — websites, apps, and reading systems. Publishers who sell into the EU market face potential fines and market exclusion for non-compliance.

In the US, Section 508 and ADA requirements continue to expand. The DOJ's 2024 web accessibility rule under Title II signals a clear direction: digital content accessibility is no longer optional.

The Four WCAG Principles: POUR

WCAG 2.1 is organized around four principles — Perceivable, Operable, Understandable, and Robust (POUR). Each principle contains guidelines and success criteria that apply differently to publishing content.

Perceivable means providing text alternatives for non-text content (alt text for images, captions for figures), ensuring content can be presented in different ways without losing meaning, and meeting color contrast requirements (4.5:1 for normal text, 3:1 for large text).

Operable means all functionality must be available via keyboard, users must have enough time to read content, and navigation must be consistent and predictable.

Understandable requires readable text (language declarations, abbreviation expansions), predictable behavior, and input assistance for forms.

Robust means content must work with current and future assistive technologies — proper semantic markup, valid code, and standard ARIA attributes.

Publisher-Specific Accessibility Requirements

For digital publications specifically, accessibility requirements go beyond general web standards. ePub 3 publications must include proper reading order (logical sequence that matches visual order), semantic structure (headings, lists, tables properly marked up), page navigation (page-list nav for print-equivalent references), language declarations (both document-level and inline for multilingual content), and ONIX accessibility metadata.

ONIX 3.0 introduced specific accessibility metadata fields that publishers must populate: accessMode (textual, visual, auditory), accessModeSufficient (what modes are sufficient to consume the content), accessibilityFeature (specific features like alternativeText, longDescription), accessibilityHazard (or the absence of hazards), and conformsTo (which WCAG level the content meets).

AI-Powered Accessibility at Scale

The challenge for publishers with large backlists is scale. Manually auditing and remediating thousands of titles is prohibitively expensive and slow. This is where AI-powered tools make accessibility achievable.

AI-powered alt text generation can analyze images in context — not just describing what the image contains, but explaining its relevance to the surrounding text. A chart in a medical journal needs different alt text than the same chart in a patient education brochure.

Automated metadata tagging can infer and apply ONIX accessibility metadata across entire catalogs, with human review focused on edge cases rather than every single title.

Structure detection algorithms can identify and fix heading hierarchy issues, missing table headers, and reading order problems across legacy ePub and PDF files.

Getting Started: A Practical Roadmap

Start with an audit of your current catalog. How many titles are ePub 3? How many have proper accessibility metadata? What percentage of images have alt text? This baseline assessment determines the scope of work.

Prioritize by impact: new publications should be born accessible from day one. High-traffic backlist titles should be remediated next. Legacy content that rarely gets accessed can be addressed last.

Build accessibility into your production workflow. Every new manuscript should go through accessibility checking before publication — not after. Tools like EPUBCheck and Ace by DAISY should be part of your standard QA pipeline.

Consider a technology partner for the heavy lifting. Converting thousands of legacy titles is exactly the kind of task where AI-powered automation delivers massive efficiency gains — 80% faster than manual remediation with equal or better quality.