Testing and Auditing Web Accessibility: A Comprehensive Approach

Testing is where accessibility theory meets practice. Without systematic testing, accessibility requirements remain aspirations rather than achievements. Effective accessibility testing combines automated scanning, manual expert review, and user testing with assistive technology to create a comprehensive picture of your digital experience's accessibility.

Why Testing Matters

Automated tools are an essential starting point, but they can only detect a fraction of accessibility issues — industry research consistently suggests that automated testing catches roughly 30 to 40 percent of WCAG violations. Automated tools can determine whether an image has an alt attribute, but they cannot determine whether that alt text is meaningful. They can measure contrast ratios, but they cannot evaluate whether the reading order is logical or whether a custom widget communicates its state correctly to a screen reader.

This is why a layered approach is necessary. Each layer of testing catches issues that the other layers miss.

Layer 1: Automated Scanning

Automated accessibility testing tools scan your pages against a set of rules derived from WCAG success criteria and flag violations. They are fast, repeatable, and excellent for catching systematic issues across large sites. Use them as your first line of defense and for ongoing monitoring.

axe DevTools is one of the most widely used accessibility testing engines. Available as a browser extension and as a library that integrates into development workflows, axe tests against a comprehensive set of rules and is designed to minimize false positives. It powers many other accessibility testing tools.

WAVE from WebAIM provides a visual overlay that highlights accessibility issues directly on the page. It is particularly useful for non-technical stakeholders because it shows issues in context rather than in a separate report. Available as a browser extension and as an online tool.

Google Lighthouse is built into Chrome DevTools and includes an accessibility audit category. Its scope is more limited than dedicated accessibility tools, but its convenience and zero setup make it a good quick check during development.

Pa11y is an open-source command-line tool that is well suited for integration into continuous integration pipelines. It can test individual pages or crawl entire sites and output results in various formats.

IBM Equal Access Accessibility Checker is a free browser extension that uses IBM's rule engine and provides guided evaluation including a keyboard checker mode.

None of these tools should be used as the sole method of accessibility evaluation. They complement manual testing — they do not replace it.

Layer 2: Manual Expert Testing

Manual testing involves a trained evaluator systematically reviewing your site against WCAG success criteria using a combination of techniques.

Keyboard testing is the most fundamental manual test. Navigate your entire site using only the keyboard. Tab through every page, interact with every control, open and close every menu and dialog. Verify that focus order is logical, that focus is always visible, that no keyboard traps exist, and that all functionality is available without a mouse.

Zoom testing involves enlarging the page to 200% and 400% and verifying that no content is lost, no functionality breaks, and no horizontal scrolling is required at 320 CSS pixels width.

Screen reader testing involves navigating your site using one or more screen readers to verify that all content is announced correctly, that headings and landmarks create a navigable structure, that form labels are properly associated, that dynamic content changes are announced, and that custom widgets communicate their state.

The three most common screen readers are JAWS (commercial, Windows), NVDA (free, Windows), and VoiceOver (built into macOS and iOS). Each has different strengths and behaviors, so testing with at least two is recommended. TalkBack is the primary screen reader on Android devices.

Cognitive walkthrough evaluates the site from the perspective of users with cognitive disabilities. Is the content clear and unambiguous? Are instructions explicit? Is navigation predictable? Can users recover from errors? Is the cognitive load reasonable?

Layer 3: User Testing with Assistive Technology

The most revealing form of accessibility testing involves real users with disabilities using their own assistive technologies. Automated tools and expert reviews can identify technical failures, but only real users can reveal practical usability barriers — the interactions that technically pass WCAG criteria but still create confusion or frustration.

Recruit participants who represent a range of disabilities and assistive technology use. Include screen reader users, keyboard-only users, users with motor impairments who may use switch access or voice control, users with low vision who rely on magnification, and users with cognitive disabilities. Observe them completing key tasks on your site and document the barriers they encounter.

The Audit Process

A formal accessibility audit typically follows a structured process: define scope, run automated scanning, conduct manual expert review, perform user testing, document all findings with severity ratings and remediation recommendations, then remediate starting with the highest-impact problems and retest after fixes.

Accessibility Statements

Under the European Accessibility Act and many other regulations, organizations are required to publish an accessibility statement. This document should describe your conformance status, list any known accessibility limitations, explain how users can report accessibility problems, and provide contact information for accessibility inquiries. It should be easily findable from every page of your site.

Ongoing Monitoring

Accessibility is not a one-time project. New content, design changes, feature additions, and third-party integrations can all introduce accessibility regressions. Integrate automated accessibility testing into your CI/CD pipeline, conduct periodic manual reviews, and establish a process for receiving and responding to accessibility feedback from users.