What is a website accessibility audit?
A website accessibility audit reviews whether people with different disabilities can find, understand and operate a website. It combines automated checks, code inspection, representative manual testing and user-task review.
Accessibility is not a single switch. A page can have valid HTML and still be hard to navigate with a keyboard, confusing when zoomed, missing a useful image description or impossible to complete with a screen reader. Test the experience as well as the markup.
What automated tools can and cannot check
Automated tools are valuable for finding repeated patterns at scale. They can flag missing document titles, some heading and landmark problems, missing image alternatives, certain contrast failures, broken links and invalid attributes.
They cannot reliably decide whether:
- an alternative text describes the purpose of an image;
- a link label makes sense out of context;
- a keyboard user can understand where focus has moved;
- a form error explains how to recover;
- a complex interaction is predictable and usable;
- the content itself is clear for the people who need it.
Record automated results as leads for investigation, not as a pass or fail verdict for the whole website.
Choose representative pages and journeys
Testing only the homepage gives a false sense of coverage. Choose pages from each important template and include the journeys people need to complete.
| Sample | Why it matters | Example check |
|---|---|---|
| Homepage and navigation | Shared header, footer, menus and landmarks affect every visitor. | Move through the main routes with a keyboard and zoomed view. |
| Content and landing pages | Heading structure, images, links and readable content vary by template. | Find a section, understand its purpose and follow the next action. |
| Forms and search | Labels, errors, status messages and focus determine whether tasks can be completed. | Submit valid and invalid data without a mouse. |
| Commerce or account flow | Repeated steps and dynamic updates create higher task risk. | Complete a representative journey and recover from an error. |
| Dialogs and interactive components | Focus, names, roles and state changes are often missed by page scans. | Open, use and close each component with keyboard and assistive technology. |
Include a small number of high-risk pages even when their template is unusual. Explain the sampling method so another reviewer can repeat it.
Check titles and language
Every page should identify itself clearly. Check that the document title is present, descriptive and distinct, and that the visible heading agrees with the page's purpose.
Set the correct language for the document and for genuine language changes within the content. Clear language, short paragraphs, descriptive labels and an explained abbreviation help many people, not only users of assistive technology.
Do not rely on a title such as “Page” or “Untitled”. It gives people navigating browser tabs, history or a screen reader no useful orientation.
Check headings and structure
Headings provide a map through the page. Use one clear main heading where appropriate, then organise sections in a logical order. Do not choose a heading level because it has the right visual size; style the correct semantic level instead.
- make headings describe the section that follows;
- avoid empty headings and headings used only for decoration;
- do not skip levels simply to create visual spacing;
- give repeated regions useful landmarks and labels;
- check the structure with a heading or landmark navigation view.
Check images and alternatives
Choose the alternative based on the image's role, not its file name or appearance.
| Image type | Useful treatment |
|---|---|
| Informative | Describe the information a person needs from the image. |
| Functional | Describe the action of a linked or button image, not just what it looks like. |
| Decorative | Use an empty alternative when the image adds no information and is hidden from assistive technology. |
| Complex | Provide a concise alternative and a nearby full explanation, data table or equivalent. |
| Text in an image | Make the same meaningful text available as real text wherever possible. |
Check icons, charts, logos, product swatches and images inside links. A technically present alt attribute can still be unhelpful. Start with useful image alternatives and a navigable heading structure when checking page-level accessibility.
Check colour, text, zoom and reflow
Colour must not be the only way information is communicated. Error, selected, required and successful states need text, shape, position or another perceivable cue as well as colour.
- check foreground and background contrast for text and meaningful controls;
- check links and focus states against their surroundings;
- zoom the page and increase text size without losing content or controls;
- test narrow viewports and reflow without unexpected horizontal scrolling;
- check text spacing, line height and containers that clip enlarged text.
Test real content, not only a component with short placeholder text. Long names, validation messages and translated copy expose failures quickly.
Check forms, errors and status messages
Every input needs a clear label and an understandable relationship with its instructions. Required fields, permitted formats and examples should be available before submission rather than revealed only after an error.
Submit the form with missing, invalid and corrected values. Check that:
- errors identify the field and explain how to fix it;
- focus moves to the most useful place without trapping the user;
- the invalid state is conveyed without colour alone;
- success, loading and saved states are announced where needed;
- dynamic content does not unexpectedly move the user or erase their work.
Check media and downloads
Give video meaningful captions and provide transcripts or equivalent text where appropriate. Make audio controls usable from the keyboard and avoid unexpected autoplay, especially when sound begins without an explicit action.
Label downloads with their format and purpose, make documents accessible in their own right and do not hide essential instructions inside an inaccessible file. Test embedded players and third-party components as part of the actual journey.
Prioritise and report accessibility findings
Describe the user impact and the conditions needed to reproduce it. Include the URL or component, steps, expected result, observed result, evidence, likely owner and a suggested repair.
| Priority | Typical finding | Response |
|---|---|---|
| Critical | A key task cannot be completed with a keyboard or assistive technology. | Assign immediately and provide an accessible route while fixing. |
| High | A shared form, navigation component or repeated content pattern blocks many people. | Repair the shared cause and retest representative journeys. |
| Medium | A repeated clarity, contrast or structure issue affects a section. | Schedule a component or template correction. |
| Low | An isolated improvement with limited task impact. | Include it in planned maintenance and track the decision. |
State the test environment and scope. An audit report is more trustworthy when it says what was tested and what remains outside the sample.
Run an accessibility audit with Clearcrawl
- 01Crawl the public site.
Collect the URLs, response status, titles, headings, image alternatives and internal links.
- 02Identify templates.
Group pages so repeated components and representative layouts are visible.
- 03Filter repeated patterns.
Find title, heading, alt-text and broken-link issues that may have a shared cause.
- 04Choose a manual sample.
Include important templates, journeys, dynamic components and high-risk pages.
- 05Run specialist and manual tests.
Check keyboard, focus, zoom, contrast, assistive technology and user tasks.
- 06Record user impact.
Give each finding evidence, an owner, a priority and a clear repair.
- 07Fix shared causes and verify.
Recrawl the affected patterns and repeat the manual task after the release.
Clearcrawl helps reveal sitewide patterns and choose a representative sample. It does not replace specialist accessibility testing, manual review or feedback from people who use assistive technology.
Common accessibility audit mistakes
Automated results cover only some requirements and cannot judge every interaction.
Forms, account flows, dialogs and content templates often contain the real barriers.
An empty or generic description can be present and still fail to convey the image's purpose.
Keyboard users need a clear indication of where they are on the page.
Owners need to know who is blocked, how to reproduce it and what a useful repair looks like.
Frequently asked questions
What is a website accessibility audit?
A website accessibility audit reviews whether people with different disabilities can find, understand and operate a website. It combines automated checks, code inspection, representative manual testing and user-task review. It is an evidence-gathering exercise, not a guarantee that every user or requirement has been covered.
Can an automated tool complete an accessibility audit?
No. Automated tools can identify repeatable patterns such as missing alternative text, some contrast failures and structural problems. They cannot reliably judge whether an image description is useful, whether a keyboard journey makes sense or whether an error message helps a person recover.
What is the difference between WCAG A, AA and AAA?
WCAG conformance levels group success criteria by increasing breadth and strictness. Level A covers the most fundamental barriers, AA adds commonly expected requirements, and AAA is the most demanding level. The appropriate target depends on the product, audience, policy and applicable requirements; do not treat a level as a complete accessibility guarantee.
How often should accessibility be audited?
Test accessibility before launch, after changes to shared components or important journeys, and whenever evidence shows a barrier. A scheduled review can catch drift, but continuous checks during development and manual testing after meaningful releases are more effective than one annual inspection.
Can a crawler help with accessibility?
Yes. A crawler can reveal repeated patterns across pages, including missing titles, heading issues, image alternative gaps and broken internal links. It helps choose representative pages and find shared causes, but specialist tools and manual testing are still needed to assess the experience.
ClearcrawlFind patterns, then test the experience
Map templates, find repeated issues, build a representative manual sample, fix shared causes and verify code and user task.
7 free days. See what Clearcrawl can find, then choose the plan that fits your site.