Use this web accessibility testing checklist to catch common problems before release. It combines automated scans with keyboard-only checks, focus and form testing, visual review, reduced-motion checks, and a practical screen-reader smoke test that developers can repeat for every page, component, and redesign.
Overview
Accessibility testing is most effective when it is treated as a recurring part of frontend development rather than a final inspection. Automated tools can identify useful signals, such as missing alternative text or certain contrast problems, but they cannot determine whether a user can complete a task comfortably with a keyboard or understand a complex interaction with assistive technology.
A reliable workflow therefore uses several kinds of checks:
- Automated checks for repeatable issues in markup, properties, and color combinations.
- Browser inspection for computed styles, accessibility trees, focus behavior, and responsive states.
- Manual interaction for keyboard navigation, forms, dialogs, menus, scrolling, and error recovery.
- Assistive-technology checks for a small set of realistic screen-reader tasks.
Test representative pages and components, not only the homepage. A checkout flow, authenticated dashboard, modal dialog, data table, search interface, and validation-heavy form can expose very different problems. Keep a short record of the page, browser, input method, issue, severity, and reproduction steps so regressions are easier to verify.
What to track
1. Keyboard navigation
Disconnect the mouse or avoid using it, then complete the page's primary tasks with the keyboard. Use Tab and Shift+Tab to move through controls, Enter or Space to activate them, and arrow keys where a widget requires them.
- Every interactive control receives focus in a logical order.
- Focus does not become trapped unexpectedly, except inside an intentionally modal interaction.
- Users can reach and operate menus, accordions, tabs, carousels, dialogs, and custom controls.
- There is a visible focus indicator with enough contrast against its surroundings.
- Keyboard users can dismiss overlays and return focus to the control that opened them.
- No content or action depends solely on hover, dragging, or pointer movement.
Pay particular attention to sticky headers, cookie notices, autocomplete fields, and custom dropdowns. These often alter focus order or hide the control that currently owns focus.
2. Structure, headings, and landmarks
Inspect the page as a document, not only as a visual design. Headings should describe the content hierarchy and should not be chosen merely because a particular size looks right. Check that the page has a clear main region and that repeated areas such as navigation, search, complementary content, and the footer are identifiable.
Review the accessibility tree in browser developer tools when the visual DOM and the exposed structure may differ. A visually obvious label, button, or status message should also have a meaningful accessible representation. Avoid adding ARIA where native HTML already provides the required behavior; incorrect roles and states can create more confusion than missing decoration.
3. Forms and validation
An accessible forms checklist should cover the entire task, including the initial instructions, data entry, validation, correction, and successful completion.
- Each field has a persistent, programmatically associated label.
- Instructions and examples are available without relying on placeholder text alone.
- Required fields and expected formats are communicated clearly.
- Error messages identify the field, explain the problem, and suggest a correction.
- Errors are discoverable without requiring the user to scan the whole page.
- Focus moves appropriately after submission, especially when a summary of errors appears.
- Success, loading, and asynchronous updates are communicated in a way assistive technology can detect.
Test invalid input deliberately: leave required fields empty, enter values in the wrong format, submit twice, and correct one error while leaving another. Confirm that the user's entered data is not discarded unnecessarily.
4. Color, zoom, and visual presentation
Do not use color as the only way to communicate status, selection, or error. For example, pair an error color with text or an icon that has an accessible name. Check normal text, large text, disabled-looking controls, links, focus indicators, charts, and form boundaries against their backgrounds.
Increase browser zoom and test a narrow viewport. Look for clipped text, overlapping controls, horizontal scrolling caused by layout mistakes, and dialogs that cannot be reached or dismissed. Accessibility and responsive testing often reveal the same underlying problems: fixed dimensions, hidden overflow, and controls that do not adapt to content.
5. Motion, media, and alternatives
Test the interface with reduced-motion preferences enabled. Animations should not prevent completion of a task, conceal important information, or make controls difficult to use. Provide a way to pause or stop motion that starts automatically when appropriate.
Check that images have useful alternative text when they convey information and empty alternative text when they are purely decorative. Captions, transcripts, descriptive labels, and text alternatives should be considered for video, audio, charts, and icon-only actions. An icon may be visually clear but still need an accessible name.
6. Screen-reader smoke tests
A screen-reader smoke test is not a replacement for broader testing, but it can reveal missing names, confusing announcements, and broken state changes quickly. Pick one common task, such as searching for an item, opening a menu, submitting a form, or completing a checkout step.
- Navigate by headings or landmarks to confirm the page has a useful outline.
- Move through interactive controls and listen for names, roles, values, and states.
- Open and close a dialog or menu, checking where focus goes.
- Trigger a validation error or asynchronous update and verify it is announced or otherwise discoverable.
- Complete the task without relying on visual position alone.
Behavior can vary by browser, operating system, screen reader, and version. Record the environment used and treat a smoke test as evidence about that combination, not as proof that every assistive-technology setup behaves identically.
7. Testing methods at a glance
| Method | Useful for | Limitations |
|---|---|---|
| Automated accessibility scanner | Repeatable markup, naming, and some contrast or state checks | Cannot judge task completion, wording, or every visual and interaction issue |
| Browser developer tools | Accessibility tree, computed styles, focus, layout, and responsive behavior | Requires interpretation and does not represent every user workflow |
| Keyboard-only review | Focus order, operability, traps, dialogs, and custom widgets | Does not test spoken output or all visual presentation concerns |
| Screen-reader smoke test | Accessible names, announcements, landmarks, states, and task flow | Results depend on the browser, operating system, and assistive technology |
| Human task testing | Real-world comprehension, efficiency, and recovery from mistakes | Needs representative tasks and participants; it is not fully automatable |
Cadence and checkpoints
Run a lightweight accessibility check during development, not only after deployment. A practical cadence is:
- During component work: test keyboard behavior, visible focus, names, states, and error handling as each interactive component is built.
- Before a pull request: run the project's automated checks and manually test the changed user flow at a typical desktop and narrow viewport.
- Before release: review the primary tasks on representative pages, including keyboard navigation, forms, zoom, reduced motion, and a screen-reader smoke test.
- Monthly or quarterly: sample important journeys again, review open accessibility issues, and confirm that test instructions still match the current interface.
- After a redesign or dependency change: repeat checks for shared components, navigation, dialogs, forms, and any code that changes focus or announcements.
Include accessibility checks in the same workflow used for frontend regression testing. A small, repeatable checklist is more useful than an ambitious audit that nobody repeats. For broader release debugging, pair this review with your frontend performance audit checklist and use browser tools to inspect layout and network behavior when a page fails under real conditions.
How to interpret changes
When a scan reports more issues, first determine whether the change reflects new defects, newly covered pages, a tool configuration change, or a change in the underlying component library. Compare the affected route and component with the previous version rather than treating the issue count as a complete quality score.
Prioritize problems that block a primary task: an unreachable submit button, a keyboard trap, a dialog that cannot be closed, an unlabeled form control, or an error that cannot be found. Then address issues that create repeated friction, such as inconsistent focus styles, unclear instructions, misleading headings, or status updates that are easy to miss.
For every issue, record the user impact, affected component, reproduction steps, expected behavior, and a suggested test that prevents regression. A fix is not complete until the original task works with the input methods and environments used to report the problem. If a CSS or JavaScript change affects layout or interaction, revisit both the visual and accessibility checks; a change that looks minor can alter focus order, clipping, or announcements.
When to revisit
Revisit this web accessibility testing checklist whenever a page, component, content model, browser target, or assistive-technology setup changes. Schedule a monthly or quarterly review for important user journeys even when the codebase appears stable. Also repeat the review after introducing a new modal, menu, date picker, rich-text editor, autocomplete field, chart, authentication flow, or validation pattern.
To make the next review actionable, keep a small test log with these fields:
- Page or workflow tested
- Browser, operating system, and assistive technology
- Keyboard result and focus observations
- Form and error-state result
- Automated findings and tool configuration
- Open issues, owners, and retest date
Start the next release review with the highest-value task, test it without a mouse, inspect focus and announcements, then run the automated checks. Fix blockers first, retest the complete workflow, and update the checklist when the product introduces a new interaction pattern. This turns accessibility from a one-time gate into a practical part of ongoing frontend quality.