Turn a Website Accessibility Audit Into a Fix Plan

Learn how to scope a website accessibility audit, combine testing methods, prioritize fixes, interpret reports, and retest changes with confidence.

Turn a Website Accessibility Audit Into a Fix Plan

Turn a Website Accessibility Audit Into a Fix Plan

A website accessibility audit is a structured review of barriers in content, design, code, navigation, forms, media, documents, and interactive features. A reliable audit combines automated testing with manual review and, where appropriate, assistive-technology or user testing. An automated score alone cannot confirm WCAG conformance or legal compliance.

The useful outcome is a clear list of problems, affected pages or components, user impact, recommended fixes, responsible owners, and retest results. This guide explains how to scope the work, compare testing methods, evaluate the report, and turn findings into an actionable remediation plan.

What a Website Accessibility Audit Actually Tells You

A website accessibility audit evaluates whether people with different disabilities can find information, navigate pages, complete tasks, understand feedback, and use interactive features. It examines visual, auditory, cognitive, speech, neurological, and physical access barriers across a website or web application.

Review typeWhat it doesWhat it cannot establish alone
Automated checkerScans for detectable patterns such as missing labels, some contrast failures, and invalid markup.Whether every task is usable or whether the site conforms overall.
Monitoring scanRepeats automated checks and flags changes on selected pages.Whether new interactions, documents, or third-party features work for real users.
Accessibility auditCombines defined scope, automated checks, manual review, documentation, prioritization, and retesting.A guarantee that every user, technology, or legal requirement is covered.
Legal reviewAssesses obligations under a specific law, policy, contract, or jurisdiction.It is not replaced by a technical audit.

Canadian government accessibility material commonly refers to WCAG 2.1 Level A and Level AA success criteria in its assessment methodology. WCAG provides technical guidance, but applicable legal requirements still depend on context. WCAG 2.1 Level A and Level AA criteria are a useful reference, not legal advice.

A checklist can help a team prepare, but an audit adds scope, evidence, ownership, and verification. Use a website accessibility checklist as a preparation tool, not proof that a site has no barriers.

Scope the Audit Around Real User Journeys

The right scope follows the tasks people need to complete, the templates that produce pages, and the components shared across the site. A one-page scan cannot represent a checkout flow, logged-in portal, PDF library, or responsive menu that was not included.

Define the scope before testing starts

A written scope should identify the audience, important tasks, URLs, page templates, content types, responsive states, browsers or devices, assistive technologies, conformance target, exclusions, and expected deliverables. Ask for these details before approving an audit proposal.

Include the homepage, key landing pages, search, service or product templates, forms, account areas, checkout or donation flows, and shared components such as menus, dialogs, filters, cookie notices, and validation messages. Also identify images, video, audio, PDFs, embedded maps, and third-party widgets.

For a small business website, a practical starting scope includes the homepage, a representative service page, navigation, the contact or quote form, mobile layouts, and important downloads. For a nonprofit, add donation, volunteer, event, membership, and newsletter journeys. For eCommerce, include categories, products, filters, cart, checkout, account functions, order confirmation, and support pathways.

Use representative templates, not an arbitrary page count

One page can reveal a shared component problem, but it does not prove that every page has equivalent headings, images, links, or content. Select pages that represent different templates, content lengths, forms, tables, media, error states, and responsive layouts.

Ask the auditor to explain how the sample represents the full site. A useful scope map lists each selected URL, its template, its business purpose, the user journey it supports, and the shared components it exercises. If a component appears on 80 pages, one representative test can expose a systemic issue, but the report should identify the other affected pages or templates.

Include authenticated areas, documents, and third-party tools

Include logged-in areas when customers, members, staff, donors, or applicants need them to complete an important task. Provide test accounts with appropriate permissions and identify different roles, such as administrator, member, customer, or applicant, because the interface and available actions may differ.

Include PDFs when they contain essential instructions, forms, pricing, policies, menus, or reports without an accessible HTML equivalent. The review should cover reading order, headings, tables, form fields, document language, links, and whether the HTML alternative contains the same practical information.

List booking, payment, map, chat, social, consent, analytics, video, and customer-support tools separately. Record which parts your team controls, which parts belong to a vendor, and what alternative route exists if the embedded experience is inaccessible. A third-party boundary does not remove the need to identify the barrier or explain its effect on the user journey.

Expected audit workflow and deliverables

  1. Discovery and scope: agree on users, tasks, templates, environments, conformance target, exclusions, test accounts, and content types. Deliverables should include a scope document and page or journey inventory.
  2. Baseline testing: run suitable automated checks and record the tool, date, pages, rules, and known limitations. This creates a repeatable starting point, not a conformance certificate.
  3. Manual evaluation: review keyboard operation, focus, headings, names and states, forms, errors, zoom, reflow, media, documents, and realistic tasks. Deliverables should identify the steps used to reproduce each issue.
  4. Assistive-technology or user testing: add this layer for complex interfaces, high-value journeys, known access needs, or major redesigns. Record the environment and task so results can be interpreted accurately.
  5. Reporting and prioritization: group duplicate findings, identify shared components, map findings to affected URLs, and assign owners and priorities. Deliverables should include an issue register and an executive summary.
  6. Remediation: resolve content, design, development, configuration, document, or vendor issues. Track each fix against its original finding instead of treating a new scan as proof that all work is complete.
  7. Retesting: reproduce the original issue, test the corrected state, check related templates and responsive states, and record the final status. Deliverables should include a retest log and unresolved-risk list.

Define the audience, tasks, URLs, templates, states, devices, technologies, conformance target, and exclusions before testing. A structured website planning process can help teams connect audience research, customer journeys, sitemap planning, wireframes, architecture, and feature requirements before implementation.

Choose the Right Mix of Accessibility Testing Methods

Developer testing website navigation with keyboard beside responsive mobile device

No single method finds every accessibility problem. Automated tools provide breadth and repeatability, while manual and user testing reveal whether people can complete important tasks. Government of Canada guidance notes that older tools may not work properly with modern components and recommends testing with people with disabilities for better coverage. Accessibility testing guidance supports a layered process.

MethodUseful for findingImportant limitsUse it when
Automated scanningSome missing labels, empty links, invalid structure, contrast issues, duplicate IDs, and other machine-detectable patterns.It cannot reliably judge meaning, task completion, helpful alternative text, or every focus interaction.You need broad baseline coverage or repeatable regression checks.
Manual keyboard reviewFocus visibility, order, traps, menus, dialogs, custom widgets, and task completion without a pointer.It requires a tester who understands the intended task.The site contains forms, menus, dialogs, filters, booking, checkout, or other interactive flows.
Assistive-technology reviewHeadings, landmarks, labels, names, states, announcements, and dynamic updates.Results vary by browser, operating system, technology, settings, and tester experience.Components update dynamically or expose information through a screen reader or other assistive technology.
Responsive and mobile testingZoom, reflow, touch access, clipped content, mobile menus, and small-screen task completion.A desktop pass does not predict mobile usability.Mobile visitors use core forms, navigation, purchasing, booking, or account features.
Testing with people with disabilitiesReal-world friction, unclear language, inefficient workflows, and unexpected barriers.It complements rather than replaces technical testing.The audience is diverse, the task is consequential, or the interface is being substantially redesigned.

Practical keyboard testing

Use Tab and Shift+Tab from a clean page load. Check that focus is visible, follows a logical order, and reaches every interactive control with an understandable name.

Open menus, filters, dialogs, date pickers, carousels, and forms. Confirm that each control can be entered, operated, and exited. When a modal opens, focus should move into it, remain there while required, and return to a sensible location when it closes. No control should trap the keyboard.

Repeat the test with validation errors, empty results, success messages, and an interrupted task. A form that works in its default state can still fail when an error appears.

Responsive and mobile testing

Test important tasks at narrow and wide viewports, with enlarged text and zoom. Look for horizontal scrolling, clipped labels, overlapping controls, hidden focus indicators, inaccessible menus, and forms that become difficult to complete.

Assistive-technology testing is especially useful for dynamic updates, custom controls, account areas, complex forms, booking, filtering, donation, and checkout. User testing adds value when the task is consequential, the audience has known access needs, or the team is making a substantial change.

For additional implementation context, review accessibility testing and implementation guidance alongside your defined scope and evidence.

Common Findings by Website Component

Failures become easier to fix when described in the language of the component and user task. These are common patterns to investigate, not claims about any particular website.

ComponentCommon problemRemediation direction
ImagesMissing, duplicated, or unhelpful alternative text.Describe informative images, name functional images, and mark decorative images appropriately.
HeadingsSkipped levels or headings chosen only for visual size.Organize content by meaning and identify repeated regions consistently.
ColourText, focus indicators, or error states are difficult to distinguish.Use sufficient contrast and communicate status with more than colour alone.
FormsMissing labels, unclear required status, or errors disconnected from fields.Use persistent labels, clear instructions, and understandable error summaries.
Menus and dialogsKeyboard users cannot open controls, focus disappears, or state is not exposed.Support keyboard operation, expose state, manage focus, and test open and closed states.
Carousels and mediaContent changes unexpectedly or captions, transcripts, and controls are missing.Give users control and provide equivalent stable information.
PDFsMissing reading order, headings, tables, form fields, or language metadata.Repair the document or provide an accessible HTML alternative.
Third-party embedsBooking, payment, map, chat, or social tools introduce barriers.Record the vendor boundary, request remediation, or provide an alternative.

For organizations serving diverse communities, WCAG-compliant design can make audit recommendations more actionable by embedding inclusive practices into navigation, forms, content structure, and development decisions.

What a Useful Accessibility Audit Report Contains

A strong report gives developers, designers, content editors, and business owners enough information to act without repeating the investigation. It should be both a decision document for leadership and a usable backlog for the people making changes.

  • Reference: finding ID, URL, template, component, device state, or journey.
  • Issue and impact: what prevents access, who is affected, and which task becomes difficult.
  • Technical reference: applicable WCAG criterion or internal requirement.
  • Reproduction: exact actions, device or browser state, and input needed to observe the issue.
  • Expected behaviour: what should happen.
  • Recommended fix: content, design, code, configuration, or vendor action.
  • Ownership and reach: responsible team and whether the problem repeats across templates.
  • Priority and status: release stage, owner, due date, and retest outcome.

Hypothetical report entry

Finding A-014: Booking dialog does not receive focus

Location: Appointment booking component on the service template.

Issue: The dialog opens visually, but keyboard focus remains behind it.

Impact: A keyboard user may be unable to start the booking task.

Reproduction: Tab to the booking button, press Enter, then press Tab and observe focus.

Recommended fix: Move focus into the dialog, keep it there while open, provide a named close control, and return focus to the trigger after closing.

Owner: Front-end development with design review.

Priority: Immediate release blocker if booking is a core task.

Retest: Open, complete, submit, invalidate, and close the dialog at desktop and mobile widths.

Prioritize Remediation When Resources Are Limited

Fix the barriers that block important tasks or affect many users first. Then weigh user impact, site-wide reach, business importance, implementation effort, and applicable organizational or legal risk so the team can sequence the remaining work transparently.

Task blockage is usually the strongest signal. An inaccessible login, checkout, donation, booking flow, or primary navigation can prevent a person from using the service at all. Reach then separates a shared header defect from a one-off problem on an obscure page. Business importance matters because a barrier on the main quote form deserves faster attention than one on a rarely used campaign page.

Effort helps sequence work, but it should not be used to postpone a severe barrier indefinitely. A small change that restores access to a core form can outrank a complex redesign. Risk includes contractual, organizational, policy, and jurisdiction-specific considerations, which should be documented separately from the technical finding.

  • Immediate release blocker: an inaccessible login, checkout, donation, booking, navigation, or shared component. Assign an owner and retest the full task.
  • Next-release work: a substantial barrier on a common template or a fragile workaround. Schedule design and development together.
  • Content correction: poor alternative text, unclear headings, vague links, or inaccessible documents. Give it to the content owner.
  • Planned improvement: lower-impact inconsistencies, legacy documents, or third-party limitations. Record the limitation and define a path forward.

For each finding, record the affected task, number or type of pages, user impact, owner, proposed release, effort estimate, and retest condition. That turns an audit from a static score into a managed remediation backlog.

Build Accessibility Into Design, Development, and Maintenance

Address accessibility before a pattern is repeated throughout a site. Discuss user journeys in requirements, review headings and states in wireframes, check contrast and focus in visual design, build semantic controls, and test realistic tasks before launch.

Accessibility works best as a cross-functional requirement. Content editors own meaningful alternative text and document structure, designers define readable hierarchy and visible states, developers implement semantic controls and focus management, and project owners decide which user journeys must work at launch.

For custom-coded pages, interactive interfaces, member portals, dashboards, booking, or lead capture, the custom web page design service lists capabilities such as custom-coded pages, interactive UI and UX, member portals, dashboards, appointment booking, and contact and lead-capture pages. Attach accessibility requirements to each workflow, not only the homepage.

Retest the Fixes and Prevent Regression

Retesting is a second evaluation, not merely a new scan. Reproduce the original finding, verify the fix in affected templates and responsive states, repeat the manual task, run suitable automated checks, and record whether the issue is resolved, partly resolved, not reproducible, or still open.

A practical retest record includes the original finding ID, changed component or URL, deployment date, test environment, steps repeated, result, remaining limitation, and reviewer. If a shared component changed, test at least one representative page from each important template that uses it.

New content, plugins, templates, redesigns, embedded tools, and landing pages can introduce barriers. Compare website maintenance plan responsibilities with accessibility regression work assigned as a specific responsibility.

Frequently Asked Questions About Website Accessibility Audits

Can an automated scan confirm WCAG conformance?

No. It identifies some detectable patterns but cannot reliably judge meaning, task completion, helpful alternative text, every focus interaction, or error recovery.

How many pages should an audit cover?

There is no universal count. Include important journeys, major templates, shared components, representative content, responsive states, and relevant authenticated, document, media, and third-party experiences.

Should logged-in areas and PDFs be included?

Yes, when people need them for an important task. Provide suitable test accounts for authenticated areas and include essential PDFs unless equivalent information is available in accessible HTML. Otherwise, document the exclusion clearly.

When should people with disabilities participate?

They add particular value for high-value journeys, complex interfaces, public services, diverse audiences, and major redesigns. Their testing should complement technical review.

How often should a site be retested?

Retest after fixes are deployed and repeat relevant checks when templates, content systems, plugins, documents, interactive features, or third-party services change.

Does passing an audit guarantee legal compliance?

No. An audit reports technical findings against a defined scope. Legal duties vary by jurisdiction, organization, service, contract, and audience.

Choose the Next Practical Step

List the pages, templates, journeys, forms, interactive features, documents, mobile states, authenticated areas, and third-party tools that matter most. Then request a testing plan covering methods, report format, remediation ownership, exclusions, and retesting. The Website Design Calgary service describes responsive website design, redesign, SEO, security, and CMS options for organizations planning a new or upgraded website.