Accessibility Testing Strategy

Accessibility Testing Strategy

Plan a testing approach that catches real accessibility barriers — because automated tools find roughly 30% of issues. The other 70% require human judgment.

The Testing Pyramid

Layer 1: Automated Testing (Catches ~30%)

What it finds:

• Missing alt text
• Colour contrast failures
• Missing form labels
• Missing language attribute
• Empty headings and buttons
• Duplicate IDs

What it misses:

• Whether alt text is meaningful
• Whether focus order is logical
• Whether content makes sense
• Whether interactions work with keyboard
• Whether the experience is usable

Tools: axe, Lighthouse, WAVE, Pa11y, jest-axe When: on every build, in CI/CD pipeline Who: automated — no human needed

Layer 2: Manual Expert Testing (Catches ~50%)

What it finds:

• Keyboard navigation issues
• Focus management problems
• Logical reading order failures
• Meaningful alt text assessment
• Cognitive accessibility barriers
• Interaction pattern issues

Method:

• Tab through every interactive element
• Test with screen reader (VoiceOver or NVDA)
• Test at 200% zoom
• Test with system preferences enabled
• Review content for plain language

When: before each release, on major feature changes Who: someone trained in accessibility testing

Layer 3: User Testing With Disabled People (Catches remaining ~20%)

What it finds:

• Real-world usability barriers that experts miss
• Assistive technology compatibility issues
• Unexpected workflows and workarounds
• Cognitive and emotional barriers
• Issues that only appear with actual AT proficiency

Method:

• Recruit participants with diverse disabilities
• Test core tasks with their own devices and AT
• Observe — don't guide or explain
• Record what works, what fails, and what frustrates

When: quarterly, or before major launches Who: users with disabilities using their own setup

What to Test First

Priority 1: Core Tasks

• Can users complete the primary tasks the product exists for?
• Test the critical path end to end
• Any failure here is a launch blocker

Priority 2: Input and Navigation

• Keyboard-only navigation through the core flow
• Screen reader navigation through the core flow
• Touch interaction on mobile for the core flow

Priority 3: Content and Comprehension

• Is content understandable at the target reading level?
• Are error messages helpful?
• Is help available and findable?

Priority 4: Edge Cases and Preferences

• Zoom to 200%
• System preferences (reduced motion, high contrast, dark mode)
• Long content, empty states, error states

Screen Reader Testing Basics

For teams new to screen reader testing:

Start With VoiceOver (Mac)

• Turn on: Cmd + F5
• Navigate: VO + arrow keys (VO = Ctrl + Option)
• Headings: VO + Cmd + H
• Links: VO + Cmd + L
• Forms: VO + Cmd + J
• Turn off: Cmd + F5

Start With NVDA (Windows)

• Download free from nvaccess.org
• Navigate: Tab, arrow keys
• Headings: H
• Links: K
• Forms: F
• Stop speech: Ctrl

What to Listen For

• Does every interactive element have a name?
• Are state changes announced (expanded, selected, checked)?
• Does focus move to the right place after actions?
• Can you understand the page reading only the headings?

Assessment Questions

1. Is automated testing running on every build?
2. Is manual expert testing happening before each release?
3. Has the product been tested by people with disabilities?
4. Are core tasks tested with keyboard and screen reader?
5. Is there a clear priority for what to test first?

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: Owl-Listener
• Source: Owl-Listener/inclusive-design-skills
• License: MIT

Install and usage instructions live in the source repository linked above.