Back to blog

Usability and Accessibility: Differences, Overlap, and Practical Testing

Shusaku Yosa

ユーザビリティとアクセシビリティの違いと設計ポイント

Usability asks how readily particular users can accomplish their goals. Accessibility addresses barriers to accessing and using information and functionality, including for people with disabilities. The two overlap, but evaluating one does not complete the other.

Avoid dividing people into two categories

It is misleading to assign usability to “general users” and accessibility to disabled users. Someone using a screen reader also needs a comprehensible, manageable process. Context, assistive technology, and task demands can change how the same interface works for a person.

W3C WAI discusses the overlap between accessibility, usability, and inclusion. Combine standards-based evaluation with examination of actual use, including use by people with disabilities.

A field can be technically reachable yet leave people unsure what to enter. Clear instructions do not solve a date picker that cannot be operated with a keyboard. Examine meaning, operation, and presentation separately before combining the findings.

Test a form with a concrete task

For a fictional consultation form, use the task “choose a suitable date and request a consultation.” The following improvements are examples to test, not a complete conformance checklist.

Problem

Consequence

Change to verify

Field names appear only inside inputs

The name disappears during entry

Keep a visible label associated with each input

Date selection depends on a mouse

Keyboard users cannot complete the request

Check every action and selection state by keyboard

Errors use red alone

People who cannot distinguish the color miss the problem

Explain the issue and correction in text

Submission has no clear confirmation

People may submit again

Communicate completion and the next step

Zoom hides the button

People enlarging text cannot continue

Test the sequence while enlarged

Check instructions, navigation order, and how controls are communicated to assistive technology. WAI's form-label tutorial covers labels associated with controls. Placeholder text should not be the only label.

Choose image alternatives by purpose

Adding a long description to every image is not the goal. Ask what the image contributes in its particular context. WAI's decorative-image guidance explains empty text alternatives for decoration.

Role

Example

Alternative text approach

Decorative

An ornament adding no meaning beyond nearby text

Use empty alt text so it does not add irrelevant speech

Informative

A diagram showing differences between plans

Convey the necessary meaning; use nearby explanation for complexity

Functional

An image acting as a submit control

Provide a name that explains the action

Nearby text may already provide an informative image's meaning. Decide whether additional text helps or merely repeats it. Empty alt text is a deliberate contextual decision, not an excuse to omit necessary information. Do not classify a functional image as decoration and remove the control's name.

Combine automated, manual, and user evaluation

Method

Useful checks

Limit

Automated checks

Some structural, contrast, and missing-name problems

Cannot fully judge meaning or successful task completion

Manual review

Keyboard use, focus, zoom, and error recovery

Reviewer experience and environments may leave gaps

User testing

Realistic tasks with intended users

Does not prove conformance beyond the tested scope

Do not coach people toward the answer during a test. Record when assistance was needed. Choose participants relevant to the intended conditions, including people using appropriate assistive technology. Recheck a revised interaction rather than assuming that the proposed fix solved the problem.

Keep a quick check separate from conformance

WCAG 2.2 specifies success criteria and conformance requirements. A short checklist, zero automated warnings, or one successful user session does not establish site-wide conformance. Define the version, level, pages, states, and evaluation method before making such a claim.

Record the screen, conditions, issue, expected outcome, and review date. Check whether someone can complete the request and recover from an error, not simply whether issue counts fell. Speed alone may penalize people who need time to review their entries.

Improve the complete task, then repeat

Follow the form from preparation through entry, error recovery, and confirmation. Improvements can support customer onboarding and first value. Use customer experience value to examine the benefit of completion alongside the effort required. Continue testing as content and interactions change.

Back to blog