Accessibility Heuristic Evaluation

Why it exists

Traditional usability heuristics focus on general cognitive friction and often miss barriers specific to disability. Recruiting users who rely on screen readers, keyboard navigation, or switch access for every design iteration is slow and expensive. Accessibility heuristics were developed to give teams a repeatable way to catch blocking issues early without requiring disabled participants at every stage.

The mental model

Think of it as a building-code inspection for disabilities. Just as an inspector checks that exits are wide enough and fire alarms include visual strobes, an accessibility heuristic review checks that an interface works without a mouse, without sound, without precise color vision, and with assistive technologies. It turns abstract compliance rules into concrete interaction questions.

How it works

A trained evaluator systematically reviews an interface against principles such as keyboard operability, screen reader compatibility, color independence, focus visibility, and error recovery. For example, can a user reach every control using only the Tab key? Do form labels remain programmatically associated with inputs so a screen reader announces them? If all color is removed, does meaning survive? The evaluator records violations by severity and maps them to relevant WCAG criteria. Running multiple reviewers in parallel improves coverage because no single expert catches every issue.

When to use it

Use accessibility heuristics during early design critiques, before code freeze, or when conducting competitive audits. They are especially useful when budget or timelines delay formal usability testing with disabled participants, or when you need to filter out showstopper barriers like keyboard traps and missing alt text before inviting users to test.

When not to use it

Never treat heuristic evaluation as a substitute for usability testing with people who have disabilities. Heuristics predict probable failures, not real-world friction. They also should not serve as a final launch gate, because expert reviewers cannot replicate the full diversity of assistive technology configurations, disability experiences, and adaptive strategies that actual users employ.

One canonical example

A team preparing an e-commerce checkout flow runs an accessibility heuristic review before recruiting participants. The reviewer discovers that credit card error messages appear only as red text below the input, with no icon or programmatic alert. The heuristic for perceivable error states flags this as high severity. Fixing it before research means screen reader users can recover from mistakes during testing instead of hitting an invisible wall that halts their progress.