Test

Build layers of evidence.

No single test tells you whether a product is accessible. These guides focus on what each method can establish — and what it can't.

Practical guides

A layered starting point when you need more than an automated scan.

Test real tasks, focus, operation and recovery from the keyboard.

Use representative tasks instead of listening to every element in sequence.

Find content and functionality that break when people need things larger.

Deliberately make mistakes and find out whether the interface helps you continue.

Check whether meaningful dynamic updates are communicated beyond the visual interface.

Look for memory, ambiguity, interruption and recovery costs automation cannot evaluate.

Record what was established and what remains unknown instead of producing an accessibility score.