Begin with the service decision
An audit intended to support a release has different priorities from one intended to establish a product-wide baseline. State the decision first so the scope can produce relevant evidence.
Choose representative coverage
Record page types, shared components, high-risk states and journeys from entry to completion. Include errors and recovery paths, not only successful completion.
Agree desktop and mobile viewports, browsers, test accounts, assistive technology and any unavailable environments before testing begins.
Make exclusions visible
A clear report states what was tested, what was unavailable and what changed during the assessment. That context prevents a tested sample from being mistaken for a whole-service guarantee.