Schedule monitoring and manual tests

Use a cadence that fits ongoing work

Approve and resource this proposed local cadence. Increase frequency when a service changes often, has a serious known barrier or supports a time-sensitive task.

• Weekly quick review: during active publishing, check upcoming releases, newly reported blockers, urgent notices and progress on assigned fixes.
• Monthly thorough review: review open issues, changed pages and files, links, event content, scanner results and overdue manual checks.
• Quarterly manual review: choose representative templates and complete key tasks with keyboard, zoom and appropriate assistive technology. Include interactive states and a sample of documents and third-party services.
• Annual evidence review: review the inventory, evaluation scope, findings, remedies, training and resource decisions; issue a report with methods, limitations and actions for the next cycle.

For a skipped or deferred review, record the reason, owner and new date. Save completed results with the review record.

Select a sample and test complete tasks

Use the inventory to group pages by template, content pattern, interaction and risk. Include recently changed or frequently used journeys, known problem areas, and a varied sample of documents and media. Explain why the sample represents the stated scope. Include an additional random sample when feasible to detect issues outside the obvious high-risk pages.

For each selected item, define the start point, expected task, interaction states and observable success. Test keyboard reachability and order, focus, labels and instructions, error recovery, zoom and reflow. Use a screen reader for relevant structures and changes in state. For a document, inspect the final published copy. For a process, test from start to completion because problems may appear between pages or inside an embedded service.

Use automated checks to identify code patterns, then verify findings and manually assess content and complete user tasks. Record cases the tools cannot evaluate and request specialist review when needed. When practical, include people with disabilities in usability work; the responsible program should arrange preparation and compensation.

Record results and their scope

A monitoring record includes review ID, date, reviewer role, scope, sample selection reason, URLs or file versions, environment, browser and assistive technology, test steps, expected result, observed result, evidence link and related issue ID. For automated checks, add tool version, ruleset, configuration, raw output and reviewed findings. Keep personal information out of general reports.

Report the size and nature of the sample, methods used, criteria checked, exclusions and open findings. Label results with the reviewed sample and date. An annual evaluation should be planned by people with suitable accessibility expertise and can follow WCAG-EM’s process for scope, exploration, sampling, evaluation and reporting.

Compare results over time only after accounting for inventory or sample changes. Route defects into the issue register and verify fixes with the original steps. See quick checks for an editor-level test and maintenance guidance for release retesting.

Use a working record

Complete a record when the activity occurs, download it and retain it in the approved team location.

• Scheduled monitoring cycle
• Manual accessibility test
• Assessment evidence and distribution review