Evidence Studio
Professional field guide
Make every finding retestable.
A finding is useful when another person can understand the barrier, locate it, reproduce it, fix it and confirm the result. This checklist keeps that evidence chain intact.
This is a workflow checklist, not a compliance standard. Adapt it to the applicable standard, audit methodology, client agreement and the needs of disabled users.
Finding-level evidence
- Tested page and state. Record the exact URL, page title, authenticated state, modal or step, and any content variation needed to reproduce the barrier.
- Test conditions. Include browser, viewport, zoom, language, relevant assistive technology and versions, contrast or forced-colour state, and the input method used.
- Human-readable problem. Describe what happens, who is affected, and which task becomes difficult or impossible. Do not paste scanner text as the final explanation.
- Applicable requirement. Connect the finding to the standard and version in scope. Preserve the specific criterion and the reasoning used; avoid implying that one criterion explains every effect.
- Reproduction steps. Write the shortest reliable sequence another tester can follow. Include expected and actual outcomes.
- Protected visual or technical evidence. Attach the relevant screenshot, DOM excerpt or other evidence. Redact personal and client data in the shareable copy while preserving the controlled original where policy permits.
- Occurrences and shared cause. Link repeated instances without hiding their locations. Only group results when the same corrective action genuinely applies.
- Remediation guidance. State the desired behavior and relevant implementation principle. Keep the development team responsible for choosing code that fits its system.
- Retest trail. Record the build or date, the same test conditions, result, remaining occurrences and any regression evidence. Do not overwrite the original finding.
Before a client export
- Check every screenshot and free-text field for secrets, personal information and unrelated customer data.
- Make link text, headings, tables and status labels understandable without relying on color.
- Separate confirmed findings, observations, limitations and items not tested.
- State the sample and scope; do not imply that an evaluated subset proves the entire product passes.
- Open the exported report with a keyboard and at least one relevant assistive technology.
Build one finding in a working demo.
Evidence Studio keeps source results, protected screenshots, decisions, occurrences and retests connected in a local project.
Try the premium demo