What accessibility evidence should a Shopify team keep?
Keep dated evidence that shows what was tested, which standard and scope were used, what failed, who owned each fix, and when a person verified the result.
Separate evidence from a score
A score is a useful trend signal, not a conformance record. Preserve the actual issue, page or component, test method, severity, and result so another person can understand what was checked.
Track fixes through verification
An issue should move from open to fixed only after the affected interaction is retested. Attach the release identifier and a short note about keyboard or assistive-technology verification when the issue requires it.
Make the log operational
Review recurring failures by component and app. If the same issue returns after releases, add it to pre-deploy checks or component tests instead of treating each occurrence as a new one-off task.
Sources and testing references
These sources describe accessibility techniques and WCAG success criteria. They do not by themselves establish legal compliance.
Frequently asked questions
What belongs in an accessibility evidence log?
Dated records of evaluations and what they found, fixes made and verified, monitoring outputs, training, vendor correspondence about barriers, and decisions taken on known issues, each with an owner. The log should let a stranger reconstruct what the store knew, did, and planned, in order.
How far back should evidence be kept?
Keep it for as long as the site operates in anything like its current form, and longer where contracts or settlement terms require. Storage is cheap and the questions evidence answers often arrive years after the work, from auditors, buyers, insurers, or counsel.
Who should own the evidence log?
One named role, not a shared folder's worth of good intentions: whoever owns accessibility operations. Contributors add entries, but one owner keeps the index complete, the format consistent, and the gaps visible before someone else finds them.
Can an evidence log hurt a store by documenting known issues?
Stores worry about this, but the greater risk is the opposite record: barriers discovered with no evidence of response. A log showing findings, priorities, fixes, and timelines demonstrates the good faith and ongoing effort that regulators and courts look for. Concealment helps no one for long.