Reading the Report
Open a scan from the Scans tab, or from the link in the email you get when it finishes. The same report downloads as a PDF to share with your team, a customer or an auditor.
Download an example report (PDF)
Issues
An issue is one failing element. The same unlabelled search box in a header shared by 400 pages is one issue that lists the pages it is on, not 400. Each issue shows:
- the WCAG success criterion it fails, with its level (A or AA);
- a severity;
- what is wrong and, where there are several ways to fix it, the options;
- the failing markup and a selector to find it;
- the pages it appears on;
- How to fix, a link to the W3C's explanation of the criterion and the techniques for meeting it.
In the PDF, issues of the same kind are grouped, so ten unlabelled form fields read as one problem with ten affected elements.
Severity
Severity reflects how strongly a failure blocks people who use assistive technology:
| Severity | Meaning | Example |
|---|---|---|
| Critical | Blocks access for some people entirely | A button with no name; an image link with no text |
| High | Seriously impairs access | Low-contrast text; invisible keyboard focus |
| Medium | Causes real friction | A list marked up so it is not announced as a list |
| Low | Minor | Rare among WCAG failures |
Flagged for review
Some results cannot be decided by a browser, for example the contrast of text drawn over a photograph, or whether a video has captions. These are listed under To check by hand. They are not failures and are never counted; a person should look at each one.
Pages
Every page the scan opened is listed. A page that could not be checked (an HTTP error, a bot-protection page, a redirect off your site) is listed with the reason. It is never counted as passing.
The criteria list
Every WCAG A and AA success criterion for the standard you chose is listed with one of four results:
| Result | Meaning |
|---|---|
| Fails | At least one confirmed failure, listed under Issues |
| Flagged for review | Something may fail it; a person needs to check the flagged items |
| No automated failures | The automated checks for it ran and found nothing. Automated checks cover part of each criterion, so this is not a pass |
| Needs manual review | It depends on meaning, content or interaction that automated checks cannot judge |
The list never marks a criterion as met or as not applicable, because automated testing can show neither. Use it to see at a glance what is broken and to plan the manual review that any conformance claim or VPAT needs.
Compared with the previous scan
From the second scan of a site onwards, each issue is compared with the previous completed scan:
- New: not present last time.
- Fixed: present last time and gone now. An issue only counts as fixed if this scan checked every page it was on before; a smaller crawl that never reached those pages does not claim a fix.
Compare like with like: keep the same scan mode for a site if you want the numbers to line up from run to run.
Dismissing an issue
If an issue does not apply to you, for example one inside a third-party widget you cannot change, dismiss it with the button on the issue and say why. A dismissed issue:
- stays visible, marked dismissed, with your reason;
- is not counted in the totals;
- stays dismissed on every later scan of the same site.
You can restore it at any time. Prefer Ignore these elements on the site for whole widgets you never want checked.
Exporting
- PDF: the full report, for people.
- JSON: every field of the run, for your own tooling.
