Guide
Reading your dashboard
One audit is a to-do list. A dashboard full of them is a different thing — it answers which pages are worst, which single rule is costing you the most, and whether last month's work actually held.
Last reviewed: 8 August 2026
Everything lands in the same place
Audits reach your dashboard from three directions — saved from the extension, posted from a CI job, or produced by a monitor — and once stored they are indistinguishable. Same rows, same aggregates, same history. You can audit a page by hand today and monitor it from tomorrow without breaking its trend line.
Narrowing what you are looking at
There are three filters, and they combine:
- Page — scope the whole view to a single URL, or leave it on all pages.
- Search — free text matched against page URLs and titles.
- Severity — keep only the impact levels you care about: critical, serious, moderate, minor.
All three live in the URL, which is the useful part. A filtered view can be bookmarked, reloaded, or pasted into a ticket, and whoever opens it sees what you saw. Reloading a scoped view does not silently drop you back to “all pages”.
Selecting the same two severities in either order produces the same URL rather than two that differ only in spelling, so links to the same view really are the same link.
The trend
The trend charts issue counts across runs, so the shape matters more than any single point. A line stepping down after a release is the fix landing; a line creeping up between releases is regressions arriving faster than they are fixed.
Two cautions. A count is not a severity — trading four critical issues for six minor ones is progress that the total will read as a small regression, so check the impact breakdown before drawing a conclusion. And a run marked partial did not cover the whole page; it is flagged rather than hidden, because a partial run that looks clean would otherwise read as an improvement.
Per-rule pages
The most useful view is not the list of issues but the list of rules. Issues are grouped by the rule they broke, so a single contrast problem repeated on ninety elements is one row saying “ninety places” rather than ninety rows you scroll past.
That grouping is what makes the dashboard actionable, because it matches how the work is actually done. Ninety contrast failures are usually one token in a stylesheet. Fixing the rule fixes the row.
Open a rule and you get what it checks, why it matters, how to fix it, and every element that failed it, with the selector and the markup. Where the rule maps to a WCAG success criterion, the page links to the official Understanding document for it — and where there is no verified link, it says the criterion in plain text rather than guessing at a URL.
Turning this into a report
The same stored audits are the evidence base for a downloadable conformance report, covered in generating a conformance report. Nothing extra needs collecting — if it is on the dashboard, it is in the report.