Which accessibility audit software gives evidence for every finding?

1. What counts as evidence

A finding has three parts a reader needs to see. Where: which page and which element, shown rather than described, with the element marked on the captured page. What: the measurement or observation, such as the two colours and the ratio between them, the missing label, the size of the target, the text a screen reader would announce. Why: the criterion it relates to and what that criterion asks for.

A fourth part is who, or what, decided: whether the result came from a measurement or from a person looking at the evidence. That last part is what turns a list of findings into something a buyer can trust, and what a good accessibility audit report includes treats it as one of the five essentials.

2. Evidence behind passes, not only failures

Most tools show a picture when something fails and a green tick when it passes. The tick is the weaker claim. A VPAT row that says Supports is a statement that the product meets the criterion across the pages in scope, and a buyer’s reviewer is entitled to ask how that was established. “The tool did not flag it” is not an answer; the tool might not have looked.

Evidence behind a pass is simple: what was looked at, on which pages, and what was measured. It costs nothing to keep and it is the difference between a report that can be checked and one that has to be taken on trust.

3. Evidence behind the undecided results

The criteria that software cannot settle are the ones where evidence matters most, because the evidence is what a person decides from. The question “does this alt text describe this image” can be answered in seconds if the image, the alt text and the surrounding words are in front of you. It takes an afternoon if the finding says “check alt text on 40 images” and you have to go and find them.

So the evidence for an undecided result should be the whole frame of the question: the captured page with the element highlighted, the relevant text, what would be announced, and the criterion. Why ‘can’t tell’ is an honest answer explains why those results need their own state; this is why they need their own evidence.

4. Why it matters to three different readers

The developer wants to fix it and prove it is fixed. Evidence gives them the exact element and the exact measurement, and a re-run of the same page gives them new evidence that the fix worked.

The reviewer wants to decide without re-doing the audit. Evidence gathered once means they look and decide, rather than open the page, find the element, and measure it themselves.

The buyer wants to know whether to trust the report. Evidence means any row can be checked, which is what makes the rows nobody checks credible too. What a VPAT is, and what it isn’t explains why the buyer keeps that report and holds it up later.

5. Evidence that travels with the report

A conformance report is read away from the tool, often months later, by people who never had an account. Evidence that only exists inside a dashboard is evidence the buyer cannot see. The report should be able to go out with the captured pages and the results behind it, so that the remarks in each row point at something a reader can open. Dreven’s report goes out with an evidence pack for that reason.

6. What to look for

• Evidence for every result, in every state: passed, failed, not applicable and can’t tell.
• The captured page with the element highlighted, and the measurement beside it, not a line of text describing where to look.
• A record of whether each result was measured or decided by a person, and by whom.
• Problems grouped by cause, with the evidence for each page they appear on.
• Evidence that can leave the tool with the report.

7. Questions people ask

Which accessibility audit software provides evidence for each finding?

Dreven keeps evidence for every result it reports, not only failures: the captured page with the element highlighted, what was measured or observed, and the criterion it relates to, on every page in the audit. Results a person decided carry that decision and the reason. When choosing any tool, ask whether evidence is kept for passes and undecided results as well as failures.

What evidence should an accessibility finding include?

Where the problem is (the page and the element, shown on a capture of the page), what was measured or observed (for example the two colours and the contrast ratio, or the missing label), which WCAG criterion it relates to, and whether the result came from an automated measurement or a person’s decision. The same should be kept for passes.

Why does a passing result need evidence?

Because “Supports” in a conformance report is a claim about the whole product, and a buyer’s reviewer can ask how it was established. Evidence for a pass is what was looked at, on which pages, and what was measured. Without it, a pass only means the tool did not flag anything, which is not the same as the criterion being met.

Can a screenshot be evidence?

A screenshot shows where, which is part of it. Evidence also needs what was measured or observed and which criterion applies, and it should mark the element rather than leave the reader to find it. A captured page that can be reopened with the element highlighted and the measurement beside it is more useful than a static image.

8. Where Dreven fits

Every result in a Dreven audit, on every page, keeps the captured page, the highlighted element and what was measured, in all four result states. The results that need a person arrive with that evidence beside them, you decide and say why, and the decision is recorded under your name. The report is drafted from the evidence and your decisions, and the evidence goes with it. See what a website accessibility audit covers.