The conformance status – and when you may deviate

Every statement says how far the website meets the requirements. This conformance status is the sentence readers, authorities and courts look at first.

The suggestion

On the publishing page, under “Release”, the page shows “Assessed status:” with a short reason. It follows from the findings of the scan, from your answers to the manual checks and from how many success criteria are not evaluated.

What the status rests on

The service puts every success criterion of the conformance target into exactly one of four states:

How to assess a criterion yourself: Answering the manual checks.

From this the service suggests:

• “fully conformant” if no criterion has a defect and none is not evaluated – the conditions are listed under What is not possible.
• “non-conformant” if a critical finding is open or the scan could evaluate too few pages.
• “partially conformant” in all other cases.

Deviating

Under “Differing conformance claim” you can choose a different “Status”. A “Reason” is then required. The suggestion, your choice and the justification are recorded together with the version – so it is clear who is responsible for which assessment. The summary before approval points this out.

A reason to deviate can be that you know of defects the automated scan cannot find – then a stricter status is right.

What is not possible

The service only suggests “fully conformant”, and only lets you choose it, if all of the following hold:

• No finding is open or assessed as a limitation.
• No question of the manual checks is answered with “not met”.
• No success criterion is not evaluated. In a scan measured against WCAG 2.2 there is no question yet for some criteria; they stay not evaluated, and the status there is “partially conformant” at most.
• No uncertain result is still waiting for your decision. As long as an occurrence is listed under “To review”, its success criterion stays not evaluated – even if you answered the question on it with “met”. The overview names these criteria: Assessing uncertain results.
• No question that blocks publication is open.
• The scan is reliable: Reading the scan report.

A tool that makes a false claim convenient would be a bigger risk for you than one that prevents it. If a finding should no longer count because it is not a defect, assess it accordingly: Reviewing findings. If a criterion is not evaluated, answer the questions on it, as far as there are any: Answering the manual checks.

What the statement says about what was not evaluated

The statement names how many success criteria were not evaluated and says that it makes no claim about them. If none is left, the paragraph on the method says instead that the remaining requirements were self-assessed.

So you can publish without having answered every voluntary question. The statement then reads “partially conformant” at most and says openly what it makes no claim about. How it does that depends on whether there are defects as well:

• There are known defects. Below the status it says “Grundlage: …” (basis) with the number of findings and failed manual checks, followed by a sentence of its own on how many success criteria were not evaluated. The list of defects follows. The paragraph on the method reads: “Von den … Erfolgskriterien wurden … nicht beurteilt; zu ihnen trifft diese Erklärung keine Aussage.” – Of the … success criteria, … were not evaluated; this statement makes no claim about them.
• No defect is known, but not everything is evaluated. Then the statement claims neither defects nor “largely met”. It reads: “Zu den beurteilten Anforderungen sind uns keine Mängel bekannt. … von … Erfolgskriterien wurden nicht beurteilt.” – We know of no defects in the requirements we assessed. … of … success criteria were not evaluated. There is no list of defects then, and the paragraph on the method only says that the statement makes no claim about the criteria that were not evaluated. For public bodies the section on accessible alternatives is then only included if you have entered some.

A hint from the scan – a recommendation, not a violation – does not count as a defect here. Hints therefore never appear in the statement, not even next to a defect; you still find them in the scan report.

If you have already published a statement

Previously, only findings and a few questions counted. Success criteria that were neither measured nor asked about went unmentioned. Now they count as not evaluated.

• Published versions stay as they are. Their text is not changed afterwards.
• The next draft may read differently. Where it said “fully conformant” before, the suggestion may now be “partially conformant” – not because your website got worse, but because the statement now names what was not evaluated.
• It is largely in your hands. Answer the questions on the criteria that are not evaluated, and they count as fulfilled by your own assessment or as a defect – and the status follows from that. The exception: in a scan measured against WCAG 2.2 there is no question yet for some criteria. They stay not evaluated, and the status is “partially conformant” at most.

In the Accessibility Conformance Report

The report for US procurement lists every criterion individually. A criterion without defects that the scan measures completely appears there as “Supports”. A criterion fulfilled by your own assessment also appears as “Supports”, always with the remark that it was self-assessed by the declaring party. What is not evaluated appears as “Not Evaluated”. More: Using the draft yourself.

Why “partially” is no stain

Most websites are partially conformant. A statement that openly names what is not yet accessible, what was not evaluated and how people can still reach their goal serves its purpose – one that claims “fully” and isn’t, does not.