WCAG process for assigning levels

This post is my own opinion and observations. I'm not writing this with my AGWG co-chair hat on, although I can't really get away from that being a major part of my perspective.

How provisions are assigned to a level is an important topic. Provisions means "success criteria" in WCAG 2, and "requirements" in WCAG 3. It does depend on the conformance approach. For example, if there is one level then it is an in/out discussion. If there are multiple levels it would need more mechanisms to categorise them.

In WCAG 2 the A, AA and AAA levels were used to signal importance, i.e. the things at level A were more important than level AA. In practice A and AA got smooshed together because they were both required by regulators, but that wasn't a forgone conclusion when WCAG 2.0 was released. In the conformance section of WCAG 2 it says that AAA is not recommended for compliance in regulations, and regulators took that on board. The group has no say over how the standard is used by regulators, but levels were a "hook" they could use to decide how much of the standard to require.

Consensus process

In WCAG 2 the level was determined by consensus of the group (as in, Accessibility Guidelines Working Group). The participants of the group are people with disabilities, people from accessibility testing companies, invited experts, and people from large W3C member organisations such as Microsoft, Google, and CVS Pharmacy.

With that consensus process, in the best cases, it was based what people could live with by discussing the details and balancing user-needs and author feasibility. In the worst case it would be the path of least objections. Anything at A or AA level had to be possible to do by any person or organisation across all possible scenarios. If it was a good thing to do but not possible in some scenarios, it would be put at the AAA level.

Principles

There were principles (recorded on the wiki).

Basically, the more important it was (across scenarios) to the user the higher the level, the more difficult/expensive it was to accomplish the lower the level.

However, even with principles there are many edge cases, the variety of digital interfaces and user-circumstances makes it impossible to apply a simple rubric.

An important factor I've not seen mentioned before is: If something is very good for the user in 90% of the cases, but irrelevant or bad for users in 10% of the cases, we can’t include that at the levels taken up by regulation.

If 10% of authors get sued over something that doesn’t help users it would undermine the use of the standard. Often we can create exceptions so it doesn’t apply to the 10%, but not always.

Regulatory uptake

There was a very, very long discussion on this for WCAG 2 in 2024. I’d like to highlight something I wrote there:

It is a stakeholder process, not a scientific one. Producing something that satisfies the various stakeholders can't be done by a pre-set formula. I think places like the EU parliament are looking for this kind of process! It means we have done the work of balancing user and industry needs which makes implementing regulations easier.

The difficulty is that if we only used a metric of what matters to users, WCAG would not be taken up by laws and regulations. Then it wouldn't be used by as many people. We’ve seen this dynamic repeatedly: An over-reach by an accessibility standard creates a push-back from industry, and regulators back-off from that standard.

A minor version of this is that some jurisdictions create exceptions to particular requirements, e.g. In some parts of Canada there are exceptions for live captions and audio description. The more that happens the less useful WCAG becomes because it is not a consistent standard internationally.

One of the key benefits of WCAG at W3C is that our group includes user-advocates and industry representatives (sometimes in the same person). We do the work of establishing the best requirement that can be achieved in practice. That is a key factor in the success of WCAG, why it gets taken up legally, and why so many people use it.

If that sounds bleak, my consolation is that we’re making progress on providing better guidance to regulators in WCAG 3, and better ways to integrate recommendations that go beyond conformance. These are really the key to having WCAG be enforced (and therefore seen by the most people) and get better accessibility advice in front of people.

I don't have comments on here, but there are posts on Linkedin or Mastadon that could be used for that...