Six fixes that would move the web faster than WCAG

I keep hearing that the fight for an accessible web has been going on for decades, and that the needle has barely moved. I think that is a fair assessment; however, I am less convinced by the explanation often reached, in that we simply have not pushed hard enough on conformance.

WCAG matters. The people who wrote it, maintain it, and teach it have done seriously good work. And public-sector rules such as the UK’s accessibility regulations put the subject on agendas that used to ignore it. Plenty of designers, developers, and accessibility specialists are doing great work every day. So, for clarity, this is an opinion about the system we have built around that work, and a thought on how we could improve more sites, faster, than waiting for everyone to absorb WCAG 2.

The same six failures, year after year

To many business owners, web managers, and developers, making a website accessible feels like standing at the foot of a sheer cliff. 

The Web Content Accessibility Guidelines (WCAG) 2.2 AA standard covers dozens of success criteria. Faced with dense jargon and an audit listing hundreds of issues that all have equal weight, many small and medium organisations simply freeze. 

Trying to make the web perfectly conformant has made basic accessibility feel impossible — so many business owners don’t even try.

Every year, WebAIM runs an automated check of the top one million homepages. In the 2026 WebAIM Million, 95.9% of those homepages had at least one detectable WCAG 2 failure. That figure had been edging down for several years; however, in 2026 it went back up.

The more useful number is about the kind of failure. WebAIM reports that 96% of all detected errors fall into six categories, and that those six have been the most common for the last seven years:

1. Low contrast text
2. Missing alternative text for images
3. Missing form input labels
4. Empty links
5. Empty buttons
6. Missing document language

These are ordinary authoring and design mistakes. They are also the barriers people hit every day: text they cannot read, images with no description, forms they cannot complete, controls with no name, pages a screen reader cannot identify.

Fixing them would not make a site conformant, and it would not cover every disabled person’s needs. Automated checks miss a great deal, including whether someone can actually finish the task. Even so, if more of the web got these six right, the everyday experience would change for a lot of people, and it would change sooner than another full pass of the guidelines.

What Alt at Source taught me

Take missing alternative text. Advocates have been asking editors to describe images for years. The errors are still on about half of the homepages WebAIM checks. The ask is reasonable. The workflow often is not. Alt text is treated as a patch after publish, at the moment when the person who knows the image has already moved on.

I tried to do something about that with Alt at Source: put the description on the asset when it is created or uploaded, so later publishing does not depend on someone remembering. Sadly, my campaign did not take off. Most of the sector ignored it, and most platform vendors ignored it too.

Although the campaign failed to gain traction, two things did happen: Craft CMS took the idea seriously and implemented it into its core, and WordPress 8 was coincidentally scheduled for release with this same feature. That is a small win. I still believe that when the tool makes the accessible choice the default, people do it. Training and audit findings are fine, but end up being weak substitute for effortless digital accessibility.

Conformance is a poor way to get the basics done

I am not arguing that audits are useless, or that edge cases do not matter. A missing name on a button and a broken identity check are different problems, and both can exclude someone. What I am arguing with is the habit of treating every failure as the same kind of emergency, then using that list to shame teams who were never going to clear it in one go.

A form can pass an automated WCAG check and still be so confusing that someone abandons it. A page can carry a conformant image and still ask a screen reader user to complete a task, a selfie, an ID check, that was never designed with them in mind.

The culture around the list has a cost. Dense, unprioritised reports frighten smaller organisations. Public telling-off does the same. People who already feel they are failing do not become careful. They postpone the work, add an accessibility widget to their website, or leave it to a specialist they cannot really afford. I have watched that for years, and I do not think more browbeating is why the six errors are still the six errors.

Legislation that everyone must meet WCAG 2 sets a high and complicated bar. That bar protects people when it is met. It also leaves a long gap in which most of the web is still failing the simplest tests. I have witnessed websites where simple issues stay unresolved for a year whilst a new website is being built or the dev team remediates the issues in planned sprints.

WCAG 3 is still a working draft, and it is several years from finished. As I understand it, the group writing it is trying to move away from the WCAG 2 model, where missing one applicable success criterion means you have failed. The drafts talk about showing progress, and about results that are not only pass or fail. If that holds, it is a welcome shift, and it deserves to be said. It also does not change what people are being asked to meet today.

So, whilst we wait, I am advocating for a plainer public standard of practice:

Do these six things properly, then keep going. This would get more sites over the first hurdle. Conformance can follow. It should not be the price of starting.

Start with the six, then keep going

If you own or build a site, you do not need a very expensive audit before you can improve it. Fix contrast. Describe images that carry meaning. Label every field. Give every link and button a name. Set the page language. That is not compliance theatre, and it is not the whole of accessibility. It is the work that currently accounts for almost all of the errors the annual scans keep finding.

If you practise in this field, you already know that. The pressure is to hand over a complete fail list and let the client feel the weight of it. I would rather we said, in public, that getting these six right is the credible first step, and that real journeys come next: can this person complete the task, in their actual circumstances, without a workaround we would never accept ourselves?

The specialists doing that already, quietly, on client work and in product teams, are the reason I still want to be in this field. The standards bodies maintaining WCAG are doing necessary work that a blog post cannot replace. My disagreement is with the idea that the way forward is another, stricter demand that everyone conform before they are allowed to say they tried.

The needle has not moved because we keep aiming at perfection. More of us doing these six simple things would change more websites faster than waiting for legislation and then aiming for full WCAG conformance.