WebAIM.org has done one of the largest if not the largest automatic accessibility audit on one million of homepages considered important enough to deserve such attention.
I like the idea of such bulk automatic accessibility tests and actually made some myself, with different tools and added metrics, just for the sake of experience and to be able to research on my own datasets.
Some key findings and comments
I will not go through all the details and I invite you to read about them on the original webpage (opens in new window), but I want to make some comments that may add some value to the original report.
Home page complexity increased by 7.7%
This means that the web is getting more complex and pages are getting more elements. There is no mention about the semantic complexity – in my opinion it matters a lot if we add a bunch of non-semantic elements just for the sake of styling or if we add same amount of semantic HTML elements like links, buttons and so on. I would personally make that as an indicator as well, to monitor complexity that is exposed to end users and their assistive technologies.
96.8% of homepages fails WCAG 2
This basically means that none can conform to WCAG 2 on level AA in my book. When we consider that automatic tests can only discover up to approximately 35% of all WCAG failures we can conclude that the remaining 3.2% for sure has an error or many. We should also mention that sometimes automatic tests over-report or mis-report, but I am quite confident that it is difficult to find a page that conforms 100%. Even those that claim that may hide some errors under the hood in my experience.
Some improvements in most common WCAG failures are obvious
It seems like missing alternative text for images and missing form input labels are being worked on way better than before. Present texts can for sure be wrong and still fail but I like to see that they are improved. It means that designers, developers and content providers have better awareness about them and that it is getting more deserved focus. I guess that some of the motivation is also because of search engine optimization, but some is for sure because designers use them in their designs (labels for sure).
Same can be said for improved contrasts, still the biggest issue with 83.9% of pages failing it, but it is improving and that is a good thing.
Incorrect use of ARIA increases accessibility errors
I have noticed multiple times that developers can overuse ARIA and make things worse. It is wise to point to the first rule of ARIA – do not use ARIA if you don’t need to – use native HTML elements when they can do the job. Only add ARIA when you are out of options – and when you do use ARIA – make sure you are using it correctly.
ARIA seems to be the solution to all problems, but if developers do not invest into learning the details, the required child elements and other important rules it does for sure introduce more harm than good for users of assistive technologies. If at least developers used proper tools to prevent poor or wrong ARIA usage – like for example lint tools and automatic testing per component – it would for sure help a lot.
Home pages with ARIA present averaged 70% more detected errors than those without ARIA.
citation from report (opens in new window).
Skip to content links are still an exception
Skipping repeatable content is an important feature for keyboard-only users and when missing it’s a quite obvious indication about missing focus on accessibility. 86% of all pages missing it is a clear sign that they might not considered WCAG at all.
Please add skip links to your pages and save keyboard only users precious time.
Seems like public sector sites lead and shopping sites are almost the worst
According to the report the category that improved the best is “Law, Government and politics” – seems like they have the least errors of all. I would like to think that this is quite expected when considering they should lead by example and quite often also do. Because they have to. After thinking about Web Accessibility Directive in Europe it is totally clear the sector has some awareness about accessibility.
Leaving adult websites out we quickly discover that shopping is the worst category. That offers a tremendous potentials for accessibility specialists in my opinion. Especially in Europe – with European Accessibility Act being extremely relevant. I think they are really missing the opportunity and reach because of their low or non-existent focus on accessibility. Consider market growth alone when having accessible experiences.
Content management systems and it’s templates can improve or worsen accessibility at a scale
At the end of the day it comes down to templates and content, but CMS itself can also make or brake the accessibility at a scale. WordPress market share and abundance of its templates should maybe be isolated from all the others and compared differently, as others combined don’t even get close quantitatively. Was also missing Webflow there, probably too fresh to be a part of the million group.
JavaScript frameworks statistics may misguide us
I don’t think that framework itself has a lot to do with accessibility – you can write React JSX and make it accessible or not. It’s true that depending on some third party components may harm your accessibility if you trust them blindly, but it has actually nothing to do with the fact that you have to know about accessibility first.
RequireJS (JavaScript module loader) being used will not make your site less or more accessible by itself. Or Handlebars and so on.
So please consider the comment that WebAIM made below the table;
This does not necessarily mean that the frameworks caused these errors, but home pages with these frameworks had more errors than on average.
WebAIM on correlation of JavaScript frameworks and accessibility (opens in new window).
JavaScript libraries analysis is no different – I would never do a direct correlation. Tools like frameworks and libraries can often be misused when developers lack accessibility knowledge and the other way around. I think WebAIM is totally sincere about it but still – I must warn readers about the possible wrong correlations.
We can conclude – still a lot to improve
It is clear to anybody in the field of accessibility – there are almost no websites that are 100% conforming to WCAG. And to be a perfectionist here can even be dangerous – fixing code in a way to satisfy automatic tests, just to get “perfect” scores of automatic auditing tools is not a proper way to get the return on investment of your development budget.
Preventing real barriers before satisfying automatic tools – making the web to be more accessible for users with different disabilities should be the main focus.
It seems like we are still in the early phases but I like the positive impacts of awareness and also legislation that will for sure help as well. The main goal should still be – inclusivity and not discrimination – making web accessible for most possible.
