UK Web Accessibility Benchmark

Executive Summary
Defining Criteria
Accessibility Vs Usability
Who Has Accessibility Requirements?
Assistive Technology
How do Common Accessibility Practices Impact Users?
Measuring Criteria
Identifying Criteria
Identifying Domains
Objective
Methodology
Results & Findings
Conclusion
Data Tables

Executive Summary                

There are nearly 14 million people in the UK with a disability (Big Hack). Over 7 million of these people’s disabilities are likely to impact the way they experience the internet. Not only is there an obvious ethical obligation to ensure websites are built to be as accessible and inclusive as possible, but there is also an enormous commercial incentive; the 2019 Clickaway Pound report estimates that 70% of users with disabilities will leave a site if they encounter accessibility issues, which amounts to the loss of online revenue to the tune of around £17.1 billion a year. At Erudite we have found through A/B testing simple accessibility improvements that we have been able to achieve significant uplifts in conversion rate, and therefore sales revenue for our clients!

This study aims to ascertain how widely the WCAG (Web Content Accessibility Guidelines) Accessibility Standards are adopted and adhered to by looking at a cross-section of URLs from the UK’s most popular domains. From this, we can learn what the most common pitfalls and failings are, how our sites stack up against industry peers, and how different industries stack up against the rest of the web.

The results of this analysis showed that generally, the most prevalent accessibility issues are among the most easily fixed. Low colour contrast was the most prevalent error type, with URLs having around 35 occurrences each on average. Empty links and images missing alt text were the next most prevalent errors. Interestingly, these are practices that have far wider benefits than accessibility alone! Improving colour contrast improves readability and UX for all users, and link anchor text and image alt text can have great SEO benefits.

This analysis found that, across all types of accessibility testing, the worst-performing industry was the adult industry. Government & Non-Profit, Banking and Financial Services, Academic and Health & Medical Services sites were consistently strong performers.

Defining Criteria

What is web accessibility?

W3C’s Web Accessibility Initiative defines web accessibility as “websites, tools, and technologies are designed and developed so that people with disabilities can use them.

More specifically, people can:

• Perceive, understand, navigate, and interact with the Web
• Contribute to the Web

Web accessibility encompasses all disabilities that affect access to the Web, including:

• Auditory
• Cognitive
• Neurological
• Physical
• Speech
• Visual

At its core, accessibility is a set of technical requirements that a web product (website, app, tool etc) must adhere to in order to be completely accessible to all users. The most well-known accessibility standard is the WCAG 2.0.

Accessibility Vs Usability

However, accessibility standards alone are not sufficient. Accessibility refers purely to technical aspects of a site and how it is coded. While ensuring accessibility requirements are catered for by our sites is an important step towards ensuring our design is truly inclusive, it does not give us a real-world view of how sites function in practice.

Usability addresses this issue and considers how sites can be navigated, interpreted and used by users with disabilities in practice. Usability criteria are typically more difficult to quantify but are no less important in assessing the true usefulness and ease of use of a site for disabled users.

Consider the following example from WAI:

Suppose you have the following image on your site, on a page that’s explaining how to charge your phone, and a visitor with a visual impairment is using a screen reader to help them interpret the page.

Imagine trying to understand the instructions if you cannot see the image clearly, but the alternative text is being read to you by a screen reader. The alternative text on the left is not helpful and is likely to delay the user’s experience unnecessarily. The alternative text on the right,  communicates the same message as the image and is much more useful to a user who cannot see the image clearly.

The important distinction  is that this image technically adheres to accessibility standards – it has alternative text. However, from a usability perspective (and in the real world), only the image on the right would suffice.

Usability is a crucial aspect of inclusivity on the internet and not a factor to be overlooked. However, in many cases, in order to have a usable site, accessibility standards must first be adhered to (to revisit our example above, the right-hand image could not have been usable without first having alternative text, which is the technical element that WCAG 2.0 requires. Usability, contrary to accessibility, has no official standard or criteria, which makes it particularly difficult to measure.

Therefore, this research will examine web accessibility criteria in order to establish an overall benchmark of accessibility practices online. Expanding this research in future to include usability criteria would give a more complete picture, but in the absence of an official usability standard or quantifiable criteria, including it is beyond the scope of this project.

Who Has Accessibility Requirements?

Accessibility affects a large proportion of people in the UK – many more than most people might imagine.

According to the disability charity Scope, there are 13.9 million people with disabilities in the UK, which fall within a broad spectrum of capabilities and requirements. The University of Cambridge classify these into the following:

• Sensory (e.g.  visual, auditory)
• Cognitive (e.g. thinking, communication)
• Motor (e.g. locomotion, fatigue, dexterity)

This range of capabilities means that users with disabilities require different features and considerations on websites in order to navigate, interact with and interpret them easily. Some users will use assistive technologies such as screen readers, some will have difficulty using a mouse, and may use their keyboard to navigate around a page or may find certain colour palettes difficult to read to name but a few.

The 2018/19 Family Resource Survey by the Department of Work & Pensions found that 19% of working-age adults and 44% of state pension age adults reported having a disability.

Some disabilities are more likely to be impacted by accessibility barriers online than others. The Clickaway Pound report found that the following impairments were most likely to be impacted: (numbers of people in each group from the Department of Work & Pensions Family Resource Survey 2018/19).

• Visual (1.7 million)
• Hearing (1.8 million)
• Manual dexterity (3.7 million)
• Neuro-diversity (3.2 million)

Further to this, of disabled state pension-age adults, 34% reported that they had dexterity impairments, with 23% reporting hearing impairments, 18% reporting visual impairments and 11% reporting some kind of neuro-diversity.

With the UK’s ageing population, (ONS 2018 estimate that in 50 years there will be an additional 8.6 million people in the UK over the age of 65) it is reasonable to assume that the number of people with impairments of this nature is going to increase too.

Assistive Technology

Many people with disabilities use assistive technologies (AT) to help them interact with the internet more easily. The Clickaway Pound report found that 71% of people with disabilities felt their experience on the internet was impacted by their disability, and 69% of those people used assistive technologies. They found the most common tools to be the following:

• Screenreader software (58%)
• Screen magnification (14%)
• Dictation software (13%)
• Other (10%) speech input & output software, scanning and reading software, specialised AT for things like dyslexia)
• Refreshable braille (5%)

(of 53% of disabled population who use AT according to Clickaway Pound)

To understand the impact of accessibility criteria in a real-world setting, we will categorise accessibility criteria according to compatibility with these types of assistive technology.

As not all people with disabilities use assistive technology, we will also categorise accessibility criteria according to which types of impairments are most likely to be impacted (visual, hearing, manual dexterity or neuro-diversity).

How Do Common Accessibility Practices Impact Users?

Users with disabilities will experience websites differently depending on the nature and extent of their disability. For example; not all users with a visual impairment will use a screen reader or pinch-zoom – many prefer to wear glasses, low colour contrast could impact users with visual impairments, but isn’t likely to impact users with a hearing impairment and great mouseless navigation is likely to be useful for some users with motor impairments, but less useful for those who prefer to use a mouse.

Whatever the disability, inaccessible sites can be incredibly frustrating for users. 71% of disabled customers said they would click away from a website they find difficult to use.

• Visually impaired – Visually impaired users can interact with websites using a range of assistive technology such as screen readers or refreshable braille, or may choose not to use any additional technology and interact with the site without intervention, or by using browser tools to increase the text size or zoom in. They may find small text or low colour contrasts between background and foreground elements particularly tricky or difficult to interpret. Images with text embedded in them (rather than having it overlaid as HTML), particularly if they are missing alt text and are important for navigation could harm the experience of these users, as could redundant, overly long or uninformative alt text. Missing skip links, poor use of structural elements (Heading structure and nesting, use of OL, UL etc.), poor or no form labelling and landmarking (e.g. use of nav, main, footer elements) can all be particularly troublesome for users with visual impairments.

• Hearing-impaired – Typically, the web caters quite well for users with hearing impairments, as it is commonplace for any user to want to navigate the internet with their sound turned off. However, it is important to bear in mind that for some people with hearing impairments, English is not a user’s first language; sign-language is. Consider this when producing copy and ensure that the main points are made clearly and concisely and that it is it easily readable for a wide range of reading abilities.

• Manual dexterity – Users with motor impairments or who lack manual dexterity may find using a mouse difficult or small touch screen difficult and may choose to use their keyboard or another form of assistive technology to navigate a site rather than a mouse. For these users, focus management (the practice of allowing users to use the tab key to jump between elements of a page) is very important. The tab order on a page should be logical (generally top to bottom and left to right) and all content should be accessible via keyboard. Where content is hidden or off-page, tabindex order can be altered to ensure it is only ‘tabbable’ when is the element is visible on the page. Use of features like skip links and correctly nested headings to ensure users can jump to the content on a page without having to tab through the navigation is also very valuable.
  
  For users who prefer to use a mouse, it is particularly important to ensure touchpoints and CTAs are large enough to tap or click on easily, with plenty of surrounding space to prevent users from making unintentional gestures. Complicated widgets or UI features should have a standard alternative, and users should have the ability to turn off the need for complicated gestures or phone tilting.

• Neuro-diverse – Neurodiversity encompasses an enormous breadth of requirements. This could include users with autism, epilepsy or learning difficulties to name a few. Each user will likely have their specific preferences, but general good practice would ensure that there are no excessive animations (particularly those that blink or flicker) and any auto-playing video or animated content can easily be turned off to prevent it being distracting or problematic for users with light sensitivity.
  
  Page copy is again, crucially important. Ensure the main points of your content are easily readable, simple and concise and break your content up with correctly nested headings, bulleted lists and supplementary imagery. Avoid figurative speech or excessive idiom where at all possible, as some users may find this difficult to understand.
  
  Where the user is required to give input or interact with a page to get a certain outcome, offer clear instruction and avoid the use of time limits where possible, as some users may find it difficult to read quickly, or may find the time pressure makes them anxious.

What is considered ‘accessible’ varies widely depending on the user and their preferences and requirements, and as such, this list is by no means exhaustive. It is intended to give a contextual background to the ways different digital practices could impact users, and the implications for users within these groups if they are not catered to. Great accessibility means great user experience for everyone, and there are many more accessibility criteria to consider that will have an impact upon your entire audience.

Measuring Criteria

To effectively assess accessibility on the internet today, we must identify the following:

• Criteria against which we will measure the accessibility of websites
• The most popular websites in the UK, so we can be sure that our sample is representative of internet usage in the UK.

Identifying Criteria

The most widely accepted accessibility standards are the WCAG2.0. This standard was adopted by the UK government as the acceptable standard of accessibility for public sector websites in September 2019, giving them until September 2020 to comply. WCAG2 also became an ISO standard in 2012, making it the internationally recognised web accessibility standard.

As previously discussed, both WCAG 2.0 and its successor WCAG2.1 contain numerous testing criteria that cannot be automated and must, therefore, be omitted for the purposes of this research.

Identifying Domains

Choosing a sample of websites in the UK to analyse poses some difficulties. Firstly, the sample needs to be large enough to span all industries, verticals and website types (editorial, ecommerce, services etc.) but also needs to represent a large proportion of the UK’s internet browsing behaviour in order to give the most accurate representation of users’ real-life experiences as possible. This means we need to establish a way to measure the relative popularity of websites in the UK and identify the most popular. Because, due to its very nature, the internet doesn’t conform to conventional international borders, identifying sites that impact internet users in the UK specifically brings an additional complication.

As individual site traffic data is not usually publicly available, we have used a combination of SEMRush, Sistrix and SimilarWeb data to identify top UK sites by Organic Traffic. The top 250 sites in the UK were extracted and categorised manually by industry.

Finally, we need to account for site size in our analysis. Measuring accessibility in terms of occurrences of failed tests per site does not give a fair representation of the site’s accessibility. A site with 100 pages and 100 failed tests is considerably different to a site with 10 pages and 100 failed tests, or 1,000 pages and 100 failed tests. Therefore, the 10 most visited pages for each site (in terms of Organic traffic in the UK, according to SEMRush) will be analysed.

Measuring Criteria

In order to assess the accessibility of such a large volume of sites, we have selected the WAVE API. WAVE checks against WCAG2.1 criteria and offers the benefit of delivering large amounts of data on each URL; returning the number of incidences of each of the criteria found on the URL (e.g. a page with 3 images missing alt text will return Alt Missing = 3). These criteria are subcategorised into the following categories;

• Error – broken accessibility features or errors that must be fixed.
• Alert – indications that there may be an accessibility issue.
• Feature – an element or feature on the page that may improve accessibility.
• Structure – structural elements that, when used correctly, will make content more easily navigated and interpreted.
• ARIA – ARIA semantic mark-up, which is often used to describe elements on a page to AT users.

WAVE’s API allows bulk analysis of sites against the following criteria:

Objective

The objective of this research is to establish a benchmark of the following:

1. How well sites are catering for individuals with accessibility needs in the UK.
2. Which industries perform above and below average and whether there are any outliers.
3. How accessibility differs between mobile and desktop sites and whether there are any industry-specific differences between desktop and mobile performance.
4. How many sites were found to be free of errors.
5. Which were the most commonly found accessibility errors and pitfalls and whether there are any industry-specific patterns.
6. What proportion of sites were found to be missing basic accessibility features, such as skip links and ARIA.

Methodology

The sample of URLs were selected my drawing a list of the 200 most popular domains in the UK (in terms of Organic visibility), using a combination of SEMRush, Sistrix and SimilarWeb data. For each of these domains, the top ten URLs (in terms of Organic traffic) were selected.

Domains were then manually categorised by primary industry.

The WAVE API was then used to measure the accessibility of each of the URLs selected. The WAVE API measures against 104 criteria (see Appendix…), which are categorised into 5 subcategories; Errors, Alerts, Structure, Features, ARIA and Contrast. For each URL, the frequency of the occurrence of each criteria on the page is reported by the API (e.g. a page with one missing Table Header will return the value  Table Header Missing = 1).

Errors are criteria which indicate that something is explicitly missing or broken, where Alerts are indications that something may be missing or broken but requires manual confirmation.

Structure, Features and ARIA do not report frequency of errors, but frequency of a certain element, feature or mark-up that is important for accessibility being present on a page. The absence of these indicates a lack of catering for users with accessibility needs.

Finally, Contrast refers to the frequency of low colour contrast appearing on a page.

NOTE: Data collection was carried out in April 2020, during the early stages of the UK Lockdown as a result of Coronavirus. At this time, many sites were unable to cope with demand and had implemented queuing systems to access content and returned server errors as a result when crawled. These sites were removed from the sample.

Results & Findings

Prevalence of Accessibility Errors

Overall, sites were found to have over 25 errors per page on average. This was found to be slightly higher on desktop than on mobile, with desktop seeing 25.72 errors/page on average and mobile seeing 24.94.

Industry-Specific Performance

Overall, performance varies significantly by industry:

• Banking and Financial Services came out on top with just 3 errors/page on average.
• Government & Non-Profit sites followed, but with a considerable jump to just over 15 errors/page on average.
• The worst performing industry was found to be Adult, for which sites averaged over 41 errors/page!

However, errors are not the only indication of accessibility issues. When looking at the average alerts per page, we note some interesting patterns:

• Across all categories, there were considerably more mobile alerts than desktop alerts.
• The adult industry was the worst-performing industry in terms of alerts.
• Banking & Financial Services is not the best performing industry, with Government & Non-Profit sites having the least alerts per page on average. Social Media, Business Services and Health & Medical Services all scored above Banking & Financial Services, all scored below 1 alert per page on average.

When looking at contrast errors we see that:

• Adult sites are still the poorest performers. However, Social Media, Travel & Leisure and Health & Medical Services also score poorly.
• Banking & Financial Services were the top performers, with slightly less than 2 contrast errors per page on average.

Mobile Vs Desktop Performance

Overall, no significant difference was found in the number of errors or contrast errors per page on mobile vs desktop devices, but as previously discussed, we did see considerable differences in the average number of alerts per page on mobile compared to desktop, particularly in the Adult and Automotive industries.

When comparing average errors per page, Social Media was shown to be, perhaps unsurprisingly, more accessible on mobile than desktop, with mobile pages having just over 15 errors/page on average and over 20 errors per page on desktop.

Travel and Leisure showed a similar pattern, with around 21 mobile errors/page on average to desktop’s 25.

The biggest outlier, however, was Business Services, which averaged slightly under 21 errors/page on mobile devices, but almost 33 errors/page on desktop.

While Social Media and Travel & Leisure aren’t particularly surprising findings, given the very ‘mobile-first’ nature of their usership, it is surprising to see that Business Services underperform so heavily on desktop devices, given that ia more B2B nature would imply more desktop traffic.

Interestingly, there were no industries found to perform significantly worse on mobile than on desktop. The biggest difference between device type in this regard was found in the Automotive industry, with just over 20 errors/page on mobile compared to just over 17 errors/page on desktop.

When looking at alerts, we consistently see far greater numbers of alerts on average for mobile sites than desktop ones. Automotive sites had the greatest discrepancy, with desktop sites having fewer than one alert on average and mobile having over 5.

Few significant differences between contrast errors on mobile and desktop were noted, except for Consumer Services sites, which, have around 37 contrast errors/page on desktop, compared to just over 20 on mobile.

Error Free Sites

Of the 191 sites surveyed, just 3 were completely free of accessibility errors; 1 from the academic industry, 1 retail site, and 2 classified as ‘other’. This is around 1.6% of the sample.

While this does appear shocking, many sites were found to have relatively low numbers of accessibility errors, with 13 (6.8%) sites averaging 1 or fewer errors per page, and 41 (21.4%) having fewer than 5 errors/page on average.

Distribution of errors/page shows that most sites have <10 errors/page on average, indicating that while the sites currently have issues present, resolving these should not be complex or resource intensive.

Most Common Accessibility Issues

The criteria we define as ‘errors’ are not the only issues that can negatively impact the experience of a user with accessibility needs on a site. Alerts indicate that there are potential issues that require further investigation, and Contrast Errors indicate an area of low colour contrast that may make the page more difficult to read and interact with for users with visual impairments.

This study found that sites surveyed had slightly over 1 error per page and slightly under 2 alerts. However, by far the most common error type found was contrast errors, with sites having just under 33 contrast errors per page on mobile and over 38 contrast errors per page on desktop.

Errors

Alerts

Most Common Accessibility Issues by Industry

Interestingly, while mobile and desktop devices show similar error and alert trends, we can see vastly different patterns when looking at domains from different industries.

Academic

Typically, academic sites had fewer of each error type than the average across all analysed sites. However, we do see that missing alt attributes seem to be far more frequent for these sites. Content of an academic nature is likely to be accompanied by diagrams or other functional images that support the written content, rather than simply being decorative. As a result, ensuring these images have high quality, descriptive alt text is particularly important to ensure the content is fully accessible.   

Adult

Adult sites were found to be the worst performers of all industries analysed. Compared to the whole sample of analysed sites, this appears to be largely driven by missing alt links, empty buttons and most notably, contrast errors. Not adding alt text to linked images has two consequences for accessibility; firstly, it means that for users of screen readers or other AT, they may not be able to understand the contents of the image they are viewing and secondly, they will not know what destination to expect if they follow the link. Empty buttons have a similar result – with no text it can be difficult for users to predict what action the button may trigger, or where it may link to.

By far the most  common issue on adult sites were contrast errors, which  severely impact the readability of text for all users but is especially problematic for visually impaired.

Automotive

Automotive sites mirrored the patterns we see across the wider sample, apart from empty links, which were less frequent on automotive sites than the whole sample average.

Banking & Financial Services

The top-performing industry in this study, sites in the Banking & Financial Services category had fewer than average occurrences of all error types. Sites in this category performed particularly well for colour contrast, with negligible contrast errors detected across this sector. It is important to note that (across all sites) the URLs tested were high Organic traffic URLs. This means that particularly for the banking sector, these are likely to be informational pages. While it is important that these pages are highly accessible, it does not mean that other areas of the site that aren’t crawlable (e.g. online banking interfaces) reflect the same levels of accessibility.

Business Services

Sites in this category had considerably fewer contrast errors than the sample average, possibly due to differences in branding and colour palettes of B2B businesses compared to consumer sites. However, business services sites had more frequent occurrences of missing alt attributes, missing labels, and empty buttons. While missing alt attributes may not affect the site’s accessibility (it is not necessary to add alt text to images with no function beyond being decorative), missing labels can make it difficult for users of AT to fill out forms (which may also negatively impact conversion) and empty buttons can make navigation unnecessarily difficult.

Consumer Services

Consumer services sites showed the opposite to Business Services, with far more contrast errors than the sample average, but fewer missing form labels. Perhaps this is a result of consumer-facing brands typically being more colourful and vibrant than B2B sites, with more frequently rotating campaigns and seasonal changes which increase the margin for error.

Government & Non-Profit

Due to legislation coming into force in September 2020 that requires all government-owned sites to comply with WCAG AAA, we expected Government & Non-Profit to be the strongest performer, but they followed Banking & Financial Services to become the second-placed category. These sites typically had considerably fewer than average numbers of contrast errors and empty links, which is indicative of the typically blander designs of most government-owned sites compared to many highly visual, branded consumer sites. However, Government & Non-Profit sites had considerably more missing Table Headers than the sample average. This may be, in part, due to the fact that these sites have more tabulated data than most. Missing table headers can make tables very difficult for users of AT to interpret.

Health & Medical Services

Sites in the Health & Medical Services industry typically showed similar error patterns to the sample average. However, contrast errors again were found to be far higher than average (around double) across these domains. It is unclear what would cause these sites to score so poorly in terms of contrast, as they are typically informational and don’t tend to have the same level of highly visual, branded content that we see in the more poorly performing consumer-facing sites.

News & Media

News & Media sites tend to underperform in a few areas; missing alt links and alt attributes, empty links and contrast errors. It is easy to see how, in the world of fast-paced publishing and real-time news reporting, adding alt text to imagery and linked images, or ensuring links have a descriptive text element or alt text could slip by the wayside. However, the quickly evolving nature of the content on these sites is exactly why these attributes are crucial for accessibility. Users of AT (Assistive technology )or those with visual impairments who rely on these attributes to navigate a site may be able to familiarise themselves with the layout and navigation without these attributes, if they frequent the site and it remains more or less the same every time they visit (for example, an online store). However, if, as with most news and media outlets, the content is constantly updating and being refreshed, it becomes even more important to provide adequate signposting and labelling to ensure accessibility.

Retail

Retail sites scored below average in the use of alt links, alt attributes, missing form labels and, as we see with other consumer-facing sites, contrast errors. As with consumer services sites, it is reasonable to assume that missing alt attributes and alt links could be in part due to the quickly evolving retail landscape, changing promotions and campaigns that make adding alt attributes too resource-heavy. In many cases on these kinds of sites, they are image-rich purely for decorative purposes (for example banner imagery). Non-functional images needed have alt text, so missing alt text on these sites may not be indicative of a serious issue in all cases. However, linked images missing an alt attribute can make navigation particularly difficult, especially where important text that describes the link destination is embedded in the image and not overlaid as HTML and included within an alt attribute.

Forms, conversely, tend to be more static in nature. This indicates that form labelling issues may not be a symptom of fast-changing content slipping through the net and highly decorative elements that do not require special accessibility considerations but are the result of them never being accessible in the first place. It is likely that resolving form labelling issues once will result in long-term accessibility improvements.  

Social Media

Social Media sites scored below average for contrast errors, empty links and empty headings. Interestingly, some Social Media sites do an excellent job of ensuring some aspects of accessibility are catered for. For example, particularly low levels of images missing alt text were found amongst these sites, which is amazing given the highly image-rich nature of these sites. Indeed, Facebook automatically adds alt text to all images uploaded and allows users to override this with their own alt text if they wish. It is worth noting that as the pages analysed here were the most popular in terms of organic traffic, the pages analysed are likely to tend towards informational and signup pages, as opposed to the platform’s actual interface. Analysis of password-protected areas of the site (profiles, newsfeeds, messaging functionality) may show a different pattern.

Travel & Leisure

Finally, Travel & Leisure sites scored most poorly on contrast errors (with almost double the average number of contrast errors per page) and empty links. Again, the highly consumer-facing, visual nature of travel and leisure sites lends itself to contrast errors, but also means that they should be relatively easy to resolve. However, empty links, especially if they are the result of linked images with no alt text, can be detrimental to site navigation. These should be resolved.

These sites also tend to have slightly above average numbers of images missing alt text. This could be indicative of an issue, or a symptom of their design being heavily reliant on decorative, but functionless, imagery.

Sites Missing Skiplinks

Skip links allow users to skip their focus to the main content of the page, without having to manually bypass boilerplate content such as the navigation and header. This is particularly useful for screen-reader users (typically those with visual impairments), or users who prefer to use the tab key to navigate around a site rather than a mouse (often users with motor impairments, but also users who simply prefer not to use a mouse). Simple to implement, skip links are an accessibility (and usability staple).

Sites Missing ARIA

ARIA (Accessible Rich Internet Applications) attributes provide descriptive labels for interactions that are not natively compatible with assistive technologies such as screen readers. These could include live-updates content feeds, UI components like sliders, remove from cart buttons on checkouts or any other (typically JavaScript driven) dynamic component that does not automatically have an alternative interaction for AT. The absence of ARIA attributes on a page could indicate that the page will be difficult for users of AT to interpret.

Again, adult sites were the poorest performers, with almost half of the tested URLs not having a single ARIA attribute, both on mobile and desktop.

Social media, Automotive were the strongest industries, with none of the tested URLs missing ARIA. Academic, Health & Medical Services, and Travel & Leisure sites followed close behind, all with less than 5% of URLs having no ARIA attributes.

Conclusion

This piece of research finds that accessibility practices are poorly adopted across the majority of URLs. Some industries (namely Banking & Financial Services), performed strongly, with very small numbers of errors being reported per URL, while others (Adult) performed exceptionally poorly, with the highest total number of errors (mainly driven by contrast errors), alerts and contrast errors. The Adult industry also had the highest frequency of missing Skip Links (with 100% of sites missing Skip Links) and missing ARIA attributes (just under half of Adult URLs had no ARIA attributes).

However, the most common accessibility issues were not exclusively related to accessibility but have much wider UX and SEO benefits. Images missing alt text, links without anchor text, colour contrast issues, redundant links and redundant titles and small text are not just accessibility issues – they are omittances of general good practice. They are also typically simple to resolve, with minimal design or developer resource to implement.

---

Data Tables

Average Frequency of Errors

Average Frequency of Alerts

Average Contrast Errors/Industry

Issue Types by Industry (Mobile Vs. Desktop)

Average Frequency of Accessibility Issues

Accessibility Issue Average Frequency by Industry