87% of top websites fail web standards check — outdated code is to blame

Independent French developer Théo Ducreux launched the ValidateHTML project, which analyzed the code of 5,000 most popular domains for compliance with web standards. Out of the sample, only 2,656 sites were intended for human visitors, while the rest were technical domains such as Google's "gstatic."

The vast majority of sites displayed invalid HTML code that violates the specifications of the World Wide Web Consortium (W3C) and the Web Hypertext Application Technology Working Group (WHATWG). These standards are designed to ensure browsers and screen readers render pages consistently across different platforms. In total, researchers found 100,305 HTML violations, including errors in the latest HTML5 specifications, as well as 18,863 CSS errors.

87.2% of websites violated web specifications in at least one place. Only 12.8% had valid HTML code, and just 2.6% were completely "clean" — showing no errors or warnings at all. More than one-third of sites failed accessibility checks, meaning they are not properly processed by screen readers used by people with visual impairments. Alternative text for images, which often contains critical information, was missing on 20.4% of websites. On 41.6% of pages, the code lacked ARIA labels, which help adapt content for blind users.

The findings highlight a widespread disregard for modern web standards among even the most popular sites. Developers often prioritize speed and visual design over code quality and accessibility, leaving millions of users with outdated or inaccessible content. The project's creator hopes the data will push web developers and site owners to adopt stricter validation practices and improve the overall health of the web.