Only 2.6% of Most Visited Websites Have Fully Valid HTML

Widespread HTML Validation Issues

An analysis of the Tranco top 5,000 websites revealed that valid HTML is an exception, not the norm. Only 2.6% of these highly visited sites had completely clean HTML on their home pages, indicating a significant deviation from web standards across the internet.

The study found a skewed distribution of errors, with the median site exhibiting 12 HTML errors and the mean reaching 37.8 errors. The worst-performing site had 2,209 errors, highlighting a broad spectrum of compliance levels among popular websites.

Root Causes of Errors

The primary reason for the accumulation of errors is attributed to browsers' robust error recovery mechanisms, which allow sites to function despite invalid code, removing the immediate incentive for developers to fix issues. The most common finding, "<X> element is not permitted as content under <X>," suggests structural problems often generated by frameworks composing components, rather than manual coding mistakes.

CSS errors were also prevalent, with 65.1% of sites having at least one. However, the median CSS error count was low, typically involving invalid property values that browsers silently drop, allowing stylesheets to render without obvious breakage.

Accessibility Concerns

The analysis also highlighted near-universal accessibility failures. Issues such as incorrect heading levels, unlabelled inputs, and links without text appeared on more than a third of the audited sites. With the European Accessibility Act becoming applicable in June 2025, these findings point to a significant compliance challenge for many websites.

Methodology

The data was collected in August 2026 by requesting the home page of each domain from the Tranco top 5,000 list over HTTPS. HTML validation was performed using html-validate with its standard preset, and CSS validation utilized csstree-validator with lightningcss for modern syntax adjudication.