August 12, 2026
HTML? More like How To Mess Up Layouts
Only 2.6% of the most visited websites have fully valid HTML
The internet mostly runs on broken code — and commenters are weirdly fine with it
TLDR: A new scan of popular websites found only 2.6% had fully clean code, with most pages breaking basic web rules in some way. Commenters split hard between “this doesn’t matter because browsers fix it” and “if accessibility and quality still fail this often, developers are just not trying.”
The web just got dragged by its own report card. According to Validate HTML, only 2.6% of the most-visited websites came back completely clean, while 9 out of 10 had at least one rule-breaking issue on their home page. The median site had 12 HTML problems, and one spectacular overachiever racked up 2,209. But if you thought this would trigger mass panic, the comments had other plans: the dominant vibe was basically, "So what?"
That’s where the real drama kicked in. One camp shrugged and said browser error-recovery is actually a feature, not a failure — if pages still look fine for most people, who cares whether the code is “perfect”? Another commenter went full nihilist with the bluntest take of the thread: "What does fully valid even mean anymore?" Ouch. On the other side, accessibility advocates were not laughing. The report says accessibility failures are everywhere, and commenters zeroed in on that, asking why basic features for disabled users are still treated like an afterthought when they’re often easy to add during development.
Then came the armchair auditors: some wanted repeat tests across countries and multiple days, suspecting websites may behave differently depending on location. And of course, there was the spicy modern jab: if your site is still messy with all the tools available, including artificial intelligence helpers, maybe it’s not hard — maybe you just don’t care. In other words: the web is broken, browsers are babysitting it, and the comments section is having a blast.
Key Points
- •ValidateHTML says it validated readable home pages from 2,655 of the Tranco top 5,000 domains and found 100,399 HTML spec violations and 17,599 CSS errors.
- •According to the article, 90.5% of measured sites had at least one HTML error, while only 69 sites, or 2.6%, were completely clean on errors and best-practice warnings.
- •The median site had 12 HTML errors, the mean was 37.8, and the highest count on a single site was 2,209, indicating a highly skewed distribution.
- •The most common HTML finding was `<style> element is not permitted as content under <div>`, appearing on 59.5% of sites.
- •The article reports that 65.1% of measured sites had a CSS error, and that accessibility issues such as skipped heading levels, unlabeled inputs, and empty-link text each appeared on more than a third of sites.