For six straight years, the web slowly got a little more accessible. In 2026, that trend reversed. The WebAIM Million — an annual scan of the top 1,000,000 home pages — found detectable WCAG 2 failures on 95.9% of home pages, up from 94.8% a year earlier. The average page now carries 56.1 accessibility errors, a 10.1% jump from 51 errors in 2025.
The likely driver isn't neglect — it's growth. The average home page now ships 1,437 HTML elements, 14.3% more than the year before: more widgets, more third-party embeds, more auto-generated markup, and more opportunities for something to go unlabeled. Complexity outran care.
The same six failures, seven years running
What makes this useful rather than just discouraging: the WebAIM Million has flagged the same six failure types as the leading causes for seven consecutive years, and together they account for 96% of all detected errors. None of them require a redesign to fix.
- Low-contrast text — now present on 83.9% of home pages, up from 79.1%. This is the single most common failure, and usually a one-line CSS fix once you know where it is.
- Missing alternative text on images — affects more than half of all pages, leaving screen-reader users and, increasingly, AI crawlers with no idea what an image shows.
- Unlabeled form inputs — present on 51% of pages, the most common reason a contact or quote-request form silently loses submissions from assistive-technology users.
- Empty links and buttons, missing document-language attributes, and empty or missing headings round out the list.
Why a small business should care beyond "it's the right thing to do"
It's also the legally exposed thing to ignore. Plaintiffs filed 3,117 website accessibility lawsuits in U.S. federal court in 2025 — a 27% jump over 2024's 2,452 cases — and the majority of defendants were not large corporations: UsableNet's tracking found 64% of sued companies had annual revenue under $25 million. A typical demand letter settlement runs $5,000–$25,000, before a single line of code gets fixed. For a Flushing or Queens small business running on a tight margin, that is not a rounding error.
The overlap with AI search most businesses miss
Here's the part that rarely gets said out loud: the fixes above are the same signals AI crawlers rely on to understand a page in the first place. ChatGPT, Perplexity, and Google's AI Overviews don't render your CSS or guess at an unlabeled button — they read the semantic structure: heading hierarchy, alt text, labeled form fields, and real link text instead of "click here." A page that's easier for a screen reader to parse is, by construction, easier for an AI engine to parse and cite. Accessibility work and GEO work point at the same underlying HTML.
Five fixes, in order of leverage
- Run a free contrast and structure audit. WebAIM's own WAVE tool is free and flags exactly which elements fail, in seconds.
- Fix contrast first. It's the most common failure and the cheapest to fix — WCAG AA requires a 4.5:1 ratio for normal text against its background.
- Add real alt text to every meaningful image — not the filename, a description of what the image actually shows or does.
- Label every form field, including the ones that look self-explanatory. A placeholder is not a label.
- For bilingual sites specifically: audit both language versions separately. An accessibility pass done only on the English pages tells you nothing about whether your Chinese-language pages pass the same checks — contrast, alt text, and form labels have to be verified on each locale independently.
None of this requires a rebuild. It requires running a free scanner, then working down a short, concrete list — which puts a small business ahead of the 95.9% of the web that still hasn't.
