We Scanned 243 Real Websites for Accessibility. 93% Failed. Here's What Breaks Most (2026 Data)
We Scanned 243 Real Websites for Accessibility. 93% Failed. Here's What Breaks Most (2026 Data)
Most accessibility statistics you see quoted online trace back to one source: WebAIM's annual analysis of home pages. It's excellent work, but the web needs more than one dataset. So we analyzed our own.
A11yScope has run 1,012 automated accessibility scans since launch. After deduplicating to one result per domain (keeping each domain's most recent scan), that gives us 243 distinct real-world websites — small-business sites, dental and medical practices, e-commerce stores, SaaS products, agencies' client sites — scanned against the automated WCAG 2.1 AA rule set in axe-core.
This is what the data says. (For how severe these failures are — how many sites carry a critical, task-blocking barrier versus a minor one — see the companion accessibility severity breakdown.)
Key findings
- Only 17 of 243 websites (7.0%) had zero detectable WCAG violations. 93.0% of sites failed at least one automated check.
- The average accessibility score was 85.7 out of 100 — most sites are not disasters, but almost none are clean.
- The average site carried 4.1 distinct violation types.
- Color contrast is the single most common failure, affecting 62.6% of all sites scanned.
- Three of the top ten violations — missing image alt text, unnamed buttons, and broken viewport scaling — are rated critical impact, meaning they can make content completely unusable for assistive-technology users.
Methodology (read this before quoting us)
- Sample: 243 distinct domains, deduplicated from 1,012 scans (latest scan per domain), collected through June 2026 via A11yScope's free scanner.
- Engine: axe-core by Deque Systems — the same engine used by Google Lighthouse — run in a fully rendered headless browser, covering 38+ automated WCAG 2.1 AA and best-practice rules.
- Selection bias, stated plainly: these are sites whose owners (or their agencies) cared enough to run an accessibility check. The real-world failure rate across the broader web is likely worse than what we report here.
- Automated checks only: automated tools detect roughly 30–40% of WCAG issues. A site with zero automated violations is not necessarily fully accessible — but a site with automated violations is definitely not.
The 10 most common accessibility violations in 2026
| # | Violation (axe rule) | Impact | % of sites affected |
|---|---|---|---|
| 1 | Insufficient color contrast (color-contrast) | Serious | 62.6% |
| 2 | Content outside landmarks (region) | Moderate | 57.6% |
| 3 | Skipped heading levels (heading-order) | Moderate | 41.6% |
| 4 | Links without accessible names (link-name) | Serious | 27.6% |
| 5 | Duplicate landmarks without labels (landmark-unique) | Moderate | 21.4% |
| 6 | Missing or multiple <main> landmarks (landmark-one-main) | Moderate | 18.5% |
| 7 | No level-one heading (page-has-heading-one) | Moderate | 11.9% |
| 8 | Images without alt text (image-alt) | Critical | 11.1% |
| 9 | Links not distinguishable in text (link-in-text-block) | Serious | 10.3% |
| 10 | Buttons without accessible names (button-name) | Critical | 9.5% |
Just outside the top ten: viewport scaling blocked for zoom (meta-viewport, 8.2%, critical), redundant image alt text (8.2%), and focusable elements inside aria-hidden containers (7.8%).
What each failure means — and the fix
1. Color contrast (62.6% of sites)
Nearly two out of three websites have text that fails the WCAG 4.5:1 contrast ratio. The usual suspects: light-gray body text (#999 on white), placeholder-colored labels, and white text over hero images. The fix is mechanical — darken the text or the background until the ratio passes. Our color contrast guide walks through the exact ratios and tooling.
2. Content outside landmarks (57.6%)
Screen-reader users navigate by landmarks (<header>, <nav>, <main>, <footer>). Content sitting outside any landmark is invisible to that navigation style. Wrap page content in semantic regions — usually one <main> plus the standard furniture.
3. Skipped heading levels (41.6%)
Jumping from <h1> to <h4> because the <h4> "looked right" breaks document outlines that screen-reader users rely on. Style with CSS; structure with sequential heading levels.
4. Links without accessible names (27.6%)
Icon-only links (social icons, hamburger menus, card arrows) with no text and no aria-label. A screen reader announces them as just "link". See how to fix empty links and buttons.
8 & 10. Missing alt text and unnamed buttons (critical)
These two are rated critical because they fully block tasks: an unnamed button can't be confidently pressed, and an unlabeled informative image conveys nothing. They're also among the cheapest fixes in all of web development — an alt attribute and an aria-label.
What this means if you own a website
The encouraging read: the most common failures are also the most fixable. Color contrast, heading order, alt text, and link names are all afternoon-sized fixes, not re-platforming projects. A site at the average score of 86 can usually reach the high 90s in a day of focused work — our step-by-step fixing guide covers the workflow.
The sobering read: accessibility regresses. Every deploy, CMS edit, and new landing page can reintroduce these exact issues — which is why one-time audits drift back to failure within months, and why we built weekly automated monitoring into A11yScope.
Check your own site
You can run the same scan we used for this study — free, no signup — at a11yscope.com. It takes about six seconds and shows every violation with a fix-ready code snippet.
---
Citation: A11yScope Accessibility Scan Data Study, updated June 2026. 243 distinct domains, 1,012 total scans, axe-core WCAG 2.1 AA automated rule set. You're welcome to cite or republish these figures with a link to this page.