1 in 3 Websites Has a Critical Accessibility Barrier: A Severity Breakdown of 244 Real Sites (2026 Data)
# 1 in 3 Websites Has a Critical Accessibility Barrier: A Severity Breakdown of 244 Real Sites (2026 Data)
It is one thing to know that most websites have accessibility issues. It is another to know how bad those issues are. A skipped heading level is a nuisance; a button a screen reader cannot name is a wall. So we went back through our scan dataset and sorted every violation by the one thing that matters most to a real user: severity.
This is a companion to our study of the most common accessibility violations. That one answered what breaks most. This one answers how serious it is.
Key findings
- 34.4% of websites have at least one critical violation — an issue that, by axe-core's own impact rating, can completely block a task for someone using assistive technology.
- 82.4% have at least one serious-or-worse violation. The large majority of sites are not failing on minor technicalities; they are failing on issues that materially degrade the experience.
- The average site carries 4.1 distinct violation types; the median is 4, and the worst single site had 13.
- Not one site in the sample scored a perfect 100 — every single site had at least some checks that failed outright or needed human review.
- The critical barriers are concentrated in a short, fixable list: missing image alt text, unnamed buttons, blocked zoom, and broken ARIA.
Methodology (read this before quoting us)
- Sample: 244 distinct domains, deduplicated from more than 1,000 scans (latest scan per domain), collected through June 2026 via A11yScope's free scanner.
- Engine: axe-core by Deque Systems — the same engine behind Google Lighthouse — run in a fully rendered headless browser against WCAG 2.1 AA and best-practice rules.
- Severity = axe-core impact: each violation carries an impact rating of
critical,serious,moderate, orminor, assigned by axe-core based on how much it blocks an assistive-technology user. We report a site as "having a critical violation" if any element on the scanned page triggered a critical-impact rule. - A note on the score: A11yScope's 0–100 score is the share of automated axe checks a page passes cleanly (passes ÷ [passes + violations + checks needing manual review]). It is not a WCAG-conformance percentage. That is why a site can have zero flagged violations and still not score 100 — there are almost always checks that require a human to confirm.
- Selection bias, stated plainly: these are sites whose owners cared enough to run a check. The broader web is likely worse.
How severe are the failures?
| Severity reached | Share of sites |
|---|---|
| At least one critical violation | 34.4% |
| At least one serious or critical violation | 82.4% |
| Zero detectable violations | 7.0% |
The headline most people quote — "most sites have accessibility issues" — undersells the problem. The more honest framing is that four out of five real websites carry a serious-or-worse barrier, and one in three carries an outright critical one.
The critical barriers: a short, fixable list
Critical-impact violations are the ones that can fully block a task. The encouraging news is that they cluster around a handful of rules — and the most common ones are among the cheapest fixes in web development.
| Critical violation (axe rule) | % of sites |
|---|---|
| Images without alt text (image-alt) | 11.1% |
| Buttons without an accessible name (button-name) | 9.4% |
| Zoom/scaling blocked by viewport (meta-viewport) | 8.2% |
| Disallowed ARIA attribute (aria-allowed-attr) | 5.3% |
| Missing required ARIA children (aria-required-children) | 4.9% |
| Form field without a label (label) | 4.1% |
| Unnamed (select-name) | 2.9% |
The pattern: roughly half of all critical barriers are an alt attribute, an aria-label, or a that was never added — and the rest are broken ARIA, usually from hand-rolled widgets that would have been accessible if built from native HTML elements. Our guide to fixing empty links and buttons and form labels guide cover the exact fixes.
The serious barriers: dominated by two issues
One step down in severity, the picture narrows even further. Two rules account for the overwhelming majority of serious-impact failures.
| Serious violation (axe rule) | % of sites |
|---|---|
| Insufficient color contrast (color-contrast) | 62.7% |
| Links without accessible names (link-name) | 27.9% |
| Links not distinguishable in text (link-in-text-block) | 10.7% |
| Focusable element inside aria-hidden (aria-hidden-focus) | 7.8% |
| Nested interactive controls (nested-interactive) | 7.0% |
| without a title (frame-title) | 6.6% |
Color contrast alone affects nearly two-thirds of all sites — it is both the most common violation overall and the single biggest source of serious-impact failures. If you fix one thing this week, fix your contrast. Our color contrast guide has the exact ratios and tooling.
What this means if you own a website
The severity lens changes the priority order. You do not have to fix everything at once — you have to fix the critical and serious issues first, because those are the ones doing real harm and carrying the most legal risk.
The good news hiding in this data: the most severe problems are also the most concentrated and the most mechanical. Add alt text. Name your buttons. Don't disable zoom. Pass contrast. Label your forms. That short list clears the large majority of critical and serious barriers on a typical site — usually in an afternoon, not a re-platforming project. Our step-by-step fixing guide walks through the workflow, and the tools comparison covers what to scan with.
Check your own site
You can run the same scan we used for this study — free, no signup — at a11yscope.com. In about six seconds it shows every violation, ranked by the exact severity levels in this report, with a fix-ready code snippet for each.
---
Citation: A11yScope Accessibility Severity Study, June 2026. 244 distinct domains, 1,000+ total scans, axe-core WCAG 2.1 AA automated rule set, violations classified by axe-core impact rating. You're welcome to cite or republish these figures with a link to this page.