Original research · 2,647,654 scans

Web Accessibility Statistics 2026

What we found after running automated WCAG 2.1 AA checks on 2,647,654 websites — roughly 2.65 million distinct domains and 5.2 million pages — between March 25 and June 7, 2026.

Free to cite with attribution to WCAG Repair (link to this page). All figures describe automated, machine-detectable WCAG 2.1 AA check failures; see the methodology for what that does and does not mean.

87.9% of websites fail at least one WCAG 2.1 AA automated check

Only 12.1% of the sites we scanned passed every automated check we run. The rest had at least one machine-detectable WCAG 2.1 AA failure.

87.9%
of sites had at least one WCAG 2.1 AA failure (2,327,900 of 2,647,654)
12
median issues per site (mean 49.8; 90th percentile 122)
36.0%
of sites had at least one critical issue (952,094 sites)
16.2%
disable pinch-zoom — about 1 in 6 sites block users from zooming

Across all sites we counted 131,732,989 individual issues. We define one "issue" as a single failing element instance (one axe-core violation node), not a rule type — a page with ten unlabeled images counts as ten issues, not one.

The most common accessibility failures

The share of all 2,647,654 scanned sites where each check appeared among the site's top violation types. Plain-English explanations describe what the check actually tests — not legal status.

Accessibility check (WCAG 2.1 AA) What it tests Severity % of sites
All page content contained by landmarks Content sits outside <header>/<main>/<footer> regions, so screen-reader users can't jump between page sections. Minor 58.5%
Minimum color-contrast ratio Text or UI fails the 4.5:1 (normal) or 3:1 (large/UI) contrast thresholds, making it hard to read for low-vision users. Major 45.2%
Page should have one main landmark No <main> region, so assistive tech has no "skip to main content" target. Minor 44.7%
Page should contain a level-one heading No <h1>, so the page lacks a clear top-level title for navigation by heading. Minor 31.3%
Links must have discernible text A link has no text a screen reader can announce (e.g. an icon-only or empty link) — this is about an accessible name, not a broken URL. Major 30.1%
Heading levels should only increase by one Heading levels skip (e.g. <h2> straight to <h4>), breaking the document outline. Minor 28.8%
<html> must have a lang attribute The page declares no language, so screen readers may use the wrong pronunciation. Major 17.7%
Zooming and scaling must not be disabled A viewport meta tag blocks pinch-zoom, preventing users from enlarging text. About 1 in 6 sites. Critical 16.2%
Documents must have a <title> No page title, so browser tabs and screen-reader page announcements are blank. Major 14.5%
Images must have alternative text An <img> has no alt attribute, so screen-reader users get nothing in place of the image. Critical 13.7%
Buttons must have discernible text A button has no accessible name (e.g. an icon-only button with no label), so its purpose is unclear to assistive tech. Critical 6.5%
Frames must have an accessible name An <iframe> has no title, so screen-reader users can't tell what the embedded content is. Major 6.3%
Headings should not be empty A heading element contains no text, leaving an empty entry in the heading outline. Minor 5.7%
Links must be distinguishable without color Links are set apart from surrounding text by color alone, which color-blind users can't perceive. Major 4.8%
Form elements must have labels An input has no programmatically associated label, so screen-reader users don't know what to type. Critical 4.6%

Each per-check percentage is a lower bound. We record the top violation types per scan, capped at 10 entries; only 3.3% of scans hit that cap, but for checks that rank below a site's top 10 the true prevalence is higher than shown.

How many accessibility issues sites have

Most sites are not borderline: among sites with any failures, a large share have dozens or hundreds of failing elements. 13.1% of all sites had 100 or more issues.

36% of websites have at least one critical accessibility issue

We map axe-core's impact levels to three tiers: critical (axe "critical"), major (axe "serious"), and minor (axe "moderate" and "minor").

36.0%
have ≥1 critical issue (952,094 sites) — avg 4.37 critical issues per site
76.0%
have ≥1 major issue (2,013,477 sites) — avg 22.12 major issues per site
82.2%
have ≥1 minor issue (2,177,439 sites) — avg 23.26 minor issues per site

What these numbers do — and don't — tell you

The vast majority of websites have accessibility problems that an automated tool can catch in seconds. Many of the most common — missing alt text, low color contrast, unlabeled form fields, blocked zoom — are also the issues that come up most often in real accessibility complaints, and they are usually quick to fix once you know where they are.

These figures measure conformance with WCAG 2.1 Level AA, the technical standard most accessibility laws reference. They are not a legal assessment: passing every automated check is necessary but not sufficient for full conformance, and we make no claim here about any site's legal status. For how WCAG maps to specific regulations, see our compliance hub and the ADA compliance overview.

Want to know where your own site stands? Run a free accessibility check — it uses the same axe-core engine and WCAG 2.1 AA ruleset behind these statistics, scans up to 10 pages, and returns every issue by severity with no signup.

Check your website free →

How we collected this data

  • Sample. 2,647,654 completed accessibility scans (about 2,647,455 distinct domains) covering 5,200,275 pages, run by the WCAG Repair scanner between March 25 and June 7, 2026. We treat one scan as one distinct domain. The sample skews toward small-business websites and is not a random sample of the web, so it is not directly comparable to web-wide censuses.
  • Pages per scan. Each scan crawls up to 10 same-domain pages, but the average is 1.96 pages and 904,816 scans covered a single page — so these results primarily reflect home pages and top-level pages, not entire sites.
  • Engine and ruleset. Pages are loaded in a real headless browser and tested with axe-core — the same engine behind Google Lighthouse and Microsoft Accessibility Insights — using the wcag2a, wcag2aa, wcag21a, wcag21aa, and best-practice rule tags.
  • What an "issue" is. One issue = one failing element instance (one axe-core violation node). A check that fails on ten elements contributes ten issues. Severity tiers map axe impact levels: critical = "critical", major = "serious", minor = "moderate" and "minor".
  • Per-check percentages are lower bounds. We store each scan's top violation types capped at 10 entries; only 3.3% of scans reached that cap. For checks that fall outside a site's top 10, true prevalence is higher than the percentage shown.
  • Automated coverage only. Automated checks detect only the machine-detectable subset of WCAG 2.1 AA. Many criteria — meaningful alt-text quality, logical reading and focus order, captions, and more — require human review. True non-conformance is therefore higher than these figures, not lower.
  • Privacy. All figures are anonymized aggregates across the full population of scans. No individual site is identified, and we publish no list of scanned domains.
  • No legal claims. These statistics describe WCAG 2.1 AA automated-check failures only. They are not an assessment of any site's compliance with the ADA, Section 508, the EAA, or any other law.

Last updated June 7, 2026. To cite: "WCAG Repair, Web Accessibility Statistics 2026 (n = 2,647,654 scans)." A link to this page is appreciated.

See where your site lands in these numbers

Run a free WCAG 2.1 AA scan — same engine, same ruleset, results in under 2 minutes.