State of E-commerce Accessibility 2026
We scan e-commerce homepages for WCAG accessibility failures as part of our own market research, using the same automated tool (a headless crawl plus axe-core) we sell as a monitoring product. As of this update, the corpus stands at 199 e-commerce homepages across fashion, home goods, food, pet, beauty, and outdoor verticals, most in the 10-200 employee range. This page reports the aggregate findings and is updated periodically as we scan more sites.
A note on the sample: this is not a random or statistically representative sample of e-commerce. We selected sites by researching brands across a range of verticals for our own outreach research, not by drawing from a fixed population such as "all Shopify stores." Treat these as real, unedited results from a broad convenience sample, not a scientific survey. Every stat below states its N; we do not publish a percentage from a sample too small to defend, and we do not extrapolate to "the internet" as a whole.
How much accessibility debt is on an average homepage
Across the 199-site sample, we logged 174 critical and 361 serious findings in total -- an average of 2.7 combined critical-plus-serious instances per homepage. (Moderate and minor findings, which we also log but weight less heavily, added another 496 instances across the sample.) A "finding" here means one axe-core rule violation rolled up by our scanner; a single rule can fire many times on one page, for example a product grid with 20 images missing alt text counts as one finding with 20 instances, not 20 findings.
The single biggest issue, by site
For each site, our report ranks findings by severity and instance count and calls out one as the top issue. Ranking all 199 sites by what that top issue was:
| Issue | What it means | % of sites (N=199) |
|---|---|---|
| image-alt | Images missing alt text | 27.1% |
| color-contrast | Text with insufficient color contrast | 17.6% |
| link-name | Links with no accessible name | 16.1% |
| button-name | Buttons with no accessible name | 12.1% |
| region | Content not contained in a landmark region | 4.0% |
| label | Form fields without a programmatic label | 4.0% |
| nested-interactive | Interactive controls nested inside each other | 2.0% |
| aria-required-parent | ARIA elements missing a required parent role | 1.5% |
| aria-required-children | ARIA container elements missing a required child role | 1.5% |
Each row links to our fix guide for that issue: what it is, who it blocks, and the exact code change that resolves it. The top four (image alt text, color contrast, link names, and button names) account for 72.9% of top findings across the sample -- the same handful of issues we cover in most of our outreach and in the product-page barriers post. Seeing the same four patterns dominate a 199-site sample, up from a 142-site sample a few weeks earlier with a nearly identical ranking, is a reasonable signal that they are structural, not cherry-picked examples.
Why this matters for a site owner
None of these issues are exotic. They are the kind of thing a component library or theme gets wrong once and then repeats across every product card, every icon button, and every form on the site. That is also why they are cheap to fix in bulk: fixing the pattern in one shared component (a product card partial, a button mixin, a form field include) usually resolves the issue everywhere it repeats, rather than requiring a page-by-page pass.
It also means a homepage-only scan, the kind we offer for free, is a reasonable first signal. If a shared component is broken on the homepage, the same component is very likely broken on product and category pages built from the same theme.
Methodology
Scans run against the public homepage of each site only (not full-site crawls), using our own scanner: a headless browser crawl plus axe-core, rolled up into critical, serious, moderate, and minor findings by our own severity model. Sites were sourced by researching e-commerce brands for cold-outreach research between June 9 and July 2026; the corpus grows as we add sites to that research, so the numbers on this page will shift release to release. We exclude any scan that failed outright (bot-blocked, DNS unreachable, etc.) from every stat on this page rather than counting a failed scan as a clean one.
barrierscan.com/blog/state-of-ecommerce-accessibility-2026/), and note the N for whatever stat you use, consistent with how we report it. Questions or a request for the underlying breakdown: hello@barrierscan.com.