How 59,400 ecommerce product pages were tested for accessibility

20 min read

The State of Ecommerce Accessibility 2026 tested one product page from each of 59,400 Shopify and WooCommerce stores, all ranked among the world’s top one million websites. 64% of those pages had at least one critical accessibility failure, the most severe kind. 96% failed at least one automated check against the Web Content Accessibility Guidelines (WCAG).

Why this is important for your store

Your conversion figures show how often visits turn into purchases. They can’t show whether a customer struggled to read a price, understand a product image or choose the option they wanted.

The State of Ecommerce Accessibility benchmark shows which of those problems turn up most often on product pages, so you know what to check on your own store and can give your agency, developer or content team something specific to investigate. The cause might be the theme, an app or a content decision, and it might not always be obvious when you use the store yourself. This review of product options that can’t be selected with a keyboard shows how a form on your product page can look right and still stop a customer choosing what they want.

What was the research designed to measure?

The study was built around one question:

How often do automatically detectable accessibility failures appear on Shopify and WooCommerce product pages from stores ranked among the world’s top one million websites?

One page per store means every store counts equally in the percentages, whether it sells 50 products or 50,000. The trade-off is that one page can’t speak for the whole store.

1. Start with stores people actually visit

The starting list came from HTTP Archive Opens in a new window , a public dataset which crawls millions of websites monthly. Its website list is drawn from the Chrome User Experience Report, which collects real-world browsing measurements. HTTP Archive also records which technologies each site uses, including its ecommerce platform.

The export was queried on 25 June 2026 and covered websites ranked in the global top one million. Every domain identified as Shopify or WooCommerce went onto the two store lists. Shopify and WooCommerce were chosen because their markup allows each store’s platform to be detected reliably.

Using HTTP Archive also means the store list comes from a known public dataset, each store’s platform was recorded before any accessibility result existed, and every store’s traffic ranking comes from the same source.

HTTP Archive supplies the ranking in bands rather than as an exact league table, so the report uses three groups:

  • Top 100,000
  • 100,000 to 500,000
  • 500,000 to one million

These groups describe relative website popularity. They are not estimates of revenue, sales or company size.

2. Select one product page from each store

Each store contributed one product page. The first ten products in the store’s public catalogue were requested, and the first one whose URL didn’t look like a gift card, coupon, shipping-protection add-on, sample or test product became the store’s page. If all ten looked like that, the first was used anyway, so some unusual products may have slipped through.

3. Confirm the page is really a product page

An address alone is not enough. A product may have been removed, or the URL may now redirect to a home page or login screen.

Every audit was checked before it was accepted. It had to contain accessibility results, and it couldn’t have redirected somewhere other than to another product page.

4. Run the same audit on every page

Each product page was audited with Lighthouse, Google’s website-auditing tool. Lighthouse’s accessibility checks come from axe-core Opens in a new window , an accessibility-checking engine. The versions of both were fixed for the whole audit, so every page was checked against the same rules.

All 59,400 pages were audited between 29 July and 4 August 2026 using:

  • Lighthouse 13.4.1
  • axe-core 4.12.1

Shopify and WooCommerce pages went through the same queue, so any difference between the platforms can’t come down to when they were tested.

What counts as an accessibility failure?

WCAG Opens in a new window groups its requirements into three levels: A, AA and AAA. In this article, an accessibility failure means Lighthouse failed the page on at least one automated check tied to a Level A or AA requirement.

57,013 of the 59,400 pages (96%) had at least one detected accessibility failure. Each page is counted once, however many failures it has.

Lighthouse also reports best-practice issues that aren’t tied to WCAG A or AA. Those were kept in the data but left out of the 96% statistic.

What does a critical finding mean?

Critical is the highest impact rating axe-core gives a rule, and it’s reserved for blockers: problems that can stop a customer using part of the page at all. On a product page, that means things like an image with no text alternative, a button a screen reader can’t name, a form field with no label, or settings that stop customers zooming in to read the text.

37,874 of the 59,400 pages audited (64%) had at least one critical failure.

5. Keep the accessibility result separate from everything else

Each audit also recorded the third-party scripts and requests on the page, which is how accessibility overlays were detected. That information was kept apart from the accessibility results, so only Lighthouse’s accessibility checks could count as a failure or affect a score.

The original audit behind every result was kept, so any figure can be traced back to the audits it came from.

6. Turn individual audits into an aggregate benchmark

The dataset used for the published figures contains:

  • 39,976 Shopify product pages
  • 19,424 WooCommerce product pages
  • 59,400 product pages in total

Headline percentages and average scores are rounded to whole numbers. The underlying calculations retain their precision, and detailed data views can show it. A small decimal difference between two results should not be treated as a meaningful difference in customer experience.

7. Keep individual stores private

Only combined results are published. No store is named or ranked, and the domains and product URLs stay in the private research dataset.

What these results cannot tell you

  • This is not a complete store audit. One selected product page was examined per included store. The figures do not assess the rest of its catalogue, cart, checkout, payment process or customer accounts.
  • The figures do not measure whether customers could buy something. They come from automated checks, not observed purchase attempts, keyboard walkthroughs or screen-reader sessions across the sample. They cannot establish how many customers were blocked or how much revenue was lost.
  • No detected failure is not proof of accessibility. Automated tools cannot assess every requirement or interaction. A page with no qualifying failures, or even a Lighthouse score of 100, can still contain barriers that require human evaluation to find. W3C’s guidance on accessibility evaluation Opens in a new window explains why tools alone are insufficient.
  • The sample data is not necessarily a representation of all ecommerce stores across the web. It covers product-page audits from Shopify and WooCommerce stores in the source dataset’s top-million ranking. It does not cover every platform or stores outside that list. Stores that could not be resolved, validated or audited may also differ from those included.
  • Each audit is a snapshot. The results describe the page as it loaded at the time of testing. Later changes, other devices and states that require a customer to interact with the page may produce different findings. What happens to keyboard focus when a cart drawer opens is one example: it only shows up once a customer interacts with the page.

For Shopify stores, the line between the product page and checkout is especially important. Shopify controls its checkout, but the product page is down to each store. Where Shopify’s checkout accessibility ends and your product page begins explains why both need checking.

The benchmark is useful for finding recurring problems and deciding where to look first. Understanding the wider customer experience takes manual checks and testing with real people.

Checking your own store

A good place to start is your best-selling product page. See how it compares with other stores on the same platform, then use the free product page accessibility guide to turn what it finds into a brief for whoever makes the changes. The full findings are in the State of Ecommerce Accessibility 2026.

Questions about the method, or spotted something that doesn’t look right? Get in touch.

Next year

This is the first State of Ecommerce Accessibility report, and there are big plans for next year to make it even bigger.