AccessLumens Research · July 2026
The State of EU E-Commerce Accessibility 2026
One year after the European Accessibility Act took effect, more than half of Europe’s biggest online stores still ship a critical barrier. And the retailer a French court ordered to fix its site? We could not get in the door.
The European Accessibility Act came into force on June 28, 2025, covering e-commerce, banking, travel, and telecom, with member states now enforcing it under real penalties. We scanned 170 of the largest online stores across eight EU countries, testing the homepage, a product page, and a transaction page on each with WCAG 2.2 AA checks plus scripted keyboard testing. The question we wanted to answer: a year in, did any of it actually work?
52.3%
Still ship a critical barrier
of 107 scored sites
42.1%
Fail WCAG 4.1.2
most-litigated criterion
27.6%
Blocked the scan
bot walls and CAPTCHAs
91.9
Average score
out of 100
The nuance that matters: the average score across the 107 scored sites was 91.9 out of 100, which sounds encouraging until you look underneath. 52.3% still carried at least one critical barrier: a control a screen reader cannot operate, a menu a keyboard cannot reach, a checkout that locks out anyone not using a mouse. A good score and a locked-out customer are not mutually exclusive.
Of the 170 sites attempted, 107 returned scannable results. 47 blocked the scan entirely with bot protection or CAPTCHAs. Another 16 served JavaScript shells with no usable content for a headless browser. That reachability picture is a finding in its own right.
The Carrefour problem
In November 2025, French disability-rights organizations filed the first lawsuits under the EAA against four grocery retailers: Carrefour, Auchan, E.Leclerc, and Picard. In June 2026, a French court ordered Carrefour to make its e-commerce site fully accessible under penalty of daily fines. We scanned all four.
Carrefour was bot-blocked. Bot-blocked on both carrefour.fr and carrefour.es. The retailer a French court ordered to remediate has deployed protections that block the very tools used to verify accessibility. We could not scan it at all.
Bot-blocked on both carrefour.fr and carrefour.es. The retailer a French court ordered to remediate has deployed protections that block the very tools used to verify accessibility.
Returned a JavaScript shell with no usable content for a headless browser.
Respectable, but still carrying non-critical barriers.
Multiple issues, though none classified as critical.
Draw your own conclusions. A year after the lawsuits and weeks after a court order, two of the four are completely unreachable by accessibility tooling, and neither of the remaining two is fully clean.
“The retailer a court ordered to fix its accessibility has deployed protections that block the very tools used to verify accessibility.”
Key findings
1. More than half of Europe’s biggest stores still ship a critical barrier.
52.3% of scored sites carried at least one critical issue, and 42.1% failed WCAG 4.1.2 (Name, Role, Value), the criterion most frequently cited in accessibility lawsuits on both sides of the Atlantic. One year of enforcement has not yet moved the enterprise tier.
2. A quarter of sites could not be audited at all.
27.6% of sites blocked the scan with bot protection: IKEA across four countries, MediaMarkt across five, Zara, Mango, El Corte Inglés, Fnac, Albert Heijn, Tesco Ireland. A wall that stops a scanner also degrades screen readers, voice control, and automated monitoring. When you fortify your checkout, you can end up keeping out the people the EAA was written to protect.
3. Accessibility overlays correlated with worse scores, not better.
The 7 sites running an overlay widget averaged 88.5, versus 92.1 for sites without one, a 3.6-point gap in the wrong direction. It is the first time this pattern has been documented at scale across EU e-commerce, under the EAA framework.
4. The sectors the law prioritizes are performing worst.
66.7% of telecom sites and 55% of banking sites carry a critical barrier. Electronic communications, banking, and transport are the sectors the EAA calls out by name, yet they sit at the bottom of the table.
5. A perfect score is achievable, and 3 retailers proved it.
Darty, Bol.com, Żabka each scored a clean 100 across every page we tested. Three countries, three sectors, three scales. Proof that conformance is a normal outcome, not a moonshot.
Three sites got it right
A perfect 100 with zero issues detected, across three different countries, sectors, and scales.
100
Darty
France
100
Bol.com
Netherlands
100
Żabka
Poland
Country by country
Scores cluster tightly between 90 and 95. The real differences show up in critical-barrier rates and in how many sites blocked the scan entirely. Sweden leads; the Netherlands has the highest critical-barrier rate; France and Spain had the worst reachability.
Critical-barrier rate by country
Share of scored sites in each country with at least one critical barrier. Lower is better.
| Country | Scored | Blocked | Avg score | % with critical | Top WCAG failure |
|---|---|---|---|---|---|
| Sweden | 15 | 4 | 94.9 | 33% | 2.4.1 Bypass Blocks |
| Poland | 11 | 2 | 93.4 | 27% | 1.4.3 Contrast |
| Netherlands | 15 | 7 | 91.7 | 67% | 2.4.1 Bypass Blocks |
| Germany | 19 | 9 | 91.4 | 58% | 2.4.1 Bypass Blocks |
| France | 12 | 9 | 91.2 | 58% | 2.4.1 Bypass Blocks |
| Ireland | 12 | 3 | 91.2 | 58% | 4.1.2 Name, Role, Value |
| Spain | 11 | 9 | 91 | 55% | 4.1.2 Name, Role, Value |
| Italy | 12 | 4 | 90.3 | 58% | 4.1.2 Name, Role, Value |
The overlay finding
Accessibility overlay widgets are marketed as a quick fix for compliance. The data says otherwise.
With an overlay
88.5
average score · 7 sites
Without an overlay
92.1
average score
Overlays correlated with a 3.6-point lower average, not a higher one. Sites using FACIL’iti averaged 85.5, sites using Eye-Able 86. This is not a new finding in the US market, but it is the first time it has been shown at scale across EU e-commerce under the EAA.
The lawsuit-relevant picture
We mapped every finding to WCAG success criteria. The rows flagged EAA priority are the criteria most central to European Accessibility Act enforcement.
Failures by WCAG success criterion
Share of scored sites with at least one failure under each criterion.
| Success criterion | Sites affected | Occurrences |
|---|---|---|
| 2.4.1Bypass Blocks | 56.1% | 89 |
| 4.1.2Name, Role, ValueEAA priority | 42.1% | 220 |
| 1.4.3Contrast (Minimum)EAA priority | 29.9% | 358 |
| 1.1.1Non-text ContentEAA priority | 29.9% | 174 |
| 1.4.10Reflow | 29% | 38 |
| 2.4.4Link Purpose (In Context) | 26.2% | 163 |
| 1.3.1Info and RelationshipsEAA priority | 24.3% | 188 |
| 2.1.1KeyboardEAA priority | 17.8% | 67 |
| 2.4.7Focus VisibleEAA priority | 15% | 490 |
What is actually broken
The ten most common failures, by share of scored sites where each appeared at least once.
| Issue | Severity | Sites |
|---|---|---|
| Content not contained within landmarksStructure & Semantics | Moderate | 63.6% |
| No skip link before main contentNavigation | Moderate | 56.1% |
| Landmarks not uniqueStructure & Semantics | Moderate | 37.4% |
| No level-one headingStructure & Semantics | Moderate | 29.9% |
| Insufficient color contrastColor & Visual Accessibility | Serious | 29.9% |
| Horizontal scroll at 320px (reflow)Layout & Reflow | Serious | 29% |
| No main landmarkStructure & Semantics | Moderate | 29% |
| Links without discernible textNavigation | Critical | 26.2% |
| Illogical heading orderStructure & Semantics | Moderate | 25.2% |
| Images missing alt textImages & Media | Moderate | 24.3% |
Keyboards cannot reach your checkout
Scripted keyboard testing by country. Missing skip links are near-universal; Ireland has a focus-visibility problem; Germany and France lead on unreachable elements.
| Country | Focus trap | Focus not visible | Skip link missing | Unreachable |
|---|---|---|---|---|
| Germany | 15.8% | 21.1% | 63.2% | 31.6% |
| France | 8.3% | 0% | 66.7% | 33.3% |
| Netherlands | 0% | 20% | 73.3% | 6.7% |
| Sweden | 6.7% | 0% | 80% | 6.7% |
| Spain | 9.1% | 9.1% | 45.5% | 18.2% |
| Italy | 8.3% | 16.7% | 50% | 8.3% |
| Ireland | 8.3% | 41.7% | 25% | 8.3% |
| Poland | 0% | 9.1% | 27.3% | 9.1% |
By page type
Homepage, product, and transaction pages scored within a point of each other.
| Page type | Pages | Avg | % critical |
|---|---|---|---|
| Homepage | 105 | 91.2 | 47.6% |
| Product page | 11 | 91.3 | 45.5% |
| Transaction page | 70 | 92.4 | 42.9% |
By sector
Telecoms and banks, the EAA's priority sectors, sit at the bottom.
| Sector | Sites | Avg | % critical |
|---|---|---|---|
| Fashion | 10 | 95.7 | 10% |
| Electronics | 7 | 93.1 | 28.6% |
| Travel / Ticketing | 14 | 93 | 35.7% |
| Home & Furniture | 3 | 91.8 | 33.3% |
| Grocery | 13 | 91.6 | 61.5% |
| Banking | 20 | 91.6 | 55% |
| Telecom | 21 | 90.8 | 66.7% |
| Retail / Marketplace | 16 | 90.7 | 68.8% |
Sector and platform samples with fewer than five scored sites are shown for completeness but are too small to rank. The worst individual scores belonged to Orange ES (78.5), Intesa Sanpaolo (82), Harvey Norman IE (82.5), REWE (84).
The platform question, and why it does not apply here
Our US Local Government study found that the CMS vendor was the strongest predictor of accessibility outcomes. We expected a similar pattern in EU e-commerce. It did not emerge. Europe’s top online stores run overwhelmingly on custom or headless architectures. Zalando, Otto, Coolblue, Decathlon, Thomann, and dozens more build their frontends from scratch. Of 107 scored sites, 103 ran custom stacks; there was no Shopify, WooCommerce, or Magento at scale among these national leaders. At the enterprise tier the EAA is actually targeting, the platform is custom and the accessibility outcome is entirely on the engineering team.
What this means for the EAA
- Enforcement has not yet transformed the sector. A year in, a majority of Europe’s biggest online stores still ship critical barriers, and a quarter cannot even be reached by automated tools. The sectors the law prioritizes perform worst.
- The landscape is not uniform. Sweden and Poland show that lower critical-barrier rates are achievable; Darty, Bol.com, Żabka prove a perfect score is not theoretical.
- A high score is not conformance. Automated and scripted keyboard analysis catches a substantial but partial subset of WCAG. Issues needing human judgment, like meaningful alt text, logical reading order, and plain-language content, sit on top of these numbers, so the true barrier count is higher than reported here.
- The EAA has teeth. Whether companies choose to act before they bite is another question.
See where your site stands against Europe’s largest retailers
Run the same WCAG 2.2 plus keyboard scan we used for this benchmark, free and with no signup, and see your score, your critical barriers, and exactly what to fix first.
Scan your site freeMethodology
We scanned 170 of the largest e-commerce, banking, travel, and telecom sites across eight EU countries (Germany, France, Netherlands, Sweden, Spain, Italy, Ireland, Poland) using the AccessLumens scanning engine (SWBS-1.3 scoring) with identical configuration for every page. Each site was scanned on up to three pages: homepage, a product or service page, and a transaction page (cart, checkout, login, or account), all under WCAG 2.2 Level AA criteria with scripted keyboard analysis (focus traps, focus visibility, reachability, skip links). Sites that blocked the scan with bot protection or CAPTCHAs (47) were logged as blocked, and sites that returned JavaScript shells with no renderable content (16) as unscannable; both are excluded from scored averages but included in reachability statistics, leaving 107 scored sites. We enforced strict robots.txt compliance, so pages a site disallows were not scanned. Product-page coverage was limited (11 scored pages) because most EU sites use dynamic routing that made automated product-URL discovery unreliable; homepage (105) and transaction (70) samples are robust, and transaction pages primarily represent login and account pages rather than live checkout. E-commerce platform detection used response headers and page markup, verified in a browser for all 170 sites after an initial pass surfaced false positives. The EAA references EN 301 549, which requires WCAG 2.1 Level AA; our engine tests WCAG 2.2 AA, a superset, so a site failing here also fails the 2.1 AA bar. Scores use our SWBS model; this is a diagnostic snapshot, not a legal conformance determination (see our disclaimer). All scans complied with robots.txt directives and were conducted without coordination with, or advance notice to, the companies scanned. For the raw dataset, contact support@accesslumens.com.