State of Digital Accessibility in Slovenia, May 2026

Nearly half of major Slovenian websites have no published accessibility statement

A new survey of eighty of Slovenia’s most-visited websites, published today to mark Global Accessibility Awareness Day, finds that 45% carry no accessibility statement at all. Another 13% publish something too thin to qualify as the Article 13 declaration the law requires. Only 34% reach the threshold of a real declaration, the kind that names the relevant Slovenian transposition act, identifies a conformance standard such as EN 301 549, and provides a usable feedback channel.

The survey covered ten sites in each of eight consumer-facing sectors. It comes nearly a year into Slovenia’s enforcement window under the European Accessibility Act, which became applicable across the EU on 28 June 2025 and is transposed into Slovenian law through ZDPSI, in force since April 2023. Public-sector sites have been subject to the older ZDSMA since 2018.

Public sector leads, retail trails

The sector picture matches the pattern visible in equivalent surveys across the EU. Public-sector sites scored highest, on a mean of 10.8 out of 16 structural and statement checks, followed by utilities at 10.0 and banking at 9.8. Retail and online commerce sat at the bottom of the list with a mean of 6.3, and ecommerce was the only sector where zero of the ten sampled sites carried a real Article 13 declaration. The order is consistent with the directives’ age: public-sector sites have been under the Web Accessibility Directive since 2018, while ecommerce only came into scope under the European Accessibility Act in June 2025.

The cheapest fix is also the most missed

Of the structural checks run on each home page, the simplest accessibility provision in the WCAG standard, a “skip to main content” link at the top of the page, was present on only 22 of 80 sites. That works out to roughly one in four. Skip links are one anchor and two lines of CSS. They sit at level A in the WCAG hierarchy, the lowest bar in the standard. Three quarters of major Slovenian sites have skipped them.

Overlay widgets remain in use

Seven sites in the sample have an accessibility-overlay widget installed. These tools, sold by a handful of US-based vendors, inject a JavaScript bar that promises one-click accessibility through font-resize and contrast toggles. The disability community has documented their shortcomings repeatedly, and EU regulators have issued multiple statements warning that an overlay does not substitute for compliance with the underlying technical standard EN 301 549. Five of the seven overlay-running sites in the survey sit in a single sector.

What the home page tells the visitor

Only two sites out of eighty mention any legal framework on the home page itself, whether ZDSMA, ZDPSI, the EAA, WCAG, or EN 301 549. For 97% of the sample, a visitor curious about a site’s accessibility status would have to find a footer link first, and on 45% of sites that footer link doesn’t exist either.

What the survey did not measure

The scan in question is HTTP-only. It does not measure colour-contrast violations, missing accessible names on buttons and links, or any other condition that requires a rendered page to evaluate. Those checks need a full browser-based audit using a tool such as axe-core, which the survey authors say will follow as a second pass. The headline structural and statement findings, however, are reliable on what they cover.

Full report

Why this exists

The European Accessibility Act became applicable across the EU on 28 June 2025. Slovenia transposed it through ZDPSI (Zakon o dostopnosti proizvodov in storitev za invalidne osebe) in April 2023, pairing with the older ZDSMA that has covered public-sector sites since 2018. We wanted to know what the published web looks like in Slovenia a year into the EAA enforcement window opening for retailers, banks, telecoms, and ticket sellers.

What was checked

Eighty popular Slovenian websites, ten per sector across eight sectors: banking, ecommerce, public-sector, telecoms-media, travel, healthcare-insurance, utilities, and sport-culture. Sixteen pass-or-fail checks per site, run against the HTML the page returns to an ordinary browser. The checks split between structural signals on the home page (language attribute, viewport, skip link, main landmark, heading hierarchy, alt-text rate, form-label rate, footer link to an accessibility page, overlay-widget detection, on-page mention of the law) and statement-page signals (statement exists, reads as a real declaration rather than vendor boilerplate, names WCAG or EN 301 549, includes a feedback channel, references the correct Slovenian transposition act and supervisory authority).

No headless-browser axe scan was run in this pass, so colour-contrast violations and unnamed-button counts are out of scope. Those checks need JavaScript to execute against the rendered page and roughly multiply scan time by twenty.

Seven sites refused the connection (HTTP 403) or timed out and are reported as unreachable rather than as failing on accessibility. That itself is a signal: the population of consumer-facing sites that block an automated fetch with a default user-agent is non-trivial, and includes a small number of services that citizens cannot really avoid using.

Findings, sector by sector

Mean checks passed (out of 16, reachable sites only):

Public-sector leads, ecommerce trails. The shape matches the pattern visible elsewhere in the EU: sectors with the longest-standing legal obligation (public-sector under the Web Accessibility Directive since 2018) do best. The EAA-newly-in-scope sectors cluster lower. Ecommerce sits at the bottom, which is uncomfortable, because ecommerce is the slice of the directive most consumers will encounter first.

The statement gap

Forty-five percent of the eighty sites publish no accessibility statement at all. Twenty-seven sites publish what classifies as a real declaration. Ten more publish something thin or partial. The funnel after that gets narrower:

• 15 of 80 statements name a recognised standard (WCAG version or EN 301 549)
• 30 of 80 publish a usable feedback channel
• 27 of 80 reference the correct Slovenian transposition act or supervisory authority

The statute-citation pattern is interesting on its own. Among the sites that name a Slovenian transposition act, 17 cite ZDSMA (public-sector law) and 6 cite ZDPSI (EAA private-services law). When sites do cite a statute, most pick the one that applies to them. The dominant failure isn’t citing the wrong law. It’s citing no law at all.

Overlay widgets

Seven sites have an accessibility-overlay widget injected via script. Five of the seven sit in a single sector. The vendor mix is roughly half from one supplier and the remainder split between two others. None of the overlay-running sites in this sample produced a statement that passes the “real declaration” classifier; some don’t produce a statement at all. The pattern fits what the disability community has been documenting against overlays for years: a marketing layer in place of the underlying accessibility work.

What the home page tells the visitor

Two sites out of eighty mention any legal framework on the home page (ZDSMA, ZDPSI, EAA, WCAG, EN 301 549). For 97% of the sample, a visitor curious about accessibility status would have to find a footer link first, and on 45% of sites that footer link doesn’t exist either.

Where this goes

Slovenia isn’t an outlier in either direction. The headline ratios sit close to the equivalent measurements coming out of other EU member states in the run-up to and after EAA enforcement: somewhere between a quarter and a third of major sites have a real Article 13 declaration, the rest are still catching up. The structural half of the picture is the interesting one. Skip links exist on 22 of 80 sites, which is the simplest WCAG win in the entire standard. The cheapest fix is also the most missed.

The next step on this dataset is a heavier-weight scan that includes axe-core against the rendered DOM, which will surface colour-contrast and accessible-name violations the HTTP check cannot. That work is more expensive and we’ll sequence it after this baseline.

Methodology footnote

All identifiable site names have been suppressed from this report. Aggregate counts, sector means, and statement-funnel data are published; per-site results sit in a separate dataset not released here.