How were the 51 hotels selected?

They were four- and five-star independent and small-group properties in Austria and southern Germany that take direct online bookings on their own domain. They were picked as a commercial prospect list, not drawn at random, so the sample skews toward well-funded, professionally built websites. That biases the result optimistically: a random sample of smaller pensions would almost certainly score worse, not better.

Does a score of 70 mean the hotel is 70% compliant with the EAA?

No. The score is an automated severity-weighted measure of the issues a scanner can detect. Automated testing reliably covers roughly a third to a half of WCAG 2.1 AA success criteria — things like contrast ratios, missing labels, and missing alt text. It cannot judge whether alt text is meaningful, whether a booking flow makes sense to a screen reader user, or whether an error message is understandable. A high automated score is necessary but not sufficient for compliance.

What was the single most common problem?

Uninformative link text — links reading "more", "here", or "read on" with no context — appeared on 48 of the 51 sites. By raw volume, low colour contrast was the biggest: 8,729 individual failing elements, 23% of everything found in the corpus.

Were booking pages worse than the rest of the site?

Barely. Booking and enquiry pages averaged 1.46 critical-severity issues per page versus 1.21 on ordinary content pages. The difference is real but small. The more useful finding is that the problems are systemic — they come from the template, the design system, and the CMS, so they appear everywhere on the site rather than concentrating in the checkout.

Which hotels were scanned?

We are not naming them. The point of the study is the pattern across the industry, not to pillory 51 individual businesses that have the same problems as everyone else. Every property in the sample was independently contacted with its own results.

Back to Blog

Best Practices

We Scanned 51 Austrian Hotel Websites. Not One Would Pass.

A June 2026 benchmark of 51 DACH hotel websites and 830 pages. Median accessibility score: 70. Sites with zero critical issues: two. Here is what actually breaks, and what it costs to fix.

20 July 2026

11 min read

Ravid, Webply

Also in:

TL;DR

Between 7 and 21 June 2026 we ran full accessibility scans on 51 hotel websites in Austria and southern Germany — 830 pages in total. Median automated score: 70 out of 100. Hotels scoring 90 or above: zero. Hotels with at least one critical-severity WCAG failure: 49 of 51. Four checks — contrast, landmarks, keyboard focus, and unnamed links — account for 73% of every problem found. None of this is exotic. Most of it is a template fix, not a rebuild.

Every accessibility vendor quotes the same statistic: the WebAIM Million, which finds that about 95% of the web's top million homepages have detectable WCAG failures. It is a good number. It is also a number about homepages, in English, mostly American, across every industry at once.

So we ran our own. Not on a million sites, and not on homepages — on 51 real hotel websites in the DACH region, scanning every significant page type on each one: room listings, offers, restaurant pages, spa pages, contact forms, and the direct booking flows.

Here is the whole dataset, and what it says.

What was measured


Sites	51 hotel websites (Austria, plus a handful in southern Germany)
Pages	833 attempted, 830 scanned successfully (99.6%)
Period	7–21 June 2026
Method	Headless Chrome, full page render, axe-core 4.10 plus deterministic keyboard, focus, and text-quality checks
Standard	WCAG 2.1 AA, mapped to EN 301 549
Property type	4–5 star independent and small-group hotels taking direct online bookings

Two things to be honest about before the numbers.

This is not a random sample. These 51 properties were assembled as a commercial prospect list — well-known, well-funded houses with professionally built websites. If anything, that biases the results optimistically. A random draw of Austrian pensions and guesthouses would not look better than this.

Automated testing has a ceiling. Scanners reliably cover somewhere between a third and a half of WCAG 2.1 AA. They catch a contrast ratio of 3.1:1 and a form field with no label. They cannot tell you whether alt="image" is a useful description, or whether a blind guest can actually complete a booking. Every number below is a floor, not a total.

Nobody passed

The score above is severity-weighted from a 100-point baseline: critical rule failures cost 8 points each, serious 4, moderate 1.5, minor 0.5, and any critical failure caps the site at 70.

The distribution is the story. It is not a bell curve with a long tail of bad sites — it is a cluster in the 60s and 70s with nothing at all above 88. The best hotel website in the sample scored 88. The worst scored 47.

That shape tells you the cause. If accessibility were a matter of budget or effort, you would expect a spread: some hotels investing, some not. Instead almost everyone lands in the same band, which is what happens when the problems come from shared infrastructure — the same handful of Austrian hospitality web agencies, the same booking-engine widgets, the same WordPress themes, the same design conventions about thin grey text on white.

49 of the 51 sites had at least one critical-severity WCAG failure. The two that did not still had serious and moderate ones; there was no clean site in the corpus. On average each hotel broke 11.8 distinct WCAG-mapped rules.

What actually breaks

The top of that list is worth sitting with, because it is not what most people expect.

Link text that says nothing — 48 of 51 sites. Links reading "mehr", "hier", "weiterlesen", or "more". A sighted guest reads them in context: the heading above says "Spa", so "more" means "more about the spa". A screen reader user navigating by link list hears "more, more, more, here, more". This is WCAG 2.4.4, and it is the single most widespread failure in the entire corpus.

Text contrast below the AA threshold — 47 of 51 sites, 8,729 individual elements. Overwhelmingly this is light grey body text on white, and white text over photographs. It is a house style in hospitality web design — understated, elegant, and unreadable to anyone over about 55 with normal age-related vision loss. Which is, incidentally, a substantial share of the guests booking five-star Alpine hotels.

No visible keyboard focus — 47 of 51 sites, 6,668 elements. Somewhere in the stylesheet is a line that removes the browser's focus ring because it looked ugly on a button, with nothing put back in its place. A keyboard-only user tabbing through the site cannot see where they are. This is WCAG 2.4.7, it is a one-line CSS fix, and it is nearly universal.

Zoom disabled — 35 of 51 sites. A maximum-scale=1 or user-scalable=no in the viewport meta tag, usually inherited from a mobile theme written years ago. On a phone, a low-vision guest cannot pinch to enlarge your rates. WCAG 1.4.4. This is a critical-severity failure and it is deleting eleven characters from one line of HTML.

Two entries on that chart — content outside landmarks, and skipped heading levels — are flagged as best practice rather than WCAG failures. They make screen reader navigation materially worse and they are worth fixing, but a regulator would not cite them. We have kept them visually distinct so the chart is not doing quiet inflation.

Almost everything is four problems

Across 830 pages, 64 distinct checks failed at least once, producing 37,802 individual failing elements. Four of those checks account for 73% of the total.

This is the most commercially useful finding in the study. The instinct when a scan report lands with 700 issues on it is that the site needs rebuilding. It almost never does. Contrast is a palette decision. Focus visibility is a CSS rule. Landmarks are five wrapper elements in a template. Unnamed links are usually one icon-button component repeated across every page.

Fix four things in the template, and the majority of a hotel's reported issues disappear at once — not because the report was wrong, but because template problems multiply across pages and template fixes divide across them the same way.

The long tail is where the real work is: the remaining 60 checks, 27% of findings, are the ones that need judgment. Unlabelled date pickers. Booking calendars that trap keyboard focus. Payment iframes with no accessible name. Those are per-component fixes, and they are the ones that sit directly in the path of a transaction.

The booking engine is not special — and that is the problem

The obvious hypothesis going in was that booking and enquiry pages would be the disaster zone: third-party widgets, complex date pickers, payment iframes.

Page type	Pages	Critical issues per page	Serious per page
Booking / enquiry / contact	69	1.46	3.19
Homepage	51	1.49	3.59
All other pages	710	1.21	2.85

Booking pages are slightly worse than ordinary content pages, and slightly better than homepages. The gap is real but small — nothing like the cliff we expected.

The correct reading is not "booking engines are fine". It is that the barriers are evenly distributed, because they come from the template rather than from any single component. There is no clean part of a hotel website that a broken booking flow is bolted onto. 28 of the 51 hotels had at least one critical-severity issue on a page in their booking or enquiry path.

That matters legally, because the booking flow is the part the European Accessibility Act most clearly regulates. A hotel taking direct online bookings with online payment is providing an e-commerce service to consumers. The obligation sits with the hotel even when the widget is SiteMinder's or Cloudbeds' — you can require a conformance statement from your vendor, but you cannot transfer the duty.

What this costs to fix

Rough effort, based on what the corpus actually contains:

Fix	Where it lives	Typical effort
Restore visible focus styles	One CSS rule, global	Under an hour
Re-enable pinch zoom	Viewport meta tag	Minutes
Darken body text and overlay text to 4.5:1	Design tokens / theme	Half a day, plus design sign-off
Add landmarks and fix heading order	Page template	Half a day
Give icon links and buttons accessible names	Shared components	Half a day to two days
Rewrite "mehr" / "hier" link text	Content, CMS-wide	Ongoing; largest by volume
Booking flow: labels, keyboard, error announcements	Booking engine or vendor	Days, or a vendor conversation

For most of the sample, the first four rows would move the automated score from the 60s into the 80s inside a week of one developer's time. That is not compliance — the judgment-dependent half of WCAG is untouched by any of it — but it removes the failures that are trivially detectable by anyone, including a complainant, a competitor, or an automated scan run by a consumer organisation.

The last two rows are the ones that decide whether a disabled guest can actually give you money.

What we are not claiming

Some discipline about the limits of this dataset:

No regulator scanned these sites. German market surveillance told us in writing that automated pre-screening of the market is not something they do. Nobody is running a crawler looking for you. The realistic trigger is a complaint from a guest, a competitor, or an advocacy organisation.
We have excluded one of our own checks from every conclusion above. Our keyboard-trap detector fired on 20 sites, and hand-verification showed it over-reports. It is not in any figure in this article.
A score is not a legal verdict. No number produced by any scanner, ours included, establishes conformance. It establishes that specific detectable failures exist.
Sample of 51. Large enough for the pattern to be clear, small enough that you should not quote it as the state of Austrian hospitality as a whole.

The one number to take away

Zero out of 51.

Not "most hotels have some issues". Not one property in a sample of well-resourced, professionally designed four- and five-star hotel websites cleared 90 on an automated check that only measures the easy half of the standard.

If you run one of them, the useful question is not whether your site has problems — the base rate says it does. It is which four template fixes remove most of them, and whether a guest using a screen reader can complete a booking.

Check your own site

Run a free scan of your hotel website and see where it lands against the 51 in this study. No credit card, no sales call.

Or try the no-signup instant check on your homepage first.

Sources

Webply deep-scan corpus: 51 DACH hotel websites, 830 pages, 7–21 June 2026, scoring model v3. Underlying data held by Webply; per-property results shared with each property.
axe-core 4.10 rule definitions and WCAG mappings, Deque Systems.
Directive (EU) 2019/882 (European Accessibility Act), Article 3(30) — definition of e-commerce services.
EN 301 549 v3.2.1 / WCAG 2.1 Level AA success criteria.

This article reports the results of automated accessibility testing and is provided for general information. It is not legal advice and does not constitute an accessibility conformance assessment. Automated testing detects a subset of WCAG 2.1 AA success criteria; a full conformance evaluation requires manual and assistive-technology testing. For advice on your obligations under the European Accessibility Act or its national implementations, consult a qualified lawyer.