Own research
Accessibility Report 2026
What 14,984 German business websites reveal about the state of accessibility
A lot gets claimed about accessibility and very little gets measured. So we looked for ourselves: 14,984 German business home pages, tested automatically against nine criteria that can be determined unambiguously by machine. The result is clearer than we expected.
14,984
sites tested
99.3 %
with at least one finding
111
with no findings at all
The headline finding
Out of 14,984 websites tested, 111 had no findings at all. That is 0.7 percent. On every other site the check found at least one barrier, on average 2.3 per page.
This is not an accusation against individual operators. It shows that accessibility has so far rarely been part of how a website gets built. Anyone starting today has no backlog to catch up on, and is ahead of almost everyone else after the first few fixes.
The nine barriers we tested
Share of tested websites where the criterion was missing or set incorrectly.
| Barrier | Share | Sites |
|---|---|---|
| No skip link to the contentAnyone using the keyboard has to tab through the entire menu on every single page before reaching the actual content. | 98 % | 14,678 |
| Main content not marked upWithout a main content region a screen reader cannot jump straight to it. The page stays one continuous stream of text. | 47.3 % | 7,086 |
| No main headingThe main heading answers what a page is even about. Screen readers announce it first and search engines evaluate it. | 29.4 % | 4,407 |
| Images without alt textImages without alt text simply do not exist for blind visitors. It often hits exactly the product photos meant to drive the sale. | 16.7 % | 2,509 |
| Pinch zoom disabled on mobileThe page forbids pinch to zoom. Anyone who needs reading glasses cannot enlarge the text. | 16.6 % | 2,482 |
| Page language not declaredWithout a declared language the screen reader often picks the wrong pronunciation and reads German text as English. | 6.3 % | 944 |
| No mobile viewportThe page does not scale on phones. Text becomes tiny and controls are hard to hit. | 5.8 % | 874 |
| Icon buttons without a nameA magnifier or cart icon with no label is announced as a button with no purpose. | 2.5 % | 368 |
| Form fields without labelsAn input with no label leaves open what belongs in it. In a checkout that costs the sale. | 1.7 % | 262 |
Broken down by platform
The differences between systems are large and follow no simple rule. They depend less on the operator than on the default theme: whatever the base template brings along shows up on thousands of sites.
| Platform | Tested | No main region | Zoom blocked | No H1 | Images without alt |
|---|---|---|---|---|---|
| WordPress | 3,775 | 40 % | 15 % | 26 % | 18 % |
| WooCommerce | 2,566 | 52 % | 36 % | 31 % | 13 % |
| Jimdo | 865 | 88 % | 24 % | 31 % | 5 % |
| TYPO3 | 680 | 48 % | 18 % | 28 % | 13 % |
| Wix | 466 | 4 % | 0 % | 38 % | 0 % |
| Shopware | 322 | 22 % | 8 % | 54 % | 33 % |
| Shopify | 256 | 13 % | 11 % | 15 % | 14 % |
We leave out the skip link here. It was missing on 94 to 100 percent of pages on every platform, so it separates nothing.
Three things that stood out
The skip link is the blind spot
98 percent of all tested pages do not have one. It is also the change with the best ratio of effort to effect: one line of HTML, and keyboard users skip the entire menu on every subpage.
Shops block zoom particularly often
On WooCommerce sites, pinch zoom was disabled in 36 percent of cases, well above average. It hits exactly those customers who need reading glasses and still like to order on their phone.
Site builders are not uniformly worse
Wix pages almost always had a proper main region and never blocked zoom, but more often lacked a main heading. With Jimdo it was the other way round. Declaring a platform accessible or inaccessible across the board gets it wrong in both directions.
How we measured
We tested the home page in each case, fetched as an ordinary visitor, and evaluated the HTML as delivered. The nine criteria were chosen so they can be determined without interpretation: either there is a skip link or there is not.
The sample comes from publicly available listings of German businesses. We approached 22,553 addresses, of which 5,242 were unreachable and further entries incomplete. That left the 14,984 pages evaluated here. These are small and medium businesses, not a cross-section of all German websites.
What this survey explicitly is not: a full test against the European requirements. Contrast, detailed keyboard operation, focus order and comprehensibility cannot be measured this way. So the real number of barriers is higher than what is shown here, not lower.
Collected in August 2026. We deliberately name no individual companies, only aggregate figures.
Please cite freely
The figures may be used freely, for articles, talks or your own case inside a company. We appreciate a source note linking to this page.
Where does your own site stand?
Our quick test checks the same criteria and a few more, right in the browser and in under a minute. You see immediately which of the barriers listed here appear on your home page.