53.9% of Websites With Accessibility Overlays Still Score Below 90 vs 60.7% Without

A small business owner gets a letter warning that her website is not accessible to people with disabilities. Within a day she is offered a fix that takes five minutes: paste one line of code, and a toolbar appears in the corner of every page, offering bigger text, higher contrast and a screen reader mode. She installs it and considers the job done.

That toolbar is an accessibility overlay: a script added on top of a website that adds visitor controls and tries to repair accessibility problems after the page loads. 255,919 homepages run one. The question for any owner is how much it changes what Google's own accessibility test finds on the page.

Overlays come with a lift of about 3 points in automated scores, and most failures remain. 53.9% of websites with an overlay still score below 90, the passing line, against 60.7% without one. Nearly two thirds still fail color contrast, and 94.2% fail at least one automated check.

We used the September 2026 HTTP Archive, a public dataset that tests millions of websites every month, and took every mobile homepage that had been through Google's Lighthouse accessibility test: 14,717,394 homepages. Lighthouse is Google's free website testing tool, built into the Chrome browser. Its accessibility test runs automated checks such as color contrast, image alt text, form labels and link names, and gives each page a score from 0 to 100, with 90 or more counted as a pass. The test runs after the page and its scripts have loaded, so an overlay's changes are included in the score.

We split the homepages into those running any of the accessibility overlay and toolbar products HTTP Archive recognizes, 255,919 homepages, and those running none. For each group we compared the average score, the share scoring below 90, and the share failing the color contrast and image alt text checks. We repeated the comparison within three popularity groups from Google's Chrome UX Report: the 10,000 most visited websites, websites ranked 10,000 to 1 million, and websites outside the top million.

Websites with an overlay average 87.8 out of 100, against 84.7 for websites without one, a gap of 3.1 points. That moves the share failing the 90-point line from 60.7% to 53.9%, so just under half of overlay sites clear it. For an owner who installed a toolbar to solve accessibility, that means better-than-even odds the page still fails the test.

The gap to a perfect score is wider. 94.2% of websites with an overlay fail at least one automated accessibility check. An overlay works on top of the page; the problems the test finds are in the page itself, in its colors, its form code and its images.

62.6% of websites with an overlay fail the color contrast check, which looks for text too faint to read against its background, against 68.7% without one, about six points better. Contrast is set by the colors in a site's design. Most overlay toolbars offer a high-contrast mode that a visitor has to find and switch on, and the page's own colors stay the same for everyone else.

Image alt text, the description a screen reader reads aloud for each image, shows a bigger gap: 19.8% of overlay sites have images with no alt text, against 26.3% without an overlay. Some overlays add alt text to images automatically, which fits that gap. The test checks that a description is there; a description written by the business says what the image means to a customer.

The same pattern shows up in every popularity group. Among websites ranked 10,000 to 1 million, overlay sites average 87.1 against 82.9 without, the widest gap. Among websites outside the top million, where most small business sites sit, overlay sites average 87.8 against 84.8, and 53.8% still score below 90.

Websites outside the top million decide the overall result, because 243,802 of the 255,919 overlay homepages sit there. For those sites, an overlay goes with an average score of 87.8 instead of 84.8 and a contrast failure rate of 62.7% instead of 68.9%, an improvement that leaves most of the failures in place.

If you already run an overlay, test your homepage with Google's Lighthouse test while the widget is active: open the page in Chrome, right-click, choose Inspect and run the accessibility report. Odds are you still have contrast, form label and image description problems to fix in the site itself. Those fixes are usually quick: darker text colors, a label on every form field and a written description for every meaningful image.

If you are shopping for accessibility help, put the budget into fixing the site's own design, code and content first, then decide whether a toolbar adds anything on top. A website accessibility consultant can find and fix the problems in the page itself, which is what Google's test and your visitors both see.

Need help with this? See our web design services.

Embed this research

Cite this study