If you run an online store, you have probably seen the pitch. Paste one line of code into your theme, a small icon appears in the corner of every page, and your site becomes accessible, often "compliant" within a couple of days. Some of these tools are free, most cost a monthly fee, and several are a single click away in the Shopify and WordPress app stores.

These tools are called accessibility overlays. For a small store with no developer on hand, the offer is tempting. It is also one of the most widely criticised products in the accessibility field, and understanding why will save you money and help your customers.

What an overlay actually does

An overlay is a script that runs on top of your site after it loads. Most do two things.

The first is a toolbar: the icon opens a panel with buttons to enlarge text, raise contrast, pause animations, change the font, or switch on a built-in "screen reader mode" that reads the page aloud.

The second is automatic repair. The script scans the page as it loads and tries to patch what it finds: it guesses descriptions for images that have none, adds labels to unlabelled buttons and form fields, and changes the code that assistive technology reads, all without touching your theme.

Both sound helpful. Both run into the same problem: they work around the site instead of fixing it.

The toolbar duplicates what people already have

The people an overlay is meant to help already have their own tools, set up the way they like them. Someone who is blind uses a screen reader they have used for years, with its own voice, speed and shortcuts. Someone with low vision has browser zoom, the magnifier in their operating system, or high-contrast settings that apply to every site they visit, not only yours.

A text-size button on one store adds little for them. A second screen reader inside the page can get in the way of the one they are already using. In a 2023 joint statement, the European Disability Forum and the International Association of Accessibility Professionals pointed out that overlays largely repeat features built into browsers and operating systems, and can override the settings users have chosen for themselves.

Automatic repairs are guesses

Software can tell that an image has no description. It cannot know what the image is for. On a product page, the useful description of a photo is "linen shirt in sage green, relaxed fit, shown from the back", and only the person selling the shirt knows which of those details matter. An automatic caption that says "a person wearing a shirt" gives a screen reader user almost nothing to decide on.

The same goes for labels. A script can see that a button has no name, but it has to guess whether that magnifying-glass icon opens search or zooms the product photo. A wrong guess is worse than no label, because it sounds confident.

And many of the problems that stop people buying are not things a script can patch at all: a size picker that cannot be reached with the keyboard, a pop-up that traps focus so the checkout behind it is out of reach, an error message that only appears in red. Those live in how the theme is built, and they have to be fixed there.

Disabled users say they often make things worse

The clearest evidence comes from people who use assistive technology every day. In a study presented at ASSETS 2024, a major accessibility research conference, researchers surveyed and interviewed blind and low-vision web users about their experiences with overlays. As many as 42% of the people surveyed had stopped using a site because of its overlay, most often because the site had become less accessible than it was before. Participants described layout and navigation breaking, overlays clashing with their own screen readers, and pages becoming slow or interrupted by prompts.

Accessibility professionals reach the same conclusion. In WebAIM's 2021 survey of accessibility practitioners, 67% rated overlays as not at all or not very effective. Among respondents with disabilities the figure was 72%.

There is a privacy question too. Some overlays detect whether a visitor is using assistive technology. The European Disability Forum and IAAP warned that this information can reveal that a person is disabled, which is not something a store needs to know or should be collecting.

Regulators have stepped in

In April 2025 the US Federal Trade Commission approved a final order requiring accessiBe, one of the largest overlay vendors, to pay $1 million. The company had claimed its widget could make any website compliant with the Web Content Accessibility Guidelines. The FTC said those claims were false, misleading or unsupported. It also found that the company had presented articles and reviews as independent opinions without disclosing its connections to the people who wrote them. The order bars accessiBe from claiming its automated products can make any website compliant unless it has evidence to back that up.

In Europe, the joint statement is equally direct: overlays do not make a website accessible or compliant with European accessibility legislation. That matters for any store selling into the EU now that the European Accessibility Act applies to many online shops.

What works instead

The alternative is less exciting than a single line of code, but it is the only approach that holds up: fix the problems in the site itself.

  1. Find out what is actually wrong. An automated scan across your whole store shows the problems a machine can detect, page by page, and the patterns that repeat.
  2. Fix them in the template. Most store problems come from a handful of theme files: the header, the product card, the cart. One fix there clears the same problem on every page.
  3. Write your own image descriptions. You know what matters about each product. A sentence per photo, written once, is better than any guess.
  4. Try your checkout with only a keyboard. Tab through it from product page to payment. Anywhere you get stuck, a keyboard or switch user gets stuck too.
  5. Leave people's own tools alone. If you already have an overlay installed, test the site with it switched off. If the site works without it, the widget is not adding anything. If it doesn't, the widget was hiding problems rather than solving them.

None of this needs a large budget. It needs someone to look at the site the way your customers use it, and to fix what they find at the source. That is slower than pasting in a script, and it is the version that works for the people it is meant for.

Sources