Skip to content

Settling Doesn't End It

In federal court, nearly half of 2025’s digital accessibility cases were brought against a defendant who had been sued before.

Your client forwards you a demand letter. Their website allegedly fails WCAG, the plaintiff is someone they’ve never heard of, represented by a firm they’ve never heard of, and the letter proposes a resolution. Your client wants one thing from you: make it go away, quickly, and let them get back to work.

That instinct is reasonable. On the available data, it is also the thing most likely to put them right back here.

What the filings show

UsableNet reviews digital accessibility complaints each year — the filings themselves, in federal and key state courts — and publishes what it finds. Its 2025 review reports that in federal court, 46% of cases involved repeat defendants: companies that had already faced an ADA web accessibility claim.

Before we lean on that number, two things you should know about it, because you’d find them yourself.

UsableNet sells accessibility remediation. This figure runs toward their commercial interest. We’re citing it anyway, because it’s the only count of its kind we’ve found — but you should read it knowing who published it.

And their report is inconsistent about it. The body text says 46% of federal cases. The chart accompanying it says 45%. More importantly, the underlying raw count — 1,427 repeat-defendant filings — is presented against two different universes in the same report: the body frames it against “more than 5,000” federal and state filings, while the chart calls it “45% of all federal cases.” Those can’t both be right, and the difference isn’t rounding — it’s a question of which pile the number came out of. So we cite the range, 45–46%, we attach it only to federal cases, and we rest nothing on the raw count — we’ve named it here only to show you why we won’t use it. If a number’s own publisher can’t say what it’s a fraction of, neither can we.

What survives that scrutiny is still the finding: something close to half of federal digital accessibility cases in 2025 were brought against a defendant who had been through this before. (“Digital accessibility” is UsableNet’s scope — websites, mobile apps, and online services, not websites alone.)

The mechanism UsableNet describes, from reading the complaints, fits in one sentence: “a settlement, limited remediation, a new plaintiff, and another filing—often within months.” And: “Plaintiff firms track litigation history and revisit companies that appear to have addressed only surface-level issues.”

That second sentence is the one to sit with. Prior litigation is not a spent event. It’s a signal — and it’s being read as one.

Why resolving the claim doesn’t resolve the exposure

There’s a structural reason the pattern repeats, and your client will follow it faster than most.

Under Title III, a private plaintiff gets injunctive relief only. DOJ put it plainly in a filing this February: “Title III of the ADA does not provide for money damages for private plaintiffs.” So a negotiated resolution is exactly that — the end of one plaintiff’s claim. It is not a judicial determination that the website conforms to anything, and by itself it does not change a line of code.

Which means the site that generated the first complaint still has the defects that generated it. Those defects remain available to the next person who encounters them.

Money changed hands. Nothing else necessarily did.

There is a federal case where you can watch this happen

You don’t have to take a vendor’s word for the pattern. In February, the United States filed a Statement of Interest in Alcazar v. Fashion Nova (N.D. Cal., No. 4:20-cv-01434-JST, ECF 214, filed Feb. 2, 2026), urging the court to reject a proposed class settlement.

Read it for the litigation anatomy alone. DOJ’s account of plaintiff’s counsel: “Plaintiff’s counsel has brought in excess of 500 such suits, with the vast majority ending in a non-disclosed individual settlement.” And of the named plaintiff: after this case, he “filed 15 more class action lawsuits over three months, each alleging four identical accessibility barriers for blind or visually-impaired people—word for word,” then five more the following year.

That’s the model, described by the government, on the docket. It is not personal, it is not random, and it is not going to stop because one defendant paid one plaintiff.

DOJ’s objection to the settlement itself is the part worth quoting to a client:

“Based on concern that the proposed injunctive relief for class members is not meaningful—it is a mere recitation of the obligation to make visually delivered materials available to individuals who are blind or low vision with no confirmation or enforcement mechanism—the United States objects to the proposed settlement agreement.”

The whole of the injunctive relief was a single sentence promising the site would be modified “as needed to achieve substantial conformance with WCAG 2.1.” No monitoring. No enforcement mechanism. Class counsel’s right to audit the site was optional, at their own expense — and DOJ noted it “appears to have expired on December 14, 2025.”

We’ll be precise about what DOJ did and didn’t say, since the temptation is to overstate it. DOJ did not say the settlement was worthless: it acknowledged that “the common settlement fund would secure monetary relief for the California Class.” Its objection was narrower and sharper — that the injunctive relief secured no guarantee of accessibility, particularly for class members outside California, who were getting only that relief and nothing else.

In other words: the money was real. The fix wasn’t.

The detail that should end the conversation

DOJ hired an expert to test the website the settlement administrator built — the site blind class members had to use to claim their money.

It failed. On the same kind of barriers the underlying lawsuit was about.

The first question on the claim form was “Are you legally blind?” The form labels weren’t programmatically associated with the radio buttons, so a blind claimant using a screen reader never heard the question — they were “simply told their options were ‘Yes’ or ‘No,’ and they would have to guess the question.” Screen reader users heard two “Submit” buttons where sighted users saw one; only one worked. Submit it wrong and the error message wasn’t announced at all, so “a claimant could think that their form was accepted when it was not.” There was no keyboard focus indicator on the “Submit Claim” button.

The settlement machinery built to compensate blind people for an inaccessible website was inaccessible to blind people. DOJ’s word for it was “ironically.”

If you need one artifact to explain to a client why a settlement is not the same thing as a fix, it’s this one.

The widget your client already bought is not a defense

Many clients will tell you they handled this. They installed a plug-in that promised compliance in one line of code.

EcomBack — a remediation vendor counting lawsuits against rival widget vendors, so read it accordingly — counted 983 lawsuits in 2025 against websites that had an accessibility widget present, 24.90% of the 3,948 federal and state filings it tracked, up from 722 (22.65%) in 2024. (That’s EcomBack’s own basis, federal plus state. It is not interchangeable with anyone else’s count, and we’re not going to combine it with one.)

The regulator has been more direct than a vendor can be. On April 22, 2025, the FTC approved a final consent order against accessiBe, which had claimed its accessWidget plug-in could make any website WCAG-compliant. The order requires a $1 million payment and bars accessiBe from representing that its automated products can make any website WCAG-compliant, or can ensure continued compliance over time, unless it has evidence to support the claim. The Commission voted 3–0.

Two things that order is not, and we won’t stretch it. It’s a consent order — not an adjudication, no liability finding, no court ruling. And it binds accessiBe, not the market: we have not found any court holding that overlays fail as a matter of law, and we’re not going to imply one exists.

What it is: the federal government, on the record, telling a company to stop selling the thing your client thinks they bought.

A note on WCAG, since we’re being careful

It’s worth knowing that DOJ, in the same filing, declined to bless the standard everyone cites: “The United States does not endorse WCAG as the appropriate or necessary standard for the provision of auxiliary aids and services under Title III of the ADA.” It applied WCAG 2.1 only because the settlement itself had elected it.

There is no Title III technical standard. WCAG 2.1 AA is the standard courts and parties reference — not one a statute adopts. Anyone who tells you their product delivers “ADA compliance” is describing a legal conclusion as if it were a measurement.

What actually ends it

The uncomfortable thing to tell a client who wants this over with is that the only durable exit is the one they were trying to avoid: fix the website.

Not a widget. Not a settlement that leaves the code untouched. Actual remediation against WCAG 2.1 AA, on the pages where the barriers live — the forms, the checkout, the search, the booking flow. That’s what changes what the next tester finds.

If you want to see what a client’s exposure looks like before you advise them: run their URL through our free report. No login, no credit card, no sales call. It returns the barriers we can detect automatically, and a fixed price to remediate them.

One honest limit on that. Automated detection finds a subset of the problem — it can’t evaluate whether alt text is meaningful, whether focus order makes sense, or whether a keyboard user can escape a modal. Note how DOJ’s own expert worked: five tools, including an axe accessibility tool — the same engine family our scanner is built on — plus JAWS, NVDA, and iOS VoiceOver, driven by a human with thirty years in the field. That combination is the real test. A clean automated scan is a floor, not a verdict — it means nothing machine-detectable was found, not that a site is accessible. Any scanner that claims to certify compliance is making the kind of automated-compliance claim the FTC put a million-dollar order against.

We’re not lawyers, and none of this is legal advice. It’s what the filings say. You’ll know better than we do what to do with it — but if the answer involves someone actually fixing the site, that’s the part we do.

Sources

The two government documents above are the spine of this piece. The two vendor figures are labeled as such, and we’ve told you which way each one cuts. Every quotation was checked against the primary document, not a summary of it.

← All Articles