When You Reject Cookies, You Might Be Agreeing to Arbitration–Alba v. Harbor Freight

Starting in 2024, Harbor Freight deployed the following cookie banner:

The court accepts Harbor Freight’s claim that the “2024 Cookie Banner was displayed on each page on the Website until the user interacted with it, even if the user had previously visited the Website and interacted with the prior cookie banner.”

In 2025, Harbor Freight added an arbitration clause to the TCU. The TCU also contains a class action waiver.

The named plaintiff, Alba, says he visited the website at least once a year since 2020, including after the arbitration clause was added to the TCU. Alba allegedly selected the “Reject All Cookies” button.

What Terms Apply

The court accepts Harbor Freight’s evidence of its user interface:

Harbor Freight has satisfied its burden of proof and demonstrated that Alba was shown the 2024 Cookie Banner. It is undisputed that Alba visited the Website as recently as October 2025. Harbor Freight’s unrebutted evidence that the 2024 Cookie Banner was shown to all users following October 31, 2024, is therefore sufficient to demonstrate that Alba must have been shown the revised banner during his recent visits to the Website….

Alba did not submit a declaration or any other evidence denying that the Website displayed the 2024 Cookie Banner when he visited it in 2025

Alba pointed to an article that says that normally, in the EU, a website doesn’t redisplay the cookie banner to consumers who have already made a selection. The court says this article doesn’t raise any questions about the activities taking place in the US.

Formation

Turning to formation, the court characterizes the cookie banner as a “sign-in-wrap,” which takes us to the now-familiar Chabolla considerations.

Transaction Context. Favors Alba, because “simply browsing a website is insufficient to put the user on heightened notice to anticipate terms of service.”

Visual Design. The court says the offer notice was reasonably conspicuous:

First, the font size used for the advisory paragraph, hyperlinked Terms and Conditions, and buttons was the same. Second, the advisory paragraph displayed in black text against a white background, whereas the Terms and Conditions and Privacy Policy were each bolded, underlined, and in red font. Thus, the Terms and Conditions “stood out from the surrounding text to indicate it was clickable.” Third, the hyperlink to the Terms and Conditions was directly to the left of the three buttons and contained within the pop-up banner, meaning that a Website user could not click on any of the buttons without also seeing the Terms and Conditions displayed on their screen. Finally, there was nothing else cluttering the cookie banner, and the advisory paragraph expressly notified users that “[b]y clicking ‘Accept All Cookies,’ ‘Reject All Cookies,’ ‘Cookie Settings,’ or continuing to browse our website, you acknowledge and agree to our Terms and Conditions of Use.”

A few things:

• The cookie banner excerpt was presented in isolation, not in the context of the entire page. I find these kinds of excerpts unhelpful and maybe misleading. Is the offer conspicuous when looking at the whole page?
• Even in the excerpt, the font is small. It’s blurry in the screenshot I present above (taken from the court opinion).
• Although there isn’t a ton of text, the consumer would have to wade through the cookie discussions to see that the decision to proceed has implications far beyond cookies.

The court distinguishes Chabolla because, in that case, the offer was below the acceptance button such that someone might not look there (here, the text is to the left of the acceptance button). The court says it doesn’t matter that the offer disclosure was at the end of the paragraph block because the TCU links were “bolded, underlined, and in red font, meaning they stood out even without reading the entire paragraph.” Alba also complained that the links to the TCU and privacy policy were “mashed together,” but the court says consumers would recognize the space between them and, either way, still see the TCU reference.

Assent. Alba didn’t challenge assent separately from the disclosure visbility. But I’m stuck on this. Harbor Freight claimed that taking ANY of the following actions was assent: “clicking ‘Accept All Cookies,’ ‘Reject All Cookies,’ ‘Cookie Settings,’ or continuing to browse our website.” This seems extremely expansive. For example, forming the TCUs by continuing to browse is a browsewrap, not a sign-in-wrap. And if someone clicks “cookie settings,” presumably to investigate the website further, it seems unfair to presume they want to form the contract yet.

Also, I’m willing to bet that a majority of consumers presented with this interface would be shocked to learn that “reject all cookies” meant “accept all terms,” including an arbitration clause in the TCU. Yes, that is technically what the disclosures say, but it requires a lot of mental work to put all of this together. It also raises potentially interesting logic puzzles, like what if the cookie policies’ rules about declined cookies and the TCU terms are in conflict with each other? (Nothing in this case suggested that, but I’m raising the possibilities).

Implications

For another case upholding TOS formation via cookie banners, see LB v. LinkedIn.

While I think jamming Alba with the TCU terms when he declined cookies was a raw deal, I’ll note the plus side: Harbor Freight wisely used the engagement with the cookie banner for more than resolving the cookie policy. If you’re going to force users to click, go for the gusto and get them to agree to the TOS at the same time like Harbor Freight did.

The court orders Alba’s case to arbitration. However, another plaintiff in the case took all relevant actions before the arbitration clause was added to the TCU. That plaintiff’s claims are stayed until the arbitration resolves.

For more on cookie banners, you might be interested in Kate Klonick, Ban Cookie Banners: A Case Study in Tech Regulation. I don’t agree with all of her arguments, but I heartily endorse the article’s bottom line that the EU and other countries should reconsider cookie banner mandates (and, in my opinion, all other government-mandated interstitials that break the web’s “click-and-go” paradigm).

Case Citation: Alba v. Harbor Freight Tools USA, Inc., 2026 WL 2161637 (N.D. Cal. Juy 27, 2026)