what particularity is required when an ad campaign has zillions of possibly algorithmic variants?

Ledesma v. Hismile, Inc., --- F.Supp.3d ----, 2025 WL 3785960, No. 24-cv-03626-KAW (N.D. Cal. Sept. 23, 2025)

Blogging because it’s one of the first cases I’ve seen that has to address questions raised by algorithmically modified ads that are different for different users. Hismile allegedly engaged in fraudulent marketing of its teeth whitening products, “which promise to deliver instant and dramatic results.” Plaintiffs brought the usual California claims. The judge grants the motion to dismiss, with leave to amend (except as to a nationwide class for breach of warranty/unjust enrichment, which is out for good).

Hismile allegedly advertises its products through social media, particularly on TikTok, Instagram, and Facebook, with falsified before-and-after images, misleading celebrity endorsements, deceptive/undisclosed influencer marketing, and “customer reviews” by its own employees.

For example, ads for one product allegedly show the product’s purple serum while it is still on the models’ teeth, giving an illusion that the purple serum cancels out the yellow tones (consistent with their advertising focusing on “color correction” and the “color wheel”), but “fully rinsing off the product causes the color-correcting effect to disappear entirely.” They also allegedly used unnaturally bright lighting and models who already have very white teeth to exaggerate the before-and-after effect of another product. Celebrities paid to endorse allegedly already have very white teeth and are not bona fide users. They also allegedly made false claims of clinical proof, in contradiction to the science indicating minimal effectiveness.

Hismile’s primary argument was that plaintiffs failed to identify the specific ads they saw sufficient to satisfy Rule 9(b). (I’m not convinced that Rule 9(b) should apply to false advertising statutory claims, which were designed to change all the key elements of common-law fraud, but most courts routinely apply it.) The court agreed, but indicated its willingness to accept a somewhat more detailed pleading.

Plaintiffs argued that it was enough to describe their experiences and provide example ads. E.g., plaintiff Tanaka “relied on before-and-after images and videos on Defendants’ Instagram and TikTok, customer reviews, and customer reactions on Defendants’ website and on social media.”  It’s true that “courts have found that pleadings are insufficient where the complaint included a number of representative advertisements, but it was unclear which specific advertisement the plaintiff had seen and relied upon in making their purchase. Likewise, courts have often found it insufficient to simply point to a particular misleading and fraudulent statement or phrase that appeared in various advertisements. Courts have also found it insufficient to merely provide representative advertisements without stating that those were the same advertisements that the plaintiffs saw and relied upon.”

However, plaintiffs argued that they alleged exposure to a long-term advertising campaign, allowing their claims to proceed under In re Tobacco II Cases, 46 Cal. 4th 298 (2009), which stated that when “a plaintiff alleges exposure to a long-term advertising campaign, the plaintiff is not required to plead with an unrealistic degree of specificity that the plaintiff relied on particular advertisements or statements.” The circumstances of the advertising campaign may make it “impossible” to identify the specific advertisement that persuaded an individual to purchase a product.

The court reasoned that “this appears to be a case where Plaintiffs could potentially allege a pervasive, targeted advertising campaign over a period of time, all of which pushes the same message: that Defendants’ products will ‘instantly and dramatically whiten’ teeth.” They alleged a multi-million dollar advertising campaign on social media; they also alleged posts of fifteen or more advertisements per day. One plaintiff alleged seeing approximately sixty advertisements before deciding to purchase the products; requiring a plaintiff to “specifically identify each and every one of these sixty advertisements hardly seems practical or practicable.”

This is especially true for modern social media advertising. At the hearing (not in the pleadings, which is key), plaintiffs noted that the ads include a “seemingly infinite variations of what the ads can look like,” with multiple ads including the same video but in different orders. “One video may include the yellow rubber duck clip followed by a scientist clip, while another video may have the scientist clip come first followed by the yellow rubber duck clip or a yellow banana clip. In short, the same yellow rubber duck clip may be in hundreds of different of ads, making it difficult to identify which advertisement an individual may have seen.” The realities of social media exposure to “numerous 30-second or shorter advertisements, each of which may have focused on demonstrating that whitening worked through color-correction technology,” had to be taken into account.

Bottom line: “To find that Rule 9(b) requires a plaintiff to meet such a high standard would be the same as insulating a defendant from liability simply because they have created so many different types of advertisements that are then repeatedly pushed onto social media users. This would not be a fair result.”

However, the complaint wasn’t enough as currently pled. “Plaintiff must still plausibly allege that this is the type of advertising campaign that would not require them to identify the specific advertisements they viewed,” with allegations about its duration or their exposure; allegations about the strategy of using the same clip in multiple advertisements; and/or allegations that defendants’ social media accounts include thousands of false advertisements.  

Also, with respect to some categories of claims— “before or after videos, videos with scientists and dentists explaining color theory, and videos demonstrating color theory by wiping off purple paint from yellow objects”—there was more specific information, but some plaintiffs alleged that they relied on influencer endorsements without identifying who the influencer was and what was stated:

Significantly, Plaintiffs do not appear to allege that all influencer endorsements are false, such that every influencer endorsement would constitute false advertising. Likewise, some Plaintiffs relied on customer reviews, but do not specify who made these reviews or what they stated. Again, Plaintiffs do not allege that all positive reviews are fake, nor do they suggest that reviews from real customers would be actionable. To the extent Plaintiffs intend to rely on influencer endorsements or reviews, Plaintiffs will need to provide sufficient allegations to demonstrate that the endorsements or reviews they relied upon were false.

The court also commented, looking forward to an amended complaint, that claims of “instant” and “dramatic” whitening might well be non-actionable puffery; “[s]tatements that characterize the speed of an action with terms like ‘fast’ are frequently held to be puffery.”