dotStrategy Co. v. Facebook Inc

Opinion

1 2 3 4 5 6 UNITED STATES DISTRICT COURT 7 NORTHERN DISTRICT OF CALIFORNIA 8

10 DOTSTRATEGY, CO., 11 Plaintiff, No. 20-00170 WHA

12 v.

13 FACEBOOK, INC., ORDER DENYING MOTION TO CERTIFY CLASS 14 Defendant.

15 16 17 INTRODUCTION 18 This Section 17200 putative class action centers on the intersection of fake accounts and 19 advertising on the world’s largest social media services company. Plaintiff alleges defendant 20 represented it would not charge plaintiff when defendant determined that a fake account 21 clicked on plaintiff’s advertisements, but defendant has failed to refund it even after defendant 22 removed an account from its platform because it was fake. Plaintiff now seeks to represent a 23 class of similarly aggrieved advertisers. Because plaintiff fails to establish a presumption that 24 class members were exposed to the allegedly misleading statements, the motion for class 25 certification must be DENIED. 26 STATEMENT 27 Plaintiff dotStrategy, Co., a for-profit corporation headquartered in Conway, Arkansas, 1 31:6–17). See Internet Corp. for Assigned Names and Numbers, Welcome Registry Operators 2 (June 6, 2021, 8:00 PM), https://www.icann.org/resources/pages/registries/registries-en. 3 Plaintiff sold the right to operate a website with the .buzz domain name. For example, some of 4 the businesses or entrepreneurs who have purchased .buzz domain names from plaintiff 5 included those in the beekeeping industry; the “geo” community, city people who used their 6 .buzz websites to blog about “where they are and what they are doing,” as in the colloquialism, 7 “what’s the buzz”; and the cannabis industry, taking advantage of another colloquialism 8 (Doshier Dep. 39:19–40:19). 9 At all material times, defendant Facebook, Inc., the largest social media services 10 company in the world, provided a free social networking service, facebook.com, for users to 11 connect and share photos, videos, and other content online. As of December 2019, Facebook 12 had 2.5 billion monthly active users and an average of 1.66 billion daily active users. In the 13 fourth quarter of 2019, “fake” accounts accounted for approximately five percent of the 2.5 14 billion monthly active users, or 125 million “fake” accounts. 15 Before placing ads on Facebook, advertisers, like all Facebook users, had to agree to 16 Facebook’s terms of service. The terms of service incorporated by reference Facebook’s self- 17 serve ad terms and, since about April 2018, Facebook’s commercial terms, which applied to 18 use of Facebook for advertising. At all relevant times, Facebook’s terms of service have 19 prohibited fake accounts. For example, the terms in effect beginning January 2015, stated 20 (Dkt. No. 125-25 at 3 (emphasis added)): 21 3. Safety

22 We do our best to keep Facebook safe, but we cannot guarantee it. We need your help to keep Facebook safe, which 23 includes the following commitments by you:

24 * * *

25 2. You will not collect users’ content or information, or otherwise access Facebook, using automated means (such as 26 harvesting bots, robots, spiders, or scrapers) without our prior permission. 27 1 In addition, the section of the terms labeled “Registration and Account Security” 2 required users to “provide their real names and information,” and prohibited users from 3 “provid[ing] any false information on Facebook” (ibid.). 4 Then as now, Facebook made money through advertising. It collected large amounts of 5 data about its users and then exploited that data to offer its advertiser-customers the ability to 6 present highly targeted and customized advertising campaigns to its users, and the ability to 7 evaluate the performance of their ad campaigns with detailed data collected by Facebook. 8 Small-and-medium sized businesses who used Facebook’s self-serve ad tools, like plaintiff, 9 comprised the vast majority of advertisers on Facebook. The self-serve advertising tools 10 allowed advertisers to, inter alia, choose the target audience for the ad campaign based on 11 demographics, interests and behaviors. 12 Facebook charged for advertising based on a combination of (1) an auction system and 13 (2) the billing method chosen by the advertiser. Each time an ad could be shown to a user who 14 fell into the target audience of more than one ad, Facebook’s algorithms held an auction to 15 determine which ad would be shown to the user. The ad with the highest total value won the 16 auction; a combination of three factors determined total value (Dkt. No. 105 at ¶ 39): 17 • Bid: The bid placed by the advertiser for that ad (in other words, what the advertiser is willing to pay to achieve their 18 desired outcome). There are multiple ways to manage your [advertiser’s] bid in the ad auction. 19 • Estimated action rates: An estimate of whether a 20 particular person engages with or converts from a particular ad (in other words, the probability that showing an ad to a 21 person leads to the desired outcome of the advertiser).

22 • Ad quality: A measure of the quality of an ad as determined from many sources including feedback from 23 people viewing or hiding the ad and assessments or low- quality attributes in the ad, such as too much text in the 24 ad’s image, withholding information . . . . 25 In addition, advertisers on Facebook chose between three different billing methods: 26 (1) cost-per-click: Facebook charged the advertiser when a user clicked on its ad; (2) cost-per- 27 action: Facebook charged only when a user performed an action specified by the advertiser, 1 Facebook charged based on the number of impressions, an “impression” was when the ad 2 appeared on the user’s screen (Dkt. No. 120-2 at 7; Dkt. No. 125-1 at ¶ 22). Under the cost- 3 per-impression billing regime, Facebook charged based on the number of impressions, e.g., per 4 one thousand impressions, regardless of whether the user clicked anything on the ad (Dkt. No. 5 125-1 at ¶ 22; Dkt. No. 125-36 at 21). 6 Despite the prevalence of fake accounts on its platform, approximately five percent of the 7 2.5 billion monthly active users at the end of 2019, Facebook made statements like these: 8 • Facebook is a community where everyone uses the name they go 9 by in everyday life. This makes it so that you always know who 10 you’re connecting with. 11 • On Facebook, you’ll only pay to reach the right people who’ll love 12 your business. 13 • If we detect or are alerted to suspicious or potentially invalid click 14 activity, a manual review is performed to determine the nature of 15 the activity. You will not be charged for clicks that are determined 16 to be invalid. [Invalid clicks include] clicks generated through 17 prohibited means, such as fake accounts, bots, scrapers, browser 18 add-ons or other methods that don’t follow Facebook terms. 19 From December 2013 through May 2018, plaintiff placed fifty-five advertising 20 campaigns to promote its .buzz domain name registry business for which Facebook billed it 21 approximately $ 8,000.00 total. Plaintiff alleges that Facebook charged plaintiff when fake 22 accounts clicked on plaintiff’s ads. The complaint alleges the above statements misled plaintiff 23 to believe that Facebook would not charge it for clicks on its ads by fake accounts. It asserts a 24 claim for misleading and deceitful business practices under Section 17200 of California’s 25 Unfair Competition Law. 26 A prior order dismissed plaintiff’s first amended complaint because it made “no 27 allegation that plaintiff reasonably relied upon any statement by Facebook that it would ‘not 1 (Dkt. No. 89 at 8–9). The order invited plaintiff to move for leave to amend stating: “Perhaps 2 it will be easy to cure this shortfall by specifically alleging that plaintiff in fact relied on the 3 specific statement that Facebook would not charge ‘for clicks that are determined to be 4 invalid’” (id. at 8). 5 Plaintiff moved for leave to file an amended complaint, stating: “[The complaint] 6 includes further allegations that Plaintiff reasonably relied on Facebook’s statement that it 7 would not charge for clicks determined to be invalid . . . . As noted by the Court, this 8 allegation ‘is the crux of plaintiff’s case’” (Dkt. No. 95 at 7). 9 Facebook opposed, arguing that the following disclaimer, which appeared in its self- 10 serve ad terms, precluded reasonable reliance on the invalid clicks statement (Dkt. No. 97 at 6): 11 Facebook cannot control how clicks are generated on your ads. We have systems that attempt to detect and filter 12 certain click activity, but we are not responsible for click fraud, technological issues, or other potentially invalid 13 click activity that may affect the cost running ads. 14 The order disagreed, finding the disclaimer ambiguous: “stating that Facebook is not 15 ‘responsible for click fraud’ is ambiguous, as it does not clearly disclaim liability for clicks 16 here at issue” (Dkt. No. 104 at 8). The order also rejected Facebook’s argument that plaintiff 17 had failed to allege falsity (id. at 9–10): 18 Facebook trie[d] to drive a wedge in plaintiff’s theory of liability by drawing a distinction between fake 19 accounts and invalid clicks. It argue[d] that plaintiff ha[d] not alleged facts demonstrating that its promise to not 20 charge advertisers ‘for clicks that are determined to be invalid’ [was] anything other than true. Again, the main 21 statement at issue is the following:

22 ‘If we detect or are alerted to suspicious or potentially invalid click activity, a manual review is 23 performed to determine the nature of the activity. You will not be charged for clicks that are determined to be invalid.’ 24 The order rejected most of the statements in the proposed complaint as nonactionable, but 25 granted plaintiff leave to amend based on the invalid clicks, right people, and real names 26 statements (id. at 9, 13). 27 1 Plaintiff now moves to certify the following class: 2 All persons or entities within the United States who, from December 1, 2013, to the present (“Class Period”), paid Facebook 3 for advertising based on impressions delivered to or actions generated through fake Facebook accounts. 4 This order follows full briefing and a hearing held telephonically. 5 ANALYSIS 6 Certification of a class action is governed by Federal Rule of Civil Procedure 23. The 7 plaintiff must show that the proposed class action satisfies each of the four prerequisites of 8 Rule 23(a) and one of the three requirements of Rule 23(b). Rule 23(a) requires the plaintiff to 9 show: 10 (1) The class is so numerous that joinder of all members is impracticable; 11 (2) There are questions of law or fact common to the class; 12 (3) The claims or defenses of the representative parties are typical of the 13 claims or defenses of the class; and 14 (4) The representative parties will fairly and adequately protect the 15 interests of the class. 16 In addition, plaintiff here seeks certification under Rule 23(b)(3), which requires that 17 questions of law or fact common to class members predominate 18 over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and 19 efficiently adjudicating the controversy. 20 The district court is required to do a rigorous analysis to determine if the requirements of 21 Rule 23 are satisfied. Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350–51 (2011). 22 “Frequently that rigorous analysis will entail some overlap with the merits of the plaintiff’s 23 underlying claim. That cannot be helped. The class determination generally involves 24 considerations that are enmeshed in the factual and legal issues comprising the plaintiff’s cause 25 of action.” Id. at 351 (cleaned up). 26 To certify a class under Rule 23(b)(3), “a district court must find by a preponderance of 27 the evidence that the plaintiff has established predominance under Rule 23(b)(3).” Olean 1 Wholesale Grocery Cooperative, Inc. v. Bumble Bee Foods LLC, 993 F.3d 774, 784 (9th Cir. 2 2021). 3 1. DECEPTIVE ADVERTISING CLAIMS UNDER SECTION 17200. 4 Section 17200 prohibits “any unlawful, unfair or fraudulent business act or practice and 5 unfair, deceptive, untrue or misleading advertising and any act prohibited by [California’s false 6 advertising law].” Claims of false or misleading advertising under Section 17200 “are 7 governed by the reasonable consumer test.” Williams v. Gerber Prods. Co., 552 F.3d 934, 938 8 (9th Cir. 2008) (citation omitted). 9 Section 17200 prohibits “not only advertising which is false, but also advertising which, 10 although true, is either actually misleading or which has a capacity, likelihood or tendency to 11 deceive or confuse the public.” Kasky v. Nike, Inc., 27 Cal. 4th 939, 951 (2002) (citation 12 omitted). Thus, to state a claim under Section 17200 “based on false advertising or 13 promotional practices, it is necessary only to show that members of the public are likely to be 14 deceived.” Ibid. (citations omitted). 15 In addition, a private plaintiff must have “standing,” that is, the plaintiff must have 16 “suffered injury in fact and [have] lost money or property as a result of the” deceptive 17 advertising. Cal. Bus. & Prof. Code § 17204. Section 17204’s requirement that the plaintiff 18 have suffered economic injury “as a result of” the deceitful advertising means that the plaintiff 19 must show actual reliance on the allegedly deceptive or misleading statements. In re Tobacco 20 II, 46 Cal. 4th at 326. 21 Reliance is proved by showing that the defendant’s misrepresentation or nondisclosure was an immediate cause of the 22 plaintiff’s injury-producing conduct. A plaintiff may establish that the defendant’s misrepresentation is an immediate cause of the 23 plaintiff’s conduct by showing that in its absence the plaintiff in all reasonable probability would not have engaged in the injury- 24 producing conduct.

25 While a plaintiff must show that the misrepresentation was an immediate cause of the injury-producing conduct, the plaintiff 26 need not demonstrate it was the only cause. It is not necessary that the plaintiff’s reliance upon the truth of the fraudulent 27 misrepresentation be the sole or even the predominant or decisive in influencing his decision. Moreover, a presumption, or at least 1 an inference, of reliance arises wherever there is a showing that a misrepresentation was material. A misrepresentation is judged to 2 be material if a reasonable man would attach importance to its existence or nonexistence in determining his choice of action in the 3 transaction in question, and as such materiality is generally a question of fact unless the fact misrepresented is so obviously 4 unimportant that the jury could not reasonably find that a reasonable man would have been influenced by it. 5 Id. at 326–27. 6 2. PLAINTIFF HAS MOVED THE GOALPOSTS. 7 The order dismissing plaintiff’s first amended complaint referred repeatedly to the 8 invalid clicks statement as the “main statement,” and “the crux of plaintiff’s entire case” (Dkt. 9 No. 89 at 2, 9). In moving for leave to amend, plaintiff acknowledged that the invalid clicks 10 statement was the crux of its case (Dkt. No. 95 at 7). 11 The order granting plaintiff leave to file a second amended complaint did so explicitly on 12 the basis of the three statements quoted above, but repeated that the invalid clicks statement 13 was the “crux” of plaintiff’s case, the “main issue,” and “the main statement” (Dkt. No. 104 at 14 1, 2, 9). The order explicitly rejected reliance on several of the statements, and implicitly 15 rejected reliance on the “myriad of other Facebook statements” quoted in the complaint (Dkt. 16 No. 104 at 3). 17 Yet plaintiff now asserts (Dkt. No. 129 at 3): 18 All statements upheld by the Court as bases for the § 17200 claim 19 are expressions of Facebook’s “authenticity policy,” the “cornerstone” of Facebook’s community, which the [complaint] 20 references numerous times . . . . Facebook’s authenticity policy is central to the claims in the [complaint] which this Court allowed to 21 be filed over Facebook’s opposition. 22 This order rejects plaintiff’s attempt to shift its theory of the merits at the class 23 certification stage. Plaintiff misconstrues the order granting leave to file a second amended 24 complaint. The order did so based on the invalid clicks, real names, and right people 25 statements only; it rejected as nonactionable, explicitly or implicitly, all the other statements in 26 the proposed complaint. Simply because the order did not explicitly reject each of the “myriad 27 of other Facebook statements” quoted in the proposed complaint did not mean the order 1 based on plaintiff’s vague reference to Facebook’s authenticity policy. Again, in moving for 2 leave to file an amended complaint, plaintiff acknowledged that the invalid clicks statement 3 formed the crux of its case (Dkt. No. 95 at 7). The invalid clicks statement is nowhere in 4 Facebook’s terms or authenticity policy, nor in any of the terms that apply specially to 5 Facebook advertisers. 6 At the hearing on the instant motion, plaintiff argued that one of the statements upheld by 7 the previous order (at the motion to dismiss stage) is synonymous with a statement in the terms 8 of service. The prior order determined that plaintiff had made a sufficient showing that the 9 following statement could be false or misleading: “Facebook is a community where everyone 10 uses the name they go by in everyday life. This makes it so that you always know who you’re 11 connecting with” (Dkt. No. 104 at 13). Plaintiff argues that statement is synonymous with the 12 following statement: “Use the same name that you use in everyday life. Provide accurate 13 information about yourself” (Dkt. No. 105 at ¶ 65). 14 Plaintiff argues (1) the two statements are synonymous, and, (2) therefore, because the 15 previous order upheld the former statement, the latter statement is also independently 16 actionable. The argument fails at both steps. First, the statements are not synonymous. The 17 first statement is a categorical representation; it represents unequivocally that “Facebook is a 18 community where everyone uses the name they go by in everyday life. . . .” (emphasis added). 19 A reasonable consumer could understand that representation to mean that there are no fake 20 accounts on Facebook because a fake account would necessarily contravene that categorical 21 statement. Whereas the latter statement is a command by Facebook to its users, it is one term 22 of the agreement between Facebook and its users: “Use the same name that you use in 23 everyday life. Provide accurate information about yourself.” Crucially, however, this latter 24 statement made no guarantee that all accounts on Facebook complied with that term. 25 Facebook made no representation that it guaranteed enforcement of that term (or any of its 26 terms). No reasonable consumer would have been misled by the fact that Facebook required 27 its users to use the names they go by in everyday life to believe that Facebook guaranteed that 1 In contrast, however, the statement upheld by the previous order arguably made just such 2 a representation: “Facebook is a community where everyone uses the name they go by in 3 everyday life. . . .” That statement is not a term of use between Facebook and its users. 4 Instead, it is a categorical, unqualified statement about who uses Facebook and how they use it. 5 Second, even assuming the two statements were synonymous, the prior order did not hold 6 the real names statement was independently actionable. Again, the order granted plaintiff 7 leave to file its amended complaint based on the invalid clicks statement as the crux of its case. 8 Plaintiff acknowledged as much in its motion. A “crux” is “an essential point requiring 9 resolution or resolving an outcome,” or “a main or central feature (as of an argument).” Crux, 10 MERRIAM-WEBSTER, https://www.merriam-webster.com/dictionary/crux (last visited June 21, 11 2021). 12 Therefore, this order affirms the prior order that the only statements at issue are: 13 • If we detect or are alerted to suspicious or potentially invalid click activity, a manual review is performed to 14 determine the nature of the activity. You will not be charged for clicks that are determined to be invalid. 15 • On Facebook, you’ll only pay to reach the right people 16 who’ll love your business. 17 • Facebook is a community where everyone uses the name they go by in everyday life. This makes it so that you 18 always know who you’re connecting with. 19 3. CLASS EXPOSURE TO DECEPTIVE STATEMENTS. 20 In a class action under Section 17200 based on a misrepresentation theory, absent class 21 members need not satisfy the actual reliance requirement of Section 17204 “where class 22 requirements have otherwise been found to exist.” Tobacco II, 46 Cal. 4th at 324. 23 “Nonetheless, a class action cannot proceed for a fraudulent business practice under the 24 UCL when it cannot be established that the defendant engaged in uniform conduct likely to 25 mislead the entire class.” Davis-Miller v. Automobile Club of S. California, 201 Cal.App.4th 26 106, 121 (2011). Section 17200 does not “authorize an award for injunctive relief and/or 27 restitution on behalf of a consumer who was never exposed in any way to an allegedly 1 “One who was not exposed to the alleged misrepresentations and therefore could not possibly 2 have lost money or property as a result of the unfair competition is not entitled to restitution.” 3 Pfizer Inc. v. Superior Court, 182 Cal.App.4th 622, 631 (2010). Thus, “class certification of 4 UCL claims is available only to those class members who were actually exposed to the 5 business practices at issue.” Berger v. Home Depot USA, Inc., 741 F.3d 1061, 1068 (9th Cir. 6 2014), abrogated on other ground by Microsoft Corp. v. Baker, 137 S. Ct. 1702 (2017). 7 A plaintiff seeking to represent a class on a misleading advertising theory can justify a 8 presumption of class-wide exposure to the alleged misrepresentations by showing that the 9 defendant made the statements in an “‘extensive and long-term fraudulent advertising 10 campaign’ [like the one] at issue in Tobacco II, 46 Cal. 4th at 328.” Mazza v. Am. Honda 11 Motor Co., Inc., 666 F.3d 581, 596 (9th Cir. 2012). But “in the absence of the kind of massive 12 advertising campaign at issue in Tobacco II, the relevant class must be defined in such a way 13 as to include only members who were exposed to advertising that is alleged to be misleading.” 14 Ibid. “To establish a reliance presumption, the operative question has become whether the 15 defendant so pervasively disseminated material misrepresentations that all plaintiffs must have 16 been exposed to them.” Walker v. Life Ins. Co., 953 F.3d 624, 631 (9th Cir. 2020). 17 In In re Tobacco II Cases, the California Supreme Court described the extensive and 18 long-term fraudulent advertising campaign by the defendant cigarette manufacturers and 19 distributors: 20 The complaint alleged that defendants had engaged in a public disinformation strategy concerning the health effects of 21 cigarette smoking beginning in the 1960’s with magazine articles that questioned the link between cigarette smoking and lung 22 cancer. It was further alleged that other public statements by the Defendants over the years have repeated the misrepresentations 23 that Defendants were dedicated to the pursuit and dissemination of the scientific truth regarding smoking and health. 24 * * * 25 The underlying claim in the instant case is that defendants 26 have engaged in a long-term campaign of deceptive advertising and misrepresentations to the consumers of its products regarding 27 the health risks of those products. The class, as certified, consists who were also consumers of defendants’ products during a specific 1 period of time. The nature of the claim is the same—the right to be protected against defendants’ alleged deceit—and the remedies 2 remain the same—injunctive relief and restitution. 3 Id. at 308, 324 (cleaned up). 4 In Mazza v. American Honda Motor Co., Inc., the plaintiffs alleged that Honda’s 5 advertisements misrepresented the characteristics of its Collision Mitigation Braking System, 6 promoted by Honda “as a way to make rear-end collisions less common and to minimize the 7 consequences of collision,” and omitted material information on the system’s limitations. 666 8 F.3d 581, 585–86. 9 To advertise the CMBS, Honda prepared marketing materials describing the system’s functionality. 10 In 2006, Honda released a product brochure stating that the 11 CMBS “is designed to help alert the driver of a pending collision or—if it’s unavoidable—to reduce the severity of impact by 12 automatically applying the brakes if an impending collision is detected.” The brochure described the CMBS system’s three-step 13 process of alerting, lightly braking, and strongly braking if a crash is imminent: 14 * * * 15 The brochure showed a picture of an Acura behind a truck 16 with three labels. Stage 1 was farthest from the truck and stated “RECOGNITION OF POSSIBLE COLLISION.” Stage 2 was in 17 the middle and stated “BELTS TIGHTEN AND LIGHT BRAKING.” Stage 3 was nearest the truck and stated “STRONG 18 BRAKING.” The 2007 and 2008 product brochures were similar and were available at dealerships. 19 Honda also released television commercials describing the 20 system’s operation. One ran for a week in November 2005 and another ran from February to September 2006. In the 2005 21 commercial, a voice states, “The driver is warned, and warned again. If necessary, the system even applies the brakes to lessen 22 the potential impact.” A voice in the 2006 commercial states, “The driver is warned so he can react. If necessary, the system would 23 have even applied the brakes to lessen a potential impact.”

24 From March to September 2006, Honda released a “What Might Happen” advertisement in some magazines. This 25 advertisement said that “the system can react. It can give you auditory and visual warnings, a tug on the seat belt, and when 26 necessary, even initiate strong braking.” Honda ceased mass advertising for the CMBS in 2006. 27 the CMBS’s “various alert stages can overlap depending on the 1 rate of closure of your vehicle and the vehicle ahead. . . . The system does have limitations, and will not detect all possible 2 accident causing situations.” These videos were viewable on kiosks at Acura dealerships, and dealers were encouraged to 3 show them to potential customers. The parties have not indicated how many people saw these videos, and Honda 4 discontinued the use of intranet kiosks in March 2008. Honda also operated an “Owner Link” website that contained video 5 clips describing the CMBS. Although this site was developed for car owners, the site was available to any customer via 6 www.ahm-ownerlink.com until 2008 and via www.myacura.com thereafter. Also, Acura Style magazine, a 7 periodical sent to Acura dealerships, subscribing Acura owners, and interested consumers twice each year, reported in a summer 8 2007 article that the CMBS responds “with any or all of three increasingly dramatic imperatives.” 9 Id. at 586–87 (emphasis added). 10 Our court of appeals vacated class certification, holding that the advertising campaign 11 was too limited to justify an inference that the class members were exposed to the allegedly 12 misleading advertisements. Id. at 595. 13 Following Mazza, two district judges in this district have found that misrepresentations 14 made on webpages, in combination with other disseminations, were not massive advertising 15 campaigns justifying a presumption of exposure. 16 In Ehret v. Uber Technologies, Inc., the plaintiff alleged that “Uber made 17 misrepresentations when it informed consumers that it would automatically charge a 20% 18 ‘gratuity’ when taxi rides were arranged through its app when in fact, Uber kept a substantial 19 portion of the purported ‘gratuity’ for itself.” 148 F.Supp.3d 884, 887 (N.D. Cal. 2015) (Judge 20 Edward M. Chen). Uber made the alleged misrepresentations “primarily on its website, blog, 21 and e-mail messages, rather than on the Uber app itself.” Id. at 896. Judge Chen analyzed the 22 issue of class exposure to the misrepresentations as follows: 23 Plaintiff provides evidence that Uber allegedly 24 misrepresented the 20% gratuity on its website and blog posts. But this falls short of the “decades-long” advertising campaign in 25 Tobacco II, or the highly targeted advertising campaign in Makaeff [v. Trump University, LLC, 2014 WL 688164 (S.D. Cal. 2014)], 26 which not only included advertisements and mailings but free introductory previews which were dedicated to up-selling 27 attendees on more expensive programs. In contrast, Uber’s F.R.D. 436 (N.D. Cal. 2014)] which included a television 1 commercial ad that ran for sixteen months, and in Cohen [v. DIRECTV, Inc., 178 Cal.App.4th 966] which included print 2 advertising and promotional materials. In each of these cases, as well as the instant case, there is no evidence that it was “highly 3 likely” all members of the proposed class saw the allegedly misleading statements made in the advertisements. This is 4 especially true here, where individuals may have downloaded the Uber app based on word of mouth, or used the uberTAXI service 5 because they were previous Uber users who saw that there was a new option on the Uber app and thus never visited the Uber 6 website or blog posts. . . .

7 . . . With respect to the website, [Judge Chen found] insufficient evidence of exposure. Just because the information 8 was available on the website does not necessarily imply that visitors would likely have seen it, especially when there was a 9 good deal of other information on the website. . . . It was not highlighted or especially set off to ensure that visitors would see 10 it. Instead, the website included a good deal of other information about UberBLACK and UberSUV, including pricing information, 11 flat rates, and sample fares for UberBLACK and UberSUV only— none of which involved uberTaxi and the 20% gratuity. 12 Furthermore, because individuals had various reasons for visiting Uber’s website, wholly separate from obtaining 13 information about uberTaxi, this also decreases the likelihood that visitors would have seen the 20% gratuity representation which 14 pertained to uberTAXI only. . . . Similarly, the blog post found on secondary pages of the website was only one of many blog posts, 15 covering a whole range of topics . . . . Significantly, Plaintiff presents no evidence of the likelihood that someone visiting the 16 website or blog would stay long enough to read the information or posts related to uberTAXI and the 20% gratuity representation . 17 . . . 18 Id. at 900–01 (emphasis added) (record citations omitted). Judge Chen did, however, certify a 19 class of “individuals who received e-mails advertising uberTAXI which included the alleged 20 misrepresentation that the 20% charge was for gratuity only. Unlike the website, the e-mail 21 specifically and heavily promoted the uberTAXI service; its focus only on uberTAXI was not 22 diluted by [other] information . . . .” Id. at 901. 23 In Todd v. Tempur-Sealy International, Inc., 2016 WL 5746364 (N.D. Cal. 2016) (Judge 24 Jon S. Tigar), the plaintiffs alleged that the defendants made misrepresentations in their 25 “marketing and sale of mattresses, pillows, and other bedding products containing Tempur 26 material.” Id. at *1. Judge Tigar described the exposure to the misrepresentations as follows: 27 As support for their claim that Defendants used a “tightly content’ (direct mail) each year, and generated more than 25 1 million page views per year on its website.” . . .

2 * * *

3 The same problems apply to Plaintiffs’ contention that “300 million pages” of “targeted content (direct mail)” was produced by 4 Defendants. Plaintiffs do not define “targeted content,” describe what kinds of statements were contained in the mailings or who 5 received them, or offer any reason to assume that consumer actually viewed any of the content or were likely to have done so. 6 Likewise, while Plaintiffs cite to “25 million page views per year” on Defendants’ website, they decline to specify how 7 many of those views were of webpages that included alleged misrepresentations. . . . 8 Id. at *11 (emphasis added). Judge Tigar found that the plaintiffs proved “neither that 9 Defendants’ marketing was sufficiently extensive to reach most of the class, nor that it was 10 sufficiently uniform in its use of the alleged misrepresentations” to justify an inference of 11 class-wide reliance. Id. at *12. 12 Here, plaintiff has adduced no evidence of class-wide exposure to the alleged 13 misrepresentations other than the mere existence of the misrepresentations on a few of 14 Facebook’s webpages, none of which an advertiser would have been required to view before, 15 during, or after advertising on Facebook. 16 The invalid clicks statement could be found in the business help center section of the 17 Facebook for business section of the website (Dkt. No. 105 at ¶ 41, n. 26; Dkt. No. 106 at ¶ 18 41). The right people statement could be found on a page titled “grow your business online” 19 also in the Facebook for business section of the website (Dkt. No. 105 ¶ 46, n. 30). The real 20 names statement could be found on a page titled “What names are allowed on Facebook?” in 21 the Facebook help center section of defendant’s website (Dkt. No. 105 at ¶ 42, n. 27; Dkt. No. 22 125-10 at ¶ 18). 23 Self-serve advertisers had to interact with several different webpages and interfaces to 24 place and manage their ads on Facebook: the terms of service, the self-serve ad terms, the 25 commercial terms, Facebook’s advertising policies, and the ads manager and “boosted post” 26 tools (Dkt. No. 125-36 at ¶¶ 15–17). “None of these self-serve interfaces automatically 27 display[ed] any text from articles in Facebook’s Help Center, Business Help Center, or any 1 online sales page. Users who wish[ed] to view such pages must [have] search[ed] for them or 2 navigat[ed]| to them in some other way” (id. at | 18). 3 Facebook affirmatively presented none of the webpages containing the operative 4 statements to the advertisers (Dkt. No. Dkt. No. 125-36 at 4] 18). Instead, advertisers had to 5 search for the web pages containing the statements among “thousands” of other pages in the 6 relevant sections of defendant’s website (Dkt. No. 125-10 at J] 6, 12, 16). 7 Mazza controls here. In Mazza, Honda had maintained video clips with the 8 misrepresentations on its website throughout the class period, in addition to significant other 9 advertising. Mazza, 666 F.3d at 586-87. Our court of appeals found that to be insufficient. 10 Here, the statements appeared only on several of thousands of webpages on Facebook’s 11 website containing other, none of which an advertiser would have necessarily viewed at any 12 point in buying and placing ads on Facebook. 5 13 Plaintiff does not meaningfully dispute the above but, perhaps sensing doom, tries to shift 14 our focus from the invalid clicks statement to Facebook’s authenticity policy because while 3 15 class members likely did not see the former, they certainly saw the latter. For the reasons 16 stated above, however, the authenticity policy is not the representation at issue. CONCLUSION 18 The remainder of Facebook’s arguments against certification need not be considered. 19 For the foregoing reasons, class certification must be DENIED. 20 21 22 23 IT IS SO ORDERED. 24 25 Dated: June 22, 2021 LAs PC ee 27 WILLIAM ALSUP 28 UNITED STATES DISTRICT JUDGE