United States District Court for the Southern District of Florida
Mauricio Martinez and others, on ) behalf of themselves and all others ) similarly situated, Plaintiffs, ) ) Civil Action No. 23-21394-Civ-Scola v. ) D2C, LLC ) doing business as ) Univision NOW, Defendant. )
Order Denying Motion for Class Certification
Plaintiffs Mauricio Martinez, Guadalupe Rodriguez, and Francisco Giron, in this putative class-action case, complain that Defendant D2C, LLC, doing business as, Univision NOW (“Univision” or “Univision NOW”), violated the Video Privacy Protection Act (“VPPA”) by disclosing to Meta Platforms, Inc.—formerly known as Facebook (“Meta” or “Facebook”)—information personally linking them to specific videos they had requested or obtained through Univision’s website. (Am. Compl., ECF No. 35.) The Plaintiffs now seek class certification. (Pls.’ Mot. for Class Cert., ECF No. 61.)1 Univision has responded (Def.’s Resp., ECF No. 69) and the Plaintiffs have timely replied (Pls.’ Reply, ECF No. 74).2 Additionally, Univision asks the Court to strike new evidence and argument that it says the Plaintiffs presented in their reply. (Def.’s Mot. to Strike, ECF No. 79.)3 That motion is also fully briefed (Pls.’ Resp., ECF No. 84; Def.’s Reply, ECF No. 88).4 After careful review of the briefing, the record, and the applicable legal authorities, the Court denies the Plaintiffs’ motion for class certification (ECF Nos. 60, 61) and denies Univision’s motion to strike (ECF Nos. 76, 79) as moot.
1 The Court will cite primarily to the sealed filings in this case. A redacted version of each sealed filing is also available on the docket. The corresponding publicly available docket entry for this motion, and its associated exhibits, is located at ECF No. 60. 2 The corresponding publicly available docket entry for Univision’s response, and its associated exhibits, is located at ECF No. 68, and for the Plaintiffs’ reply, and its associated exhibits, is located at ECF No. 72. 3 The corresponding publicly available docket entry for Univision’s motion to strike, and its associated exhibits, is located at ECF No. 76. 4 The corresponding publicly available docket entry for the Plaintiffs’ response, and its associated exhibits, is located at ECF No. 82, and for Univision’s reply, and its associated exhibits, is located at ECF No. 86. 1. Background “Congress enacted the VPPA in 1988 after a newspaper published a profile of Supreme Court nominee and then D.C. Circuit Judge Robert H. Bork which contained the titles of 146 films he and his family had rented from a local video store.” Ellis v. Cartoon Network, Inc., 803 F.3d 1251, 1252 (11th Cir. 2015) (cleaned up). As one of the sponsors of the bill, Representative Al McCandless explained that “people ought to be able to read books and watch films,” “protected from the disruptive intrusion of a roving eye.” S. Rep. 599, 2d Sess., at 7 (1988), reprinted in 1988 U.S.C.C.A.N. 4342-1, 4342–7. In its current form, the VPPA prohibits “[a] video tape service provider” from disclosing a “consumer’s” “personally identifiable information” (sometimes “PII”). 18 U.S.C. § 2710(b). A “consumer” is defined, under the Act, as “any renter, purchaser, or subscriber of goods or services from a video tape service provider.” 18 U.S.C. § 2710(a)(1). A “video tape service provider” is defined, in turn, as a company “engaged in the business . . . of rental, sale, or delivery of prerecorded video cassette tapes or similar audio visual materials.” 18 U.S.C. § 2710(a)(4). And, finally, “‘personally identifiable information’ includes information which identifies a person as having requested or obtained specific video materials or services from a video tape service provider.” 18 U.S.C. § 2710(a)(3). Univision is an online video-streaming service that offers both livestream access to broadcast networks and video-on-demand (prerecorded) content. During the class period—defined by the Plaintiffs as April 13, 2021, through May 8, 2023—the Plaintiffs were all paid Univision subscribers. In broad strokes, the claims of the Plaintiffs—as consumers under the VPPA—boil down to their contention that Univision—as a video tape servicer provider—violated their rights by disclosing their Facebook identification numbers (“Facebook IDs”) along with the prerecorded videos they accessed—personally identifiable information—to Meta. (E.g., Am. Compl. ¶ 3.) In pursuing redress under the Act, the Plaintiffs seek to certify a class they define as follows: All persons in the United States who purchased a subscription to Univision NOW, requested or obtained prerecorded video materials or services on Univision NOW’s website, used Facebook during the time the Pixel was active on Univision NOW’s website from April 13, 2021[,] through May 8, 2023, and whose Personal Viewing Information5 Univision NOW disclosed to Meta.
5 Though not specified in the Plaintiffs’ motion, “Personal Viewing Information” is defined in the amended complaint as each Plaintiff’s “Facebook ID, along with specific video titles and the (Pls.’ Mot. at 2.)6 The parties do not dispute, at least for the purposes of class certification, many of the relevant facts in this case. Univision is a “video tape service provider”; the Plaintiffs, as subscribers to Univision’s video-subscription service, are “consumers”; and a consumer’s Facebook ID, together with the prerecorded videos he or she requested or obtained, are “personally identifiable information.” There is also no dispute, for now, that none of the Plaintiffs consented to Univision’s alleged knowing disclosure of this information. Nor, finally, is there any dispute that Univision’s web developer—Endeavor Streaming—identified 35,845 subscribers as having viewed at least one prerecorded video on Univision’s website. Instead, the parties’ quarrels stem, primarily, from their differing (or evolving) understandings of the functionality of the mechanism by which Univision allegedly transmitted the subscribers’ personally identifiable information to Meta. As the Plaintiffs describe it, Facebook, in 2013, introduced a “Pixel,” sometimes called the “Meta Pixel,” that allowed online businesses to track their customers’ activities on their websites. (Am. Compl. ¶ 20.) This Pixel is an embedded piece of commonly used computer code, installed on a company’s website, that can collect and transmit data regarding a user’s website activities. (Id. ¶¶ 20–23; Pls. Mot. at 3–4.) While the parties do not dispute that Univision deployed the Pixel on its video-streaming website, they disagree as to how it functioned (or at least disagree as to the significance of its functionality). In their complaint and in parts of their briefing, the Plaintiffs maintain the Pixel transmitted subscribers’ personally identifiable information automatically; Univision, in contrast, identifies several conditions that had to be satisfied before the Pixel would fire. Notably, though, even the Plaintiffs’ own description of the operation of the Pixel is not always consistent. On the one hand, in their complaint and, to an extent, in their class-certification briefing, the Plaintiffs maintain that Univision set up the Pixel so that it would automatically disclose a Univision subscriber’s personally identifiable information any time someone logged into the subscriber’s account and clicked on or requested a prerecorded video. (Am. Compl. ¶¶ 2–3, 19–23, 26, 28–29, 33, 37, 43, 49, 57, 73, 77; Pls.’ Mot. at 1 (“When a Univision NOW subscriber clicked on a video, the Pixel simultaneously transmitted to Meta the user’s Facebook ID . . . along with . . . the title of the of the video that
videos’ URLs identifying specific prerecorded videos each Plaintiff requested or obtained.” (Am. Compl. ¶ 2.) 6 Because the ECF filing numbers of the sealed versus unsealed documents do not always correspond, for the sake of consistency, the Court will reference the pagination supplied by the parties in their documents. subscriber requested or obtained.”), 4, 5.) At the same time, however, the Plaintiffs also acknowledge, in their motion, that there were at least two impediments to Univision’s transmitting a subscriber’s personally identifiable information to Meta: (1) where the subscriber did not have a Facebook account; and (2) where the subscriber was using a browser that blocked the Pixel under the browser’s default settings. (Pls. Mot. at 6 n. 3.) Univision, in contrast, describes several other conditions that had to be met in order for a subscriber’s personally identifiable information to be transmitted to Meta—aside from the two conditions acknowledged by the Plaintiffs. (Def.’s Resp. at 3, 11.) As Univision’s expert, James Vint, points out, whether a subscriber’s personal viewing information was actually transmitted depended not only on whether the subscriber had a Facebook account and whether a browser’s default settings allowed the Pixel transmission, but also on (1) whether the subscriber was simultaneously logged into Facebook (Vint Decl. ¶¶ 24–31, ECF No. 69-2); (2) whether the subscriber accessed the prerecorded video on Univision’s website through the same web browser and device through which the subscriber (and not another user) was logged into Facebook (id. ¶¶ 36–37; 47–50; 59–60); and (3) whether some other browser-related component blocked the Pixel (id. ¶¶ 51–58).7 As the Court explains in more detail below, much of the controversy at the center of this case springs from these divergent views on the Pixel’s functionality. The Plaintiffs maintain that at least 17,000 subscribers, including (or in addition to) the three representative Plaintiffs, have had their personally identifiable
7 Although Vint identifies what he says are eleven distinct factors, many of them are redundant to variables already accounted for or are subsumed within one another. For example, Vint lists as separate “variables” (1) whether a subscriber has a Facebook account (Id. ¶ 23); (2) whether that subscriber was logged into their Facebook account while accessing the Univision video (id. ¶¶ 24–28); (3) whether the subscriber elected to remain signed in to Facebook during a prior browsing session (id. ¶¶ 29–31); and (4) whether the subscriber accessed Facebook only through the mobile application (id. ¶¶ 47–50). But these four variables are all accounted for under the conditions that require a subscriber to have accessed the prerecorded video through the same web browser and on the same device, as the subscriber (and not another user) used to log in to Facebook. (Id. ¶¶ 36–37, 44–45, 59–60). Similarly, Vint also identifies as variables whether the subscriber accessed Univision through its mobile application versus through its website and whether the subscriber watched content from Univision’s live channels or DVR programming versus selecting a video on demand. (Id. ¶¶ 32, 33–35.) But these factors too are already accounted for in the starting figure of 35,845 Univision subscribers who viewed a prerecorded video (and not content from Univision’s live channels) on Univision’s website. Finally, Vint lists as other factors both (1) whether the subscriber accessed Univision’s website on a browser that, by default, blocks the Pixel transmission (id. ¶¶ 38–43); or (2) if the subscriber accessed the website through a browser that didn’t block the Pixel by default, whether some affirmative action was taken on the browser to prevent the Pixel from transmitting anyway (such as blocking cookies within the browser’s settings, enabling the browser cache to self-destruct, clearing cookies, or deploying cookie blockers or anti-tracking software) (id. ¶¶ 51–58). But these are all just different ways that a browser might block the Pixel from transmitting rather than distinct factors. information disclosed to Meta by Univision. In stark contrast, Univision maintains that the Plaintiffs have failed to carry their burden of showing that even a single subscriber has had their personally identifiable information disclosed, including the three named Plaintiffs. 2. Legal Standard for Class Certification Federal Rule of Civil Procedure 23 “establishes the legal roadmap courts must follow when determining whether class certification is appropriate.” Valley Drug Co. v. Geneva Pharms., Inc., 350 F.3d 1181, 1187 (11th Cir. 2003). In view of the “awesome power of a district court” in controlling the class action mechanism, any decision to certify a class must rest on a “rigorous analysis” of the requirements of Rule 23. See Sacred Heart Health Sys., Inc. v. Humana Military Healthcare Servs., Inc., 601 F.3d 1159, 1169 (11th Cir. 2010) (cleaned up); see also Gen. Tel. Co. of Southwest v. Falcon, 457 U.S. 147, 161 (1982). While the district court’s class certification analysis “may entail some overlap with the merits of the plaintiff’s underlying claim, Rule 23 grants courts no license to engage in free-ranging merits inquiries at the certification stage.” See Amgen Inc. v. Conn. Ret. Plans & Trust Funds, 133 S. Ct. 1184, 185 L. Ed. 2d 308, 2013 WL 691001, at *7 (U.S. 2013) (cleaned up). Rather, “[m]erits questions may be considered to the extent - but only to the extent - that they are relevant to determining whether the Rule 23 prerequisites for class certification are satisfied.” See id. “The burden of proof to establish the propriety of class certification rests with the advocate of the class.” Valley Drug Co., 350 F.3d at 1187. “Under Rule 23, certification is proper where the proposed classes satisfy an implicit ascertainability requirement, the four requirements listed in Rule 23(a), and the requirements listed in any of Rule 23(b)(1), (2), or (3).” Karhu v. Vital Pharms., Inc., 621 F. App’x 945, 946 (11th Cir. 2015). Per “Rule 23(a), every putative class first must satisfy the prerequisites of numerosity, commonality, typicality, and adequacy of representation.” Vega v. T-Mobile USA, Inc., 564 F.3d 1256, 1265 (11th Cir. 2009) (cleaned up). Thus, Rule 23(a) is satisfied only where: (1) the class is so numerous that joinder of all members is impracticable; (2) there are questions of law or fact common to the class; (3) the claims or defenses of the representative parties are typical of the claims or defenses of the class; and (4) the representative parties will fairly and adequately protect the interests of the class. Fed. R. Civ. P. 23(a). Moreover, where certification is sought under Rule 23(b)(3), as it is here, a plaintiff must show, in addition to the four requirements of Rule 23(a), that “the questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.” Fed. R. Civ. P. 23(b)(3); Vega, 564 F.3d at 1265. 3. Analysis The Plaintiffs maintain they have satisfied all the requirements for class certification. As to the Rule 23(a) factors, the Plaintiffs contend, first, that based on “reasonable inferences,” the Court may determine that Rule 23(a)(1)’s numerosity requirement is satisfied. (Pls. Mot. at 5–6.) Second, they submit that Rule 23(a)(2)’s commonality requirement is met because each class member will use common evidence or address common questions to establish the same four elements under the VPPA: (1) whether Univision is a video-tape service provider; (2) whether the information transmitted by the Pixel constituted the proposed class members’ personally identifiable information; (3) whether Univision knowingly disclosed the personally identifiable information; and (4) whether the proposed class members are consumers. (Pls.’ Mot. at 6–10.) Third, the Plaintiffs posit that Rule 23(a)(3)’s typicality requirement is satisfied because, like all class members, the Plaintiffs were paid Univision subscribers, watched prerecorded videos on the website, “and thus would have had their PII transmitted to Meta.” (Pls.’ Mot. at 10–11.) And fourth, the Plaintiffs say that Rule 23(a)(4)’s adequacy requirement is met because (1) there is no conflict of interest between the Plaintiffs and the class members and (2) their counsel are experienced and committed to vigorously prosecuting this case. (Pls.’ Mot. at 11–12.) As to ascertainability, the Plaintiffs maintain that the class members who meet the first two prongs of the proposed class—Univision subscribers who requested or obtained prerecorded video materials—can be identified using Univision’s own records which logged subscribers’ viewing histories. (Id. at 15.) And, they say, the class members who meet the last two prongs—subscribers who used Facebook while the Pixel was active and had their Personal Viewing Information disclosed to Meta—can be identified through Meta’s own pixel- transmission event data. (Id. at 15–16.) Accordingly, they say, the class is both adequately defined and clearly ascertainable. Finally, as to Rule 23(b)(3), the Plaintiffs assert they have established both predominance and superiority. (Pls.’ Mot. at 12–15.) Regarding predominance, the Plaintiffs maintain that “every element” of the class members’ claims can be proved “through common evidence,” with no undermining individual questions. (Pls.’ Mot. at 12, 13.) And, as to superiority, the Plaintiffs point primarily to efficiencies to be gained and the insufficiency of the $2500 in statutory damages applicable to each violation to justify spending the time, effort, and money needed to litigate thousands of actions on an individual basis. (Id. at 13–15). In opposing certification, Univision contends the Plaintiffs have failed to carry their burden on all fronts: ascertainability, all four Rule 23(a) factors, and both predominance and superiority under Rule 23(b)(3). Although the Court finds the proposed class is ascertainable—considered a prerequisite to the 23(a) analysis—it finds the Plaintiffs have not carried their burden as to numerosity, as set forth below. See Cherry v. Dometic Corp., 986 F.3d 1296, 1302 (11th Cir. 2021) (“Class representatives bear the burden to establish that their proposed class is adequately defined and clearly ascertainable, and they must satisfy this requirement before the district court can consider whether the class satisfies the enumerated prerequisites of Rule 23(a).”) (cleaned up). Because the Plaintiffs’ miss on numerosity is dispositive, the Court declines to evaluate the other Rule 23 factors. A. The proposed class is ascertainable. The Court is not persuaded by Univision’s argument as to ascertainability: its argument is aimed at the inadequacy of the cited records to identify class members rather than whether, with more robust resources, the class would ultimately be “capable of determination” through clear and objective criteria. Cherry, 986 F.3d at 1304. Univision’s focus on the various logistical difficulties is at odds with the Eleventh Circuit’s elimination of administrative feasibility as a part of the ascertainability inquiry of a proposed class: determining a class must just be possible, even if not easy or convenient. Id. at 1303 (“membership can be capable of determination without being capable of convenient determination”) (emphasis in original). Since the Eleventh Circuit’s clarification, courts have distilled certain guiding principles. Among them, first, “the district court must find that the proposed class [is] defined using clear and objective criteria and not defined with vague and subjective criteria.” Fox v. Ritz-Carlton Hotel Co., L.L.C., 345 F.R.D. 358, 364–65 (S.D. Fla. 2024) (Singhal, J.) (citing Cherry, 986 F.3d at 1302–03). And second, the court must also “find that the class is capable of being determined,” without regard to “how the plaintiff proposes to apply the class criteria and actually identify its members.” Id. at 365. Univision does not argue that the proposed class is defined with vague or subjective criteria. Instead, it complains about the insufficiency of the records to identify class members: “Univision NOW does not, and never has, tracked such activity”; “This criterion, too, cannot be determined based upon existing records”; “This is not information that is tracked or kept by Univision NOW, nor is it information that appears to have been kept by Meta.” (Def.’s Resp. at 10.) At heart, these are simply “rebranded administrative feasibility arguments.” Fox, 345 F.R.D. at 364 (rejecting a defendant’s arguments that “it would be difficult to locate individuals who satisfy the . . . requirements in the proposed class[]” as “irrelevant in a post-Cherry world”). In focusing its analysis on how the plaintiffs propose to apply class criteria in order to identify class members, Univision fails to adduce any meaningful argument that the proposed class is not capable of being determined even if that determination would be extraordinarily difficult considering the records the Plaintiffs propose relying on. Without more, each of the criteria the Plaintiffs present is clear and capable of objective determination: all people who (1) subscribed to Univision NOW; (2) requested or obtained a prerecorded video on Univision NOW’s website; (3) used Facebook while the Pixel was active; and (4) had their viewing information disclosed to Meta. There appears to be no dispute from Univision that, at a minimum, all these criteria are “premised on clear, historical facts that don't require any level of subjectivity to ascertain.” Id. at 365. As such, even if Univision is correct—that barriers to identifying the class members may ultimately prove realistically insurmountable—the Court finds the proposed class is nonetheless adequately defined and clearly ascertainable. B. The Plaintiffs have failed to establish numerosity. Nonetheless, the Court agrees with Univision that the Plaintiffs have not carried their burden of showing numerosity. In the numerosity analysis, courts typically consider fewer than twenty-one class members to be inadequate and more than forty to be adequate, with “numbers falling in between . . . open to judgment based on other factors.” Vega, 564 F.3d at 1267 (cleaned up). The Court finds the Plaintiffs have fallen far short of this showing. The bulk of the Plaintiffs’ case for certification rests on their central theory that the mere presence of the Meta Pixel on Univision’s website resulted in the transmission of tens of thousands of Univision subscribers’ personally identifiable information from Univision to Meta.8 But this conclusory theory is
8 E.g., Am. Compl. ¶¶ 23 (“Univision NOW chose certain options . . . that track specific user activity on Univision NOW’s website for automatic disclosure to Meta, including personally identifiable information.”) (emphasis added), 28 (“[W]hen a user clicked on and requested a video . . . Univision NOW disclosed to Meta . . . the specific video name that the digital subscriber requested . . . and the digital subscriber’s [Facebook] ID to Facebook in a single transmission.”), 57 (“Univision NOW disclosed the personal viewing information of all of those subscribers”—who requested or obtained specific video content—“to Facebook in violation of the VPPA”); Pls.’ Mot. at 1 (“When a Univision NOW subscriber clicked on a video, the Pixel simultaneously transmitted to Meta the user’s Facebook ID . . . along with . . . the title of the of the video that subscriber requested or obtained.”), 5 (“[T]he Pixel automatically transmitted the fact that the subscriber requested or obtained that video to Meta”), 8, 11 (“[E]ach Plaintiff, like all Class Members, watch directly undercut not only by Univision’s unrefuted expert testimony, but by the Plaintiffs’ own concessions and evidence. Ultimately, the nonviability of the Plaintiffs’ theory of automaticity dooms their attempt to establish numerosity. In maintaining they have satisfied Rule 23(a)(1)’s numerosity requirement, the Plaintiffs point to a series of “reasonable inferences” they say lead to a showing that the class is so numerous that joinder of all members is impractical. (Pls.’ Mot. at 5–6.) Their starting point is the 35,845 subscribers, at least, in the United States, who Univision itself says viewed, during the relevant period, a minimum of one prerecorded video on Univision’s website.9 (Pl.’s Mot. at 5.) On the one hand, relying on their overarching theory, the Plaintiffs initially suggest that this 35,845 figure directly correlates to the actual number of class members, positing that “the Pixel automatically transmitted the fact that the subscriber requested or obtained that video to Meta” every single time each one of those subscribers viewed a video. (Id.) On the other hand, however, the Plaintiffs themselves also acknowledge, in direct tension with this theory, at least two impediments to a subscriber’s viewing information’s being transmitted to Meta: (1) not having a Facebook account; and (2) using a browser that, by default, blocks the Pixel. (Id. at 6 n. 3.) Based on just these two hurdles, the Plaintiffs acknowledge an estimated 50% overall reduction to their initial 35,845 class size: they point to statistics regarding the percentage of people in the United States who have Facebook accounts (68%) and the testimony of their expert, Dr. Serge Egelman, regarding the percentage of the population who use a web browser that would not block the Pixel transmission (70%), to conclude, using “basic math,” that the class would be comprised of “at least approximately 17,000 individuals.” (Id.) The Plaintiffs’ attempt to establish numerosity, however, is problematic for several reasons. First, as to the Plaintiffs’ concession that a Pixel transmission is contingent on whether a Univision subscriber has a Facebook account, their reliance on a Pew Research poll for their supposition that this would only knock out 32% of their 35,845 starting point is not particularly helpful. (Pls.’ Reply at 6 n. 3.) As the Plaintiffs concede in reply, “being logged in to Facebook”—not just having an account—“is a prerequisite to the Pixel disclosing information.” (Pls.’ Reply at 5 (emphasis added).) Moreover, as Univision’s expert points out (and, as
prerecorded videos on Univision NOW’s website . . . and thus would have had their PII transmitted to Meta . . . exactly as Univision had intended.”); Pls.’ Reply at 8 (“[T]he Pixel would automatically fire when a subscriber viewed a prerecorded video on Univision NOW.”). 9 The Plaintiffs also emphasize that this number doesn’t even account for the users who may have watched more than one video. (Pls.’ Mot. at 6.) While the Court understands this fact may increase the absolute number of VPPA violations associated with a particular class member, the Plaintiffs fail to articulate how an increased number of alleged violations per subscriber would shed any light on the absolute number of members in the proposed class. explained below, the Plaintiffs do not really dispute), even being simultaneously logged in to Facebook is still not enough to necessarily prompt a Pixel transmission: a subscriber must also have accessed the prerecorded video on Univision’s website through the same web browser and device through which the subscriber (and not another user) was logged into Facebook. (Vint Decl. ¶¶ 28 (“[i]f a user visited the [Univision] Website but was not currently logged in to Facebook on the same device and browser, the c_user”—which contains a subscriber’s Facebook ID— “would not be sent to Facebook”) (emphasis in original), 36–37, 47–50; 59–60). Accordingly, the Plaintiffs’ statistic, even taken at face value, is simply a starting point. That is, while 68% of the population in the United States purportedly “has a Facebook account”—or, according to the website the Plaintiffs reference, “ever use Facebook”—that tells the Court nothing about what percentage of the 35,845 Univision subscribers who viewed a prerecorded video were also logged in to their own Facebook accounts, through the same browser, and on the same device on which they viewed that video. The Plaintiffs’ assessment of the effect that web-browser blockers had is similarly problematic. To begin, the Court is unable to reconcile the Plaintiffs’ proffer that “70% of the United States population uses a browser that does not block the Pixel” with the record evidence they cite to support it. (Pls.’ Reply at 6 n. 3 (citing Egelman Rep. ¶ 44).) In the report testimony that the Plaintiffs rely on, the Plaintiffs’ expert opines that Google Chrome and Microsoft Edge, which, combined, make up 70% of the market share, are both likely to allow the Pixel transmission “under default configurations.” (Egelman Rep. ¶ 44.) Missing from this explanation, however, is any information about what percentage of that population actually use their Chrome or Edge browser under this default configuration. Accordingly, while the Court can accept that about 70% of the 35,845 subscribers likely used either Chrome or Edge, the Court is once again left to arbitrary speculation as to what percentage of that 70% used their browser under the default settings. Further, Univision lists several deliberate actions a user can take that would also block any Pixel transmission to Meta: enabling a browser’s third-party cookie blockers; setting a browser’s cache to “self-destruct”; clearing cookies upon the end of a browser session; and deploying add-on software that blocks third-party cookies. (Def.’s Resp. at 11 (citing Vint Decl. ¶¶ 51–58).) The Plaintiffs account for none of these variables. Further, the Plaintiffs do not put up any real challenge to all these additional conditions that Univision identifies as impeding the Pixel’s transmissions. Instead, they simply characterize those factors, in passing, as “manufactured” and “purported,” without actually disputing their significance or effect on the operation of the Pixel. (Pls.’ Reply at 1, 5.) Indeed, the Plaintiffs tacitly acknowledge the existence of these factors, maintaining only that they are simply irrelevant. As the Plaintiffs see it, determining on a case-by-case basis whether all these conditions have been met is unnecessary because the very fact of a disclosure of a subscriber’s personal viewing information would necessarily mean that all of the factors “would already have been accounted for.” (Id. at 1–2.) But in the numerosity analysis, this puts the cart before the horse. In the end, the Plaintiffs fail to supply “the means to make a supported factual finding, that the class [to be] certified meets the numerosity requirement.” Vega, 564 F.3d at 1267. Right out of the gate, as explained above, the Plaintiffs concede that their blanket contention that “[e]ach time a subscriber viewed a video on the Univision NOW website, the Pixel automatically transmitted the fact that the subscriber requested or obtained that video to Meta” is untrue.10 As the Plaintiffs themselves explicitly acknowledge, “not every single one of the at least 35,845 subscribers who watched at least one prerecorded video on . . . Univision’s website will be Class members” because either they don’t have a Facebook account or used a website browser that blocked the transmission. (Pls.’ Mot. at 5–6, n. 3.) Again, as elaborated on above, in order for the Pixel to have transmitted the relevant information to Meta, several conditions must have been met. To summarizer, in addition to viewing or selecting the prerecorded video through Univision’s website, the subscriber must have also: (1) had a Facebook account at the time the video was selected; (2) used a web browser that didn’t block the Pixel by default; (3) been simultaneously logged into the subscriber’s own Facebook account while selecting the video; (4) been simultaneously logged into Facebook on the same device that the subscriber used to select the video;
10 Despite acknowledging in their motion for class certification that there are at least some factors that would prevent the Pixel from transmitting information to Meta when a subscriber views a video, the Plaintiffs nonetheless maintain in their reply that “the Pixel would automatically fire when a subscriber viewed a prerecorded video.” (Pls.’ Reply at 8.) But to support their claim, they rely on Univision’s corporate representative’s testimony that “if a user views a video, a Pixel could be fired” (Univision Dep. at 58:10–12 (emphasis added)) and that Univision’s “intention” was that when a “video player page” loads “the Pixel would fire” (id. 63:15– 19 (emphasis added)). (Pls.’ Reply at 8.) This testimony doesn’t come even close to doing the heavy lifting the Plaintiffs require of it. If anything, it undercuts the Plaintiffs’ claims as to the immutability of the Pixel transmissions. In their reply, the Plaintiffs also maintain that, for the class to have fewer than forty members, “the Pixel would had to have fired in only 0.1% of instances where a subscriber viewed a prerecorded video.” (Pls.’ Reply at 8.) While the Plaintiffs argue “[t]hat is not a reasonable interpretation,” they don’t explain why except to point out that “Univision did not produce a single document evidencing any instance in which the Pixel malfunctioned.” (Pls.’ Reply at 8–9.) But, again, by the Plaintiffs’ own admission, there are any number of conditions—notwithstanding any “malfunction[s]”—that would prevent the Pixel from firing. (5) been simultaneously logged into Facebook using the same browser through which the subscriber selected the video; and (6) not deployed any number of browser settings or add-on software that would have blocked the Pixel. As previously explained, the Plaintiffs account for the first two conditions in their motion, reducing the potential class size, right off the bat on their own, from 35,845 to 17,000. (Pls.’ Reply. at 6 n. 3.) And though the Plaintiffs neglect to mention it their motion, the Court notes that their expert remarked that 79% of Facebook users never log out of Facebook, with only 7% logging out regularly and only 5% doing so always. (Egelman Rep. ¶ 42.) Conservatively, this would appear to further reduce the potential class by, at minimum, another 12% or to about 15,000 members. From there, however, the Plaintiffs leave the Court adrift. The Court has (1) no supportable way of guessing how many of these 15,000 were also using the same device on which they were logged in to their Facebook account as they used to access the Univision video; (2) no non- speculative basis for estimating how many of those subscribers were also logged into their Facebook account on the same web browser through which they accessed the Univision video; and then, (3) no meaningful way of surmising how many of those had not deployed any one of several mechanisms that would have blocked the Pixel. While certainly the Court may reasonably “make common sense assumptions in order to find support for numerosity,” the Plaintiffs here have made no effort to supply even an inkling of how these various factors might affect the class size here. Marko v. Benjamin & Brothers, LLC, No. 6:17-CV-1725- ORL-41GJK, 2018 WL 3650117, at *5 (M.D. Fla. May 11, 2018). They provide no way of estimating the number of users who log into Facebook on one device— say, a cell phone—and access Univision content on another—for example a laptop. Nor do the Plaintiffs hazard even a guess as to the number of users who, even if using the same device, access Facebook through either another web browser or the mobile application.11 And, as addressed above, the Plaintiffs fail to supply any indication of the percentage of users who, despite using a browser that would not block the Pixel by default, have deployed a setting or other software that would. Finally, the Plaintiffs fail to account for Univision
11 Univision’s expert, in explaining that the Pixel would not fire if a subscriber was logged into Facebook only via its mobile application, notes that “98.5% of [Facebook] users access[] the platform with a mobile device and 81.8% of those users only use a phone to access the social network.” (Vint Decl. ¶ 48.) This data point isn’t particularly helpful though because it doesn’t clarify how many users might log in to Facebook using its mobile application versus using a web browser. However, it proves the point that, without any guidance from the Plaintiffs, the percentage of subscribers satisfying any one condition, never mind all three, could readily drop the class size below the threshold of members considered presumptively impracticable to join. subscribers who might have accessed a prerecorded video on a device on which another person, perhaps another family member, has logged into Facebook using an account not associated with the Univision subscriber’s account. In such a case, according to Univision’s expert (and unrebutted by the Plaintiffs), the viewing history transmitted to Meta would not have been correlated to the Univision subscriber whose account was used to view the video. (Vint Decl. ¶¶ 59, 60.) Instead, the transmission would have linked the viewing history with the Facebook ID of whatever Facebook account happened to be logged in at the time, even if that Facebook account didn’t belong to the Univision subscriber through whose account the video was selected. This is particularly problematic because the Plaintiffs fail to explain how one might be able to differentiate between a Pixel transmission of, on the one hand, the Facebook ID of the Univision subscriber versus, on the other, the Facebook ID of whoever happened to be logged in on the device that was used by the Univision subscriber. Without “some showing” as to these several variables, the Plaintiffs have not afforded the court “the means to make a supported factual finding, that the class [to be] certified meets the numerosity requirement.” Vega, 564 F.3d at 1267 (emphasis in original). To be sure, any “inference of numerosity . . . without the aid of a shred of . . . evidence [would be] an exercise in sheer speculation.” Id.; see C-Mart, Inc. v. Metro. Life Ins. Co., 299 F.R.D. 679, 686 (S.D. Fla. 2014) (Middlebrooks, J.) (“[T]he Eleventh Circuit has made it abundantly clear that the burden to satisfy numerosity is on the plaintiff seeking to certify a class, and a plaintiff is not permitted to make a purely speculative showing that numerosity has been met.”) (cleaned up); Hayes v. Wal-Mart Stores, Inc., 725 F.3d 349, 358 (3d Cir. 2013) (“[W]here a putative class is some subset of a larger pool, the trial court may not infer numerosity from the number in the larger pool alone.”) The Plaintiffs “bear[] the burden of establishing every element of Rule 23 and a district court’s factual findings must find support in the evidence before it.” Vega, 564 F.3d at 1267. Although it may be tempting to assume numerosity based on the sheer size of the Plaintiffs’ starting point of 35,845 (or even 15,000) members, such baseless assumptions cannot carry the day. See Golden v. City of Columbus, 404 F.3d 950, 966 (6th Cir. 2005) (“[T[he party seeking class certification bears the burden of showing impracticability and mere speculation as to the number of parties involved is not sufficient to satisfy Rule 23(a)(1).”) (cleaned up); Colomar v. Mercy Hosp., Inc., 242 F.R.D. 671, 676 (S.D. Fla. 2007) (Seitz, J.) (denying certification where the plaintiff offered “no means for rationally estimating which number of the total uninsureds are class members,” as the plaintiff “cannot simply rely on the 24,000 figure and ask the Court to infer that a minimum number paid the full bill given the sheer size of the uninsured pool”). The Plaintiffs failure to supply anything more than speculation as to class size is particularly problematic in the face of Univision’s contention that the Plaintiffs, through their motion, have failed to come forward with evidence that any personally identifiable information at all has been disclosed. (Def.’s Resp. at 1.) While the Plaintiffs proffer that “Meta’s event data [will] show[] all Pixel transmissions from UNOW to Meta during the Class Period,” Univision points out that, based on the data produced by the time the motion for certification was filed, “[t]here is no record that any identifying information regarding prerecorded videos Plaintiffs allegedly viewed was transmitted or disclosed to Meta for any of the three named Plaintiffs.” (Def.’s Resp. at 12 (emphasis in original).) Indeed, as Univision explains (and the record appears to show), Meta told the Plaintiffs, in response to their subpoena, that, based on the information provided, it was unable to identify any Pixel event data for Plaintiffs Martinez and Rodriguez. (Id.) And, as to Plaintiff Giron, Meta’s information did not identify any specific video material he viewed. (Id.) Importantly, in their motion, the Plaintiffs neglect to identify any record evidence showing that each one of them watched a prerecorded video on Univision’s website while satisfying every single condition required for the Pixel to fire. So, not only do they fail on the numerosity front, their initial showing on even their very own claims is not particularly compelling. In sum, the Plaintiffs’ motion for class certification fails because they did not carry their burden of establishing numerosity. C. The Court declines to consider new information presented in the Plaintiffs’ reply. In their reply, the Plaintiffs point to newly identified records from Meta that they say “unequivocally show that Univision disclosed PII to Meta.” (Pls.’ Reply at 1.) These records include a spreadsheet of data that purportedly relates to Martinez as well as “an additional 2,800 rows of a sampling of Pixel event data” which “is expected to include multiple additional instances of Univision disclosing PII to Meta.” (Id. (emphasis added).) According to the Plaintiffs, these records, identified after they filed their motion for certification, establish numerosity because “Martinez’s Meta Event Data actually shows improper disclosures.” (Id. at 8.) Two problems. First, the Court is not inclined to consider this new information, supplied for the first time in in the Plaintiffs’ reply. Part of the problem with these newly supplied records is that the Plaintiffs fail to show that, despite their diligence, they were not able to access this evidence prior to filing their motion for class certification. Additionally, the introduction of these records now seems to move the goal posts mid-game. It shifts the focus of the Plaintiffs’ numerosity argument from being entirely on Univision’s subscriber-viewer records, to Meta’s records, which the Plaintiffs now say definitively show Univision’s repeated disclosures of personally identifiable information. The Court agrees with Univision that, to the extent this new position, based on new evidence, represents the Plaintiffs’ attempt to shift their showing of numerosity— from purported assumptions flowing solely from the sheer number of Univision subscribers who viewed prerecorded videos to purportedly multiple instances of Pixel disclosures shown by Meta’s event records—the tactic is decidedly unfair. See Douse v. Traeger, No. 22-13949, 2023 WL 5569289, at *3 (11th Cir. Aug. 29, 2023) (“[T]his Court does not consider arguments raised for the first time in a reply brief.”); Amargos v. Verified Nutrition, LLC, 666 F. Supp. 3d 1249, 1250 (S.D. Fla. 2022) (Bloom, J.) (“Arguments not properly presented in a party’s initial brief or raised for the first time in a reply brief are deemed waived.”) (cleaned up); Rindfleisch v. Gentiva Health Services, Inc., 22 F. Supp. 3d 1295, 1301 (N.D. Ga. 2014) (“As a general rule, federal courts do not consider arguments that are presented for the first time in a reply brief.”). The Court finds the new evidence and arguments presented here particularly problematic because they are not responsive to Univision’s opposition, but instead supply entirely new theories upon which the Plaintiffs seek to justify class certification. Second, even if the Court should, or were to, consider the new evidence and theories, the Plaintiffs’ new showing still fails to move the needle with respect to numerosity for several reasons. For starters, even assuming the new Meta event data did show that Univision disclosed Martinez’s personally identifiable information in at least one instance, this only gets the Plaintiffs to one “class member.” Nor is the Plaintiffs’ reliance on their expert’s opinion that other data Meta committed to producing (but hadn’t yet done so) “would include multiple instances of Univision disclosing Class members’ PII to Meta” any more helpful to their cause. It is far too speculative. Not only does their expert opine on data he has never seen, but his general description of “multiple” potential class members still fails to provide the Court with any real basis from which it could infer numerosity. Instead, it again amounts to “mere speculation as to the number of parties involved and general allegations of numerosity,” which “are insufficient to satisfy Rule 23(a)(1).” Gayle v. Meade, 614 F. Supp. 3d 1175, 1196 (S.D. Fla. 2020) (Cooke, J.). Perhaps recognizing this flaw, the Plaintiffs, in reply, continue to focus, once again, on the 35,845 Univision subscribers who viewed prerecorded videos in their endeavor to show numerosity. And, just as that theory falls short in their motion, it falls short in their reply as well. See Guarisma v. Hyatt Equities, LLC, No. 1:17-CV-20931-UU, 2017 WL 6949266, at *7 (S.D. Fla. Sept. 28, 2017) (Ungaro, J.) (finding a plaintiff failed to show numerosity when he posited that, even if only 1% of the potential 219,087 class members qualified, he would easily satisfy the numerosity requirement). In sum, the Court finds the new evidence and arguments springing therefrom are improperly submitted in reply, denying Univision a fair chance of fully addressing them in its opposition and resulting in an unnecessary multiplication of these proceedings. Accordingly, the Court declines to consider them. Conversely, even if the Court did consider them, the new evidence and argument fail, in any event, to help establish numerosity. Either way, Univision’s motion to strike that new evidence and argument (or, alternatively, for leave to file a sur-reply) is rendered moot. 4. Conclusion As set forth above, the Court denies the Plaintiffs’ motion for class certification (ECF Nos. 60, 61) and denies as moot Univision’s motion to strike (ECF Nos. 76, 79). Done and ordered in Miami, Florida, on October 1, 2024.
Robert N. Scola, Jr. United States District Judge