Opinion
21-md-02981-JD 20-cv-05761-JD
11-28-2022
In re Google Play Store Antitrust Litigation
ORDER RE CONSUMER PLAINTIFFS' CLASS CERTIFICATION MOTION AND DEFENDANTS' MOTION TO EXCLUDE EXPERT TESTIMONY RE: DKT. NOS. 280, 282
JAMES DONATO UNITED STATES DISTRICT JUDGE
This action by consumer plaintiffs is one of several antitrust cases about the Google Play Store. These cases have been consolidated into a multidistrict litigation (MDL) for centralized proceedings before this Court. Dkt. No. 1. The named plaintiffs allege, on behalf of themselves and multiple putative classes of consumers, that defendant Google illegally monopolized the Android app distribution market with anticompetitive practices in the Google Play Store.
Unless otherwise noted, all docket number references are to our district's ECF docket for the multidistrict litigation case, No. 21-md-02981-JD. For present purposes, the Court will cite to the redacted versions of filings pending resolution of a mountain of sealing requests in a separate order.
Google's motion to exclude the testimony of plaintiffs' economics expert is denied. The consumers' motion for class certification is granted in main part, subject to some adjustments of the named plaintiffs. Dkt. Nos. 251, 252.
BACKGROUND
The consumer case is itself a consolidated action. Before it was made a part of the MDL, the Court consolidated a number of related consumer cases under the caption, In re Google Play Consumer Antitrust Litigation. Consumer Dkt. No. 78. The Court appointed on an interim basis co-lead class counsel, liaison counsel, and a steering committee to manage the consumer side of the litigation. Consumer Dkt. No. 128.
References to the “Consumer docket” are to the ECF docket for Case No. 20-cv-05761-JD.
The operative complaint for the consumers is the consolidated second amended class action complaint. Dkt. No. 172 (SAC). The named plaintiffs are six consumers in the states of California, Massachusetts, New York, Washington, Wisconsin, and Georgia, all of whom purchased mobile apps through the Google Play Store or paid for in-app digital content for one of those apps. Id. ¶¶ 23-29; Consumer Dkt. No. 259. The defendants are Google, LLC, Google Ireland Limited, Google Commerce Limited, Google Asia Pacific Pte. Limited, and Google Payment Corp. (together, Google). SAC ¶¶ 32-36.
The thrust of the SAC is that Google has unlawfully acquired and maintained a monopoly in the Android app distribution market through anticompetitive practices in the Google Play Store. The Google Play Store is said to be the “dominant” distribution channel for mobile apps to Android device users. Id. ¶ 51. The Play Store features “over three million apps, including all the most popular Android apps,” compared to “just 700,000 apps offered by Aptoide, the Android app store with the next largest listing.” Id. ¶ 82. According to the SAC, “Google's market power results in enormous profits,” and “[i]n 2020 alone, the Google Play Store generated revenues of $38 billion, accounting for over 20 percent of the company's total revenue in that year of $182 billion.” Id. ¶ 86.
Plaintiffs allege that “Google has willfully and unlawfully maintained its monopoly in the Android Application Distribution Market through a series of related anticompetitive acts designed to foreclose alternative and competing Android app distribution channels.” Id. ¶ 111. The anticompetitive acts include requiring OEMs to preinstall and prominently place the Google Play Store on the Android devices they manufacture; requiring mobile network operators, in return for a share of Google's revenues, to preload the Google Play Store in a prominent position on all Android mobile devices that they distribute; and prohibiting developers who sell their apps through the Google Play Store from providing any apps that would allow consumers to download a competing app distribution store. Id. ¶¶ 112-54.
Plaintiffs say that Google's monopoly power allowed it to charge a “supra-competitive commission of up to 30% on the price of apps purchased through the Google Play Store and in-app purchases processed through Google Play Billing,” the use of which is mandated by Google for all apps that are distributed through the Play Store. Id. ¶ 84. Plaintiffs purchased Android apps and made in-app purchases “directly from Google,” and so were harmed by paying artificially inflated prices for the apps. Id. ¶¶ 208-11.
The SAC identifies three product markets -- “(1) the Licensable Mobile Operating System Market; (2) the Android Application Distribution Market; and (3) the In-App Aftermarket,” SAC ¶ 41 -- but alleges claims only with respect to the latter two. These claims are six counts against Google under Sections 1 and 2 of the Sherman Act, 15 U.S.C. §§ 1, 2, for unlawful monopolization, unreasonable restraints of trade, and unlawful tie-in in the Android Application Distribution Market and In-App Aftermarket, id. ¶¶ 221-79; four counts under the California Cartwright Act, Cal. Bus. & Prof. Code § 16700 et seq., for unreasonable restraints of trade and unlawful tie-in in the same two markets, id. ¶¶ 282-328; and one count under the California Unfair Competition Law (UCL), Cal. Bus. & Prof. Code § 17200 et seq., for unlawful, unfair, and fraudulent business practices, id. ¶¶ 329-62. The requested relief consists of “treble damages for injuries caused by defendants' violations of the federal antitrust laws and California's Cartwright Act,” restitution under the UCL, and a conduct injunction. Id. at 74.
Plaintiffs' motion for class certification, Dkt. No. 251, proposes classes that are somewhat different from those in the SAC. The SAC named a nationwide class or, in the alternative, a “repealer-state class.” SAC ¶¶ 213, 220, 280-81. The motion asks for certification of a smaller group, mainly because plaintiffs have entered into a Joint Prosecution Agreement with the Attorneys General of the 38 states and the District of Columbia, who are plaintiffs in State of Utah et al. v. Google, No. 21-cv-05227-JD, which is another constituent case in this MDL. Plaintiffs advised the Court that, “[t]o pursue consumers' claims against Google most effectively and efficiently, plaintiffs' counsel and the thirty-nine Attorneys General asserting parens patriae claims” have “agreed in the Joint Prosecution Agreement that class certification would be sought” in the consumers' case “only for consumers in states, districts and territories that have not asserted a parens patriae claim” in the States case. Dkt. No. 251 at 3. In effect, plaintiffs and the Attorneys General agreed that plaintiffs would not pursue certification on behalf of state residents represented in the Attorneys General case.
The SAC says that the “repealer-state class” consists of individuals in “those states whose laws permit indirect purchaser standing and provide for antitrust recovery to indirect purchasers.” SAC ¶ 213. Plaintiffs have abandoned the “‘repealer states' class” because “Google has consistently included a choice-of-law provision in its user agreements designating California law as controlling in litigation brought by users,” and so “California law governs the state law claims of all class members, regardless of where they reside and regardless of whether a particular state has ‘repealed' Illinois Brick.” Dkt. No. 251 at 3.
Plaintiffs propose certification of two classes for the Sherman Act, Cartwright Act, and UCL claims:
Rule 23(b)(3) Multistate Damages Class:
All persons in the following U.S. states and territories:
Alabama, Georgia, Hawaii, Illinois, Kansas, Maine, Michigan, Ohio, Pennsylvania, South Carolina, Wisconsin, Wyoming, American Samoa, Guam, Northern Mariana Islands, Puerto Rico, and the U.S. Virgin Islands who paid for an app through the Google Play Store or paid for in-app digital content (including subscriptions or ad-free versions of apps) through Google Play Billing on or after August 16, 2016, to the present.
Rule 23(b)(2) Multistate Injunctive Relief Class:
All persons in the following U.S. states and territories:
Alabama, Georgia, Hawaii, Illinois, Kansas, Maine, Michigan, Ohio, Pennsylvania, South Carolina, Wisconsin, Wyoming, American Samoa, Guam, Northern Mariana Islands, Puerto Rico, and the U.S. Virgin Islands
who currently own a mobile phone or tablet with an authorized and preinstalled version of Google's Android OS capable of accessing the Google Play Store.Dkt. No. 251 at i.
Plaintiffs filed a “corrected proposed order” that again changed the class definitions slightly, Dkt. No. 304, which will be discussed later.
Google opposes certification. Dkt. No. 273. It has also asked to exclude the testimony of plaintiffs' expert for certification, Dr. Hal J. Singer. Dkt. No. 252.
To aid the Daubert and class certification analysis, the Court held a concurrent expert proceeding, known informally as an “expert hot tub.” The hot tub featured Dr. Singer and Google's expert, Dr. Michelle M. Burtis, in a debate about the economic factors germane to the question of certification. Dkt. No. 299. The Court moderated the debate pursuant to a joint submission by the experts identifying their top areas of disagreement, which was prepared at the Court's direction. Dkt. Nos. 191, 284. The Court has used the hot tub procedure in other cases, and has found it to be an invaluable tool for vetting Daubert issues and determining questions of class certification, among other uses. See In re Capacitors Antitrust Litigation, No. 17-md-02801-JD, 2020 WL 870927 (N.D. Cal. Feb. 21, 2020); In re Capacitors Antitrust Litigation, No. 17-md-02801-JD, 2021 WL 5407452 (N.D. Cal. Nov. 18, 2021). The Court also heard argument by counsel on the class certification and Daubert motions. Dkt. No. 317.
For the pending motions, the Court will take up the Daubert challenge to Dr. Singer first. Because this challenge is made in the class certification context, reference to Rule 23 is required. The full Rule 23 determination will be made in an ensuing section.
DISCUSSION
I. LEGAL STANDARDS
Federal Rule of Evidence 702 provides that “[a] witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise if: (a) the expert's scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert has reliably applied the principles and methods to the facts of the case.” Put more plainly, expert opinions are admissible when they are relevant, supported by the evidence, based on sound methodologies, and useful to the jury on topics that ordinary people would not necessarily understand without help. The Rule 702 inquiry is “a flexible one,” with no “definitive checklist or test.” Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 593-94 (1993). The Court's task is to “ensur[e] that an expert's testimony both rests on a reliable foundation and is relevant to the task at hand.” Id. at 597. “Expert opinion testimony is relevant if the knowledge underlying it has a valid connection to the pertinent inquiry. And it is reliable if the knowledge underlying it has a reliable basis in the knowledge and experience of the relevant discipline.” Primiano v. Cook, 598 F.3d 558, 565 (9th Cir. 2010) (quotations and citation omitted).
To determine admissibility under Rule 702, the Court may consider factors such as “whether the theory or technique employed by the expert is generally accepted in the scientific community; whether it's been subjected to peer review and publication; whether it can be and has been tested; and whether the known or potential rate of error is acceptable.” Wendell v. GlaxoSmithKline LLC, 858 F.3d 1227, 1232 (9th Cir. 2017) (quotations and citation omitted). The Court may “also consider whether experts are testifying ‘about matters growing naturally' out of their own independent research, or if ‘they have developed their opinions expressly for purposes of testifying.'” Id. These factors are “illustrative, and they are not all applicable in each case,” and “Rule 702 should be applied with a ‘liberal thrust' favoring admission.” Id. (quotations and citation omitted); see also Primiano, 598 F.3d at 564 (“the trial court has discretion to decide how to test an expert's reliability as well as whether the testimony is reliable, based on ‘the particular circumstances of the particular case'”) (citation omitted).
“Ultimately, the test under Daubert is not the correctness of the expert's conclusions but the soundness of his methodology.” Elosu v. Middlefork Ranch Inc., 26 F.4th 1017, 1024 (9th Cir. 2022) (quotations and citation omitted). The Court “is ‘a gatekeeper, not a fact finder.'” Id. (quoting Primiano, 598 F.3d at 568). The Court will “exclude junk science that does not meet Federal Rule of Evidence 702's reliability standards by making a preliminary determination that the expert's testimony is reliable.” Ellis v. Costco Wholesale Corp., 657 F.3d 970, 982 (9th Cir. 2011) (citing Kumho Tire Co. v. Carmichael, 526 U.S. 137, 145 (1999)). Any objections short of that are fodder for cross-examination and not exclusion.
Dr. Singer's opinions were proffered in aid of plaintiffs' request to certify a class, and so the Daubert issues must be evaluated in light of Rule 23. The overall goal of Rule 23 is “to select the method best suited to adjudication of the controversy fairly and efficiently.” Amgen Inc. v. Conn. Ret. Plans & Trust Funds, 568 U.S. 455, 460 (2013) (cleaned up). “The class action is ‘an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only.'” Comcast Corp. v. Behrend, 569 U.S. 27, 33 (2013) (citation omitted).
To come within the exception, plaintiffs bear the burden of proving by a preponderance of the evidence that the proposed classes satisfy all four requirements of Rule 23(a) and at least one of the subsections of Rule 23(b). Id.; Olean Wholesale Grocery Coop., Inc. v. Bumble Bee Foods LLC, 31 F.4th 651, 664-65 (9th Cir. 2022) (en banc), cert. denied, S.Ct., 2022 WL 16909174 (Nov. 14, 2022). The Court's analysis “must be rigorous and may entail some overlap with the merits of the plaintiff's underlying claim,” though the merits questions are to be considered only to the extent that they are “relevant to determining whether the Rule 23 prerequisites for class certification are satisfied.” Amgen, 568 U.S. at 465-66 (internal quotations and citations omitted). The class certification procedure is decidedly not an alternative form of summary judgment or an occasion to hold a mini-trial on the merits. Alcantar v. Hobart Service, 800 F.3d 1047, 1053 (9th Cir. 2015). The decision of whether to certify a class is entrusted to the sound discretion of the district court. Zinser v. Accufix Research Inst., Inc., 253 F.3d 1180, 1186 (9th Cir. 2001).
The analysis of expert testimony for class certification has some specific elements. “When considering class certification under Rule 23, district courts are not only at liberty to, but must perform ‘a rigorous analysis [to ensure] that the prerequisites of Rule 23(a) have been satisfied.'” Ellis, 657 F.3d at 980 (quoting Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 351-52 (2011)). The Court may not certify a class just because the expert evidence is admissible. Id. at 982. The Court must directly answer the Rule 23 question of “whether the plaintiffs' evidence is capable of resolving a common issue central to the plaintiffs' claims.” Olean, 31 F.4th at 667. To do that, the Court will decide if the expert's methodology is “capable of showing class-wide antitrust impact” in light of “factors that may undercut the model's reliability (such as unsupported assumptions, erroneous inputs, or nonsensical outputs such as false positives) and resolve[] disputes raised by the parties.” Id. at 683 (emphasis omitted). While this is a rigorous inquiry, the “district court's findings at the certification stage ‘do not bind the fact-finder on the merits.'” Id. at 667 n.10 (citation omitted).
II. THE MOTION TO EXCLUDE
Google does not suggest that Dr. Singer is unqualified to be an expert witness on economics, and for good reason. Dr. Singer is a managing director at Econ One, an economic consulting firm, and is an adjunct professor at the McDonough School of Business at Georgetown University, where he teaches MBA candidates. Dkt. No. 254-4 (Singer Report). He received a Ph.D. in economics from the Johns Hopkins University in 1999, has been involved with antitrust issues as an economist throughout his career, and has published articles in antitrust journals and presented at antitrust events. Dr. Singer has testified about competition issues before the United States Congress on multiple occasions. Id. He is qualified as an expert in economics, and Google does not contend otherwise.
Google's main Daubert objection is to Dr. Singer's “pass-through formula.” Dkt. No. 252 at 6-13. The formula is an essential element of his opinions about Google's overcharges in app sales, and the artificially inflated prices consumers paid as a consequence. These opinions are rather complex, and they start with Dr. Singer's description of the relevant product market.
Dr. Singer proposes two relevant markets: an Android App Distribution Market, which is “the market for the sale and distribution of Apps for Android mobile devices,” and an In-App Aftermarket, “the ancillary aftermarket for services in support of consummating purchases of In-App Content.” Singer Report ¶ 2. Dr. Singer posits, without objection by Google, that the Android App Distribution Market is a two-sided market in that it “matches buyers (in this case consumers) and sellers (in this case app developers). Two-sided platforms benefit from ‘indirect network effects,' meaning that each additional buyer makes the platform more appealing to sellers.” Id. ¶ 51. The In-App Aftermarket is “one-sided: It is a simple transaction between a buyer (the developer) and a seller of services, including payment processing, record keeping, and unlocking of content, needed to consummate a purchase of In-App Content.” Id. ¶ 27.
Dr. Singer says that Google charges a “take rate generally of 30 percent” in these two markets, id. ¶ 29, meaning that Google “takes 30 percent of all revenues on the original sale and downloading of Apps from the Play Store and the sale of digital content within Apps,” id. ¶ 1. Dr. Singer assessed “the competitive effects of the various restrictions Google enforces (collectively, the ‘Challenged Conduct'), to extract these ‘take rates,' and, in particular, . . . assess[ed] whether, as a result of the Challenged Conduct, consumers have overpaid for the initial downloads of Apps through the Play Store and for purchases of In-App Content.” Id.
It was important for Dr. Singer to figure out whether Google's “take rate” would have been lower absent the challenged conduct, and if so, whether that would have translated into lower prices for consumers. For Play Store transactions, Google typically collects the full payment amount directly from consumers and then shares that revenue with developers according to Google's revenue-sharing agreement with each developer. The prices consumers pay for apps and in-app content are set independently by the developers in their sole discretion. See Singer Report ¶¶ 19, 21, 27, 57, 135, 137, 175, 227.
In the hot tub debate, Dr. Singer said that he used for the two-sided Android App Distribution Market the Rochet-Tirole model of pricing to estimate a “but-for take rate” that Google would have imposed on Play Store transactions “in a but-for world absent the challenged conduct.” Dkt. No. 302 (Hot Tub Tr.) at 9:24-10:3. To estimate the “but-for take rate” for the one-sided In-App Aftermarket, he utilized the Landes-Posner model, developed by economist William Landes and Judge Richard Posner. Id. at 40:8-41:14. Dr. Singer's “pass-through” rate was an input into both of these models. See id. at 9:19, 41:21-22.
The “pass-through” rate is a critical element of Dr. Singer's overcharge analysis, and is the main point of contention in the Daubert dispute. In Dr. Singer's conception, “Google sets a take rate or commission imposed directly on developers,” and the “pass-through rate” is the “portion of the supracompetitive cost imposed on developers through the take rate [that] is passed through to consumers.” Singer Report ¶ 180. Dr. Singer stated at the hot tub that “92.4 percent of the transactions in the database were all at that headline 30 percent rate,” Hot Tub Tr. at 60:1-4; in this case, there was also no “before and after” period for the challenged conduct. A traditional regression analysis was therefore not possible. Id. 62:21-63:7; Singer Report ¶ 168. This is why Dr. Singer “looked for an economic model of consumer demand that would allow [him] to make predictions of how an app developer would change its price in response to a change in the take rate, given the nature of the demand that that app developer faced.” Hot Tub Tr. at 60:12-16. He found that “the logit model captures the demand faced by app developers.” Id. at 60:18-19. Logit models are often used in merger cases to “map a change in the merging parties' costs that come about from merger synergies into a change in price.” Id. at 61:18-20.
Dr. Singer applied the logit model here, through a series of calculations, to conclude that the pass-through rate for each app is one minus an app's share in the app's chosen Play Store category. Id. at 76:5-7, 97:21-99:7; Singer Report ¶¶ 235-240. The Play Store is divided into more than 34 categories under headings such as “dating,” “entertainment,” “games,” and “sports,” and developers self-select one of these categories to market their apps to consumers in the Play Store. Dkt. No. 254-6 (Singer Reply Report) ¶¶ 75-76 & Table 2.
Dr. Singer stated at the hot tub that there were 35 categories until Google removed the “transportation” category in 2016. Hot Tub Tr. at 82:14-16.
Dr. Singer opined that a category-share-dependent pass-through formula makes intuitive sense because each developer is “competing against everyone within the category.” Hot Tub Tr. at 78:6. He concluded that “the logit model makes a very specific prediction about the relationship between an app's share within its category and its price; and in particular, the prediction is that as the app's price goes up, it should lose share within the category, reflecting the fact that all of these apps within the category are substitutes in some way.” Id. at 81:21-82:1. Dr. Singer tested the fit between the logit demand model and the transactional data available to him, and he “found a very tight fit” for every category. Id. at 82:2-3. He noted that these categories are “meaningful arena[s] of competition . . . which one can use for estimating shares for the logit model.” Id. 117:17-21.
Google has not suggested that this overall approach is “junk science” destined for the scrap heap under Daubert. At the expert hot tub, Google's expert, Dr. Michelle Burtis, forthrightly acknowledged that the Rochet-Tirole model “is used in [the] economic literature,” and “is used in the way that Dr. Singer is using it here.” Hot Tub Tr. at 16:2-7. With respect to the Landes- Posner model, Dr. Burtis stated, “Economists use these kinds of equations. I certainly wouldn't say that this is, you know, junk science in that regard.” Id. at 44:19-24.
Dr. Burtis's critique focused on Dr. Singer's “methodology for the pass-through rates,” mainly because she thought it was “not standard” and had “never seen it before.” Id. at 51:5-11. As Daubert objections, these comments do not go far. Rule 702 does not forbid new methodologies and analyses. It is true that “well-established propositions are less likely to be challenged than those that are novel, and they are more handily defended.” Daubert, 509 U.S. at 592 n.11. But general acceptance of a method is no guarantee of reliability, and is not a touchstone of admissibility under Rule 702. See id. (“Nor, on the other hand, does the presence of Daubert's general acceptance factor help show that an expert's testimony is reliable where the discipline itself lacks reliability, as, for example, do theories grounded in any so-called generally accepted principles of astrology or necromancy.”). The “gatekeeping inquiry must be tied to the facts of a particular case,” and so it is not necessarily “surprising” for expert opinions to be based on methods that are new and not been the subject of peer review. Kumho Tire, 526 U.S. at 150-51 (cleaned up).
Dr. Singer addressed the “novelty” point by explaining why he could not use the more traditional approach of a regression analysis. See id. at 60:1-4, 62:21-63:7; Singer Report ¶ 168. He applied a version of the logit model, which has been used in the merger context, after running tests to confirm that the model fit the data here. Dr. Burtis made much of the fact that the original version of the model utilized a “per unit” cost rather than an “ad valorem” cost (i.e., a “percentage of the price” cost, as is the case here), but did not say why this might be a fatal flaw or “junk science.” See, e.g., Hot Tub Tr. at 56:15-57:23. At the hot tub, Dr. Burtis stopped short of saying that Dr. Singer should have used a regression analysis, and did not identify any other model he might have used. She also did not say that it was categorically wrong for Dr. Singer to apply the logit model here, and instead simply criticized the inputs he selected. See id. at 90:25-91:3 (“if he wanted to use the logit model -- it was his choice. He wanted to use it. So if you're going to use it, do it right. Figure out the groupings of products that are truly substitutes for one another.”). That again is the stuff of cross-examination and not exclusion.
Overall, Google has not demonstrated that unreliability or invalidity warrant exclusion of Dr. Singer's opinions. Its heart may not have been in that. Google's arguments were directed far more to opposing certification than to disqualifying Dr. Singer. Every substantive point made by Dr. Burtis at the hot tub, and by Google in its Daubert motion (including, for example, Dr. Singer's failure to account for developers' marginal costs and for focal point pricing), was presented mainly to say that plaintiffs cannot establish commonality and predominance for Rule 23 purposes. Compare Hot Tub. Tr. & Dkt. No. 252 with Dkt. No. 273. That is a wholly different question from admissibility. Exclusion of Dr. Singer's opinions under Rule 702 is denied.
III. THE RULE 23(B)(3) CLASS
Plaintiffs propose certification under Rule 23(b)(2) and (3). Dkt. No. 251. The (b)(3) proposal got the most discussion in the parties' briefs, and the Court will start with that.
Under Rule 23(b)(3), a class is appropriate when “questions of law or fact common to class members predominate over any questions affecting only individual members,” and a class action is “superior to other available methods for fairly and efficiently adjudicating the controversy.” Fed.R.Civ.P. 23(b)(3). Each of the four requirements of Rule 23(a) -- sufficiently numerous parties, common questions of law or fact, typicality of claims or defenses, and adequacy of representation -- must also be met.
A. Numerosity (23(a)(1))
Rule 23(a)(1) requires that the proposed class be “so numerous that joinder of all members is impracticable.” Fed.R.Civ.P. 23(a)(1). There is no dispute between the parties that there are more than 21 million putative class members. Dkt. No. 273 at 4; Dkt. No. 289 at 13. The numerosity requirement is satisfied.
B. Commonality (23(a)(2)) and Predominance (23(b)(3))
The commonality requirement under Rule 23(a)(2) is satisfied when “there are questions of law or fact common to the class.” Fed.R.Civ.P. 23(a)(2). Because “any competently crafted class complaint literally raises common questions,” the Court's task is to look for a common contention “capable of classwide resolution -- which means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke.” Alcantar, 800 F.3d at 1052 (cleaned up). What matters is the “capacity of a class-wide proceeding to generate common answers apt to drive the resolution of the litigation.” Dukes, 564 U.S. at 350 (quotations omitted, emphasis in original). This does not require total uniformity across a class. “The existence of shared legal issues with divergent factual predicates is sufficient, as is a common core of salient facts coupled with disparate legal remedies within the class.” Hanlon v. Chrysler Corp., 150 F.3d 1011, 1019 (9th Cir. 1998), overruled on other grounds by Dukes, 564 U.S. 338. The commonality standard imposed by Rule 23(a)(2) is, however, “rigorous.” Leyva v. Medline Indus. Inc., 716 F.3d 510, 512 (9th Cir. 2013).
Rule 23(b)(3) sets out the related but nonetheless distinct requirement that the common questions of law or fact predominate over the individual ones. This inquiry focuses on “whether the common, aggregation-enabling, issues in the case are more prevalent or important than the non-common, aggregation-defeating, individual issues.” Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. 442, 453 (2016) (cleaned up). Each element of a claim need not be susceptible to classwide proof, Amgen, 568 U.S. at 468-69, and the “important questions apt to drive the resolution of the litigation are given more weight in the predominance analysis over individualized questions which are of considerably less significance to the claims of the class.” Torres v. Mercer Canyons Inc., 835 F.3d 1125, 1134 (9th Cir. 2016). Rule 23(b)(3) permits certification when “one or more of the central issues in the action are common to the class and can be said to predominate, . . . even though other important matters will have to be tried separately, such as damages or some affirmative defenses peculiar to some individual class members.” Tyson Foods, 577 U.S. at 453 (internal quotations omitted). “Rule 23(b)(3)'s predominance criterion is even more demanding than Rule 23(a).” Comcast, 569 U.S. at 34.
The “requirements of Rule 23(b)(3) overlap with the requirements of Rule 23(a)” and “courts must consider cases examining both subsections in performing a Rule 23(b)(3) analysis.” Olean, 31 F.4th at 664. The Court finds it appropriate to assess commonality and predominance in tandem, with a careful eye toward ensuring that the specific requirements of each are fully satisfied. See, e.g., Just Film, Inc. v. Buono, 847 F.3d 1108, 1120-21 (9th Cir. 2017).
“Considering whether ‘questions of law or fact common to class members predominate' begins, of course, with the elements of the underlying cause of action.” Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S. 804, 809 (2011). For cases like this one brought under the Sherman Act, 15 U.S.C. § 15, and California's Cartwright Act, Cal. Bus. & Prof. Code § 16700 et seq., “[t]he elements of a claim for such antitrust action[s] are (i) the existence of an antitrust violation, (ii) ‘antitrust injury' or ‘impact' flowing from that violation,” and “(iii) measurable damages.” Olean, 31 F.4th at 665-66 (citations omitted). “[T]o prove there is a common question of law or fact that relates to a central issue in an antitrust class action, plaintiffs must establish that ‘essential elements of the cause of action,' such as the existence of an antitrust violation or antitrust impact, are capable of being established through a common body of evidence, applicable to the whole class.” Id. at 666. The Supreme Court has noted that “[p]redominance is a test readily met in certain cases alleging consumer or securities fraud or violations of the antitrust laws.” Amchem Products, Inc. v. Windsor, 521 U.S. 591, 625 (1997).
Plaintiffs' UCL claim is derivative of the Sherman Act and Cartwright Act claims. See SAC ¶¶ 329-62.
1. Sherman Act Claims
Overall, plaintiffs have established commonality and predominance for the Sherman Act antitrust claims. The parties vigorously engaged with each other on this issue, and Google raised a host of objections that are resolved in the ensuing discussion. In the end, the record establishes that plaintiffs satisfied their burden of proof on these elements.
a. Antitrust Violation
Plaintiffs' main claim is for an unlawful monopoly under Section 2 of the Sherman Act, 15 U.S.C. § 2. To establish Google's liability for that claim, plaintiffs will need to show: “(a) [Google's] possession of monopoly power in the relevant market; (b) the willful acquisition or maintenance of that power; and (c) causal antitrust injury.” Federal Trade Commission v. Qualcomm Inc., 969 F.3d 974, 990 (9th Cir. 2020) (internal quotations omitted). For the Section 1 claim of restraint of trade, plaintiffs will need to show a contract, combination, or conspiracy that unreasonably restrains interstate commerce. 15 U.S.C. § 1.
“A threshold step in any antitrust case is to accurately define the relevant market, which refers to ‘the area of effective competition.'” Qualcomm, 969 F.3d at 992 (quotations and citations omitted). Plaintiffs' class certification motion uses Dr. Singer's Android App Distribution Market and the In-App Aftermarket as the two relevant product markets. Dkt. No. 251 at 4. Dr. Burtis accepted these market definitions for class certification purposes. Id. Consequently, there is no dispute about the relevant markets for certification purposes.
Plaintiffs have detailed the common evidence that the class members will present at trial to prove Google's anticompetitive conduct. Google does not disagree that common evidence is available to prove Google's alleged antitrust violations in the relevant markets. The Court's independent analysis confirms this is so. The question of liability will be answered by common evidence about Google's conduct with respect to all consumers in the relevant markets, such as Google's agreements with mobile carriers, OEMs, and developers. See Dkt. No. 251 at 4-10. There will be no need to make individualized inquires for any specific plaintiff, and Google does not suggest otherwise. Dkt. No. 273. Common questions predominate for the antitrust violation element of plaintiffs' claims.
b. Antitrust Injury or Impact
“‘Antitrust injury' is ‘injury of the type the antitrust laws were intended to prevent and that flows from that which makes defendants' acts unlawful.” Olean, 31 F.4th at 666 (quoting Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 489 (1977)). For class certification purposes, plaintiffs “must establish, predominantly with generalized evidence, that all (or nearly all) members of the class suffered damage as a result of defendants' alleged anti-competitive conduct.” In re Packaged Seafood Products Antitrust Litig., 332 F.R.D. 308, 320 (S.D. Cal. 2019) (quotations and citation omitted), aff'd, Olean, 31 F.4th 651. Plaintiffs rely on Dr. Singer's analysis as their common method of proving antitrust impact.
This is the certification element most hotly contested by Google. Google's main contention is that plaintiffs have “no common proof of pass-through.” Dkt. No. 273 at 9-16. In Google's view, this alone bars certification of a (b)(3) class.
To start the discussion, the “pass-through” concept needs clarification. In the antitrust context, pass-through typically refers to the passing on of overcharges through distribution networks to downstream purchasers. The classic example is found in Illinois Brick Co. v. Illinois, 431 U.S. 720 (1977), which involved a traditional, vertical supply chain in which a defendant accused of anticompetitive conduct (Illinois Brick) sold concrete blocks to masonry contractors, who in turn sold to general contractors, who then sold to plaintiff the State of Illinois, the ultimate consumer of the blocks. The plaintiff did not buy directly from Illinois Brick, which was alleged to have engaged in a conspiracy to fix the price of the concrete blocks. In that context, the Supreme Court determined that the plaintiff, as an indirect purchaser, could not pursue an antitrust claim against the defendant based on a pass-though of the anticompetitive overcharge through the distribution chain to the ultimate consumer. See Apple Inc. v. Pepper, 139 S.Ct. 1514, 1521 (2019). This was based mainly on concerns about multiple recoveries for the same anticompetitive injury, and the uncertainties inherent in determining what portion of an overcharge was passed on. See id. at 1524.
For obvious reasons, Google analogizes the Play Store to this traditional supply chain situation. See Dkt. No. 273 at 9-10 (“[p]roof of pass-through” is “complex” because “it ‘must account for the actions of innocent intermediaries who allegedly passed on the overcharge.'”). The problem for Google is that this theory has been definitively held not to apply to online app markets that do not operate as traditional, vertical supply chains.
As discussed in the background section, the parties agree that the developers set the prices for their apps and content, but Google collects the payment from the consumers, and keeps a cut of it before paying the remainder to the developers. This is the same system used in the Apple App Store, which sells apps to Apple users just as the Play Store sells them to Android users, and the Supreme Court had no trouble concluding that the traditional pass-through concept was inapposite. “There is no intermediary in the distribution chain between Apple and the consumer. The iPhone owners purchase apps directly from the retailer Apple, who is the alleged antitrust violator. The iPhone owners pay the alleged overcharge directly to Apple.” Apple, 139 S.Ct. at 1521. “The absence of an intermediary is dispositive” of the device owners' status as direct purchasers, and “[t]he overcharge has not been passed on by anyone to anyone.” Id. at 1521, 1525.
So too, here. The consumer plaintiffs paid the alleged overcharge directly to Google. Consequently, there is no barrier to certification on this score.
To be sure, plaintiffs were not always crystal clear when talking about pass-through. They have embraced Apple with a bear hug, starting with the complaint, which alleges that “[f]or every in-app purchase, just as for an initial app purchase, consumers pay Google, not the app developer. Google then taxes the transaction at the exorbitant rate of up to 30%, remitting the remainder to the developer, who is responsible for setting the purchase price to the consumer.” SAC ¶ 18; see also id. ¶ 220 (plaintiffs are “direct purchasers under Apple v. Pepper”). They emphasized in their class certification motion that they “paid Google directly for these purchases, and, accordingly, are direct purchasers who may sue Google under federal antitrust laws.” Dkt. No. 251 at 1 (citing Apple, 139 S.Ct. at 1520).
Even so, Dr. Singer said in his report that a “portion of the supracompetitive cost imposed on developers through the take rate is passed through to consumers.” Singer Report ¶ 180. He claimed that he “take[s] no position on whether proof of pass-through is necessary under the law.” Id. ¶ 222. But as previously indicated, his overall opinions are based on a formula that calculates “pass-through rates.” Id. ¶¶ 222-40.
The record amply demonstrates that there is no substantive confusion here. Dr. Singer and plaintiffs plainly understand the Play Store consumers to be direct purchasers. It would have been more precise under Apple to talk in terms of the share of the overcharge borne by the consumers. But the direct consumer payment of an overcharge is, in fact, what plaintiffs and their expert focus on.
The pass-through formula is suitable as an element of classwide proof of antitrust impact and injury. The formula was an input for both the Rochet-Tirole model (which Dr. Singer used for the Android App Distribution Market) and the Landes-Posner model (used for the In-App Aftermarket). It was derived from a logit model, which captured the demand curve faced by the developers who sell apps and content in the Google Play Store. Dr. Singer determined that the pass-through formula may ultimately be expressed as “one minus the share” an app has in its selfselected Play Store category. Dr. Singer calculated the pass-through rate for each category, then calculated that the average pass-through rate across all categories was 89.9%, meaning that consumers across the board paid an estimated 89.9% of Google's commission on Play Store transactions. Singer Report at 112 (Table 8). Using this rate in the Rochet-Tirole model, Dr. Singer “estimate[d] that in the but-for world, platform competition results in a competitive take rate of 23.4 percent, down from its observed value of 30.1 percent in the actual world.” Id. ¶ 193. “This difference results in an average overcharge to consumers of $0.30 per paid App download,” with “aggregate damages of $18.76 million.” Id. Using the 89.9% average passthrough rate in the Landes-Posner model for the In-App Aftermarket, Dr. Singer concluded that, for in-app content, “Google's take rate would fall to 14.8 percent in th[e] competitive but-for world” from a “take rate in the actual world of 29.2 percent.” Id. ¶¶ 218, 220. This would result in an average $1.34 consumer savings per transaction and an aggregate damage figure of $4.71 billion, by Dr. Singer's calculations. Id. at 98 (Table 5).
Google fires a blunderbuss of objections at the pass-through formula, none of which hit the mark. To start, Google says that Dr. Singer ignored developers' marginal costs, Dkt. No. 273 at 10-11, but that is not a fair characterization of his analysis. Dr. Singer stated that “Google's requirement that developers pay a percentage of their revenue to Google is mathematically equivalent to an increase in developers' marginal cost,” Singer Report ¶ 225, and his analyses are focused on modeling what is likely to happen when there is a change to that significant, common component of developers' marginal cost. See, e.g., Singer Reply Report ¶ 72 (“When marginal cost falls -- due, in this case, to a substantial and permanent reduction in the take rate -- developers will find that their prior prices are no longer profit-maximizing. Standard economics prescribes that developers will therefore decrease their prices until marginal revenue once again equals marginal cost.”); Hot Tub Tr. at 101:18-23 (“What we can observe is what the change in the marginal cost would be in a but-for world....And a logit model gives us a way to map that change in the marginal cost into a change in prices.”).
Google says that if a developer's marginal cost is zero, there is no pass-through, and so by failing to account for marginal costs, Dr. Singer failed to account for these hypothetical developers who would have had no pass-through because they had no marginal costs. Dkt. No. 273 at 10-11. At the hot tub, however, Dr. Burtis acknowledged that while replication costs (e.g., the cost to “make” an additional, say, digital sword) may be zero, it is highly unlikely that any developer's marginal costs would actually be zero, because developers necessarily have costs that exist outside of the digital world, such as the cost of digital storage, a computer to code on, or an engineer's time. See Hot Tub Tr. at 94:2-95:3.
Google also attacks Dr. Singer's use of the logit model for the pass-through analysis, namely the derivation of the “one minus the share” formula, which is heavily dependent on the Play Store category self-selected by each app developer. Google says that this is wrong because the categories are broadly defined, and so apps within a category are not necessarily substitutes for each other, or subject to the same competitive forces. The “Games” category, for example, hosts educational games for kids and wargames for adults, which typically would not be interchangeable or competing products. See Dkt. No. 273 at 12-13.
Why this might be fatal to certification is entirely unclear. Google made the point mainly as an observation, rather than a well-formulated analysis, and its persuasive value is negligible. As the Court observed at the hot tub, the 35 categories are what Google gives to developers, and “Dr. Singer can only work with what Google actually does.” Hot Tub Tr. at 89:11-19. Interestingly, Apple's App Store offers very similar categories as the Google Play Store. See Singer Reply Report at 31 (Table 2). It is worth noting that, in a similar antitrust action challenging Apple's App Store practices (indeed, in the underlying district court case that gave rise to the Apple v. Pepper decision by the Supreme Court), the consumer plaintiffs' expert, who won a Nobel Prize in economics, also modeled demand and supply conditions on a category-by-category basis. See Expert Report of Dr. Daniel L. McFadden, In re Apple iPhone Antitrust Litigation, Case No. 4:11-cv-06714-YGR (N.D. Cal.), Dkt. No. 443-14 ¶ 211 (model modified to reflect demand and supply conditions that are specific to each app category; observing that consumers who purchased apps in the Games category may have a different degree of price sensitivity from those who purchased apps in the Music category).
Overall, Dr. Singer's use of the Google Play Store categories is reasonable in light of the evidence, and his approach to those categories does not make his analyses incapable of showing classwide impact. See Olean, 31 F.4th at 676. Google may cross-examine Dr. Singer about its objections at trial, but they are not a reason to deny class certification.
Google offers a grab bag of “other idiosyncratic factors” that Dr. Singer is said not to have accounted for. Dkt. No. 273 at 13. These include, for example, the possibility that even if Google's take rate were lower, a developer might pocket the extra money or give it away, rather than lowering prices for consumers. Google also suggests that some developers might have other marginal costs that are exceptionally large, thereby changing the analysis, or that they may not reduce their prices to the full extent predicted by Dr. Singer because they engage in “focal point pricing,” i.e., ending their prices in $.99. Id. at 11-12.
The thrust of these observations is that Dr. Singer's methods may not totally eliminate the possibility of some individualized issues for class members. They do not need to. Rule 23 does not demand that all of the world's complexities be smoothed away. To obtain certification, the plaintiffs' burden is to show that the “common, aggregation-enabling, issues in the case are more prevalent or important than” those individual issues. Olean, 31 F.4th at 664 (quoting Tyson Foods, 577 U.S. at 453); see also In re Hyundai and Kia Fuel Economy Litig., 926 F.3d 539, 55758 (9th Cir. 2019) (en banc) (“even if just one common question predominates, ‘the action may be considered proper under Rule 23(b)(3) even though other important matters will have to be tried separately.'”) (cleaned up). Put more simply, predominance does not demand perfection, and Google has not demonstrated that Dr. Singer's analysis falls short on this score. Moreover, Dr. Singer stated that his method can be customized to fit particular situations. His same methodology can be run “at an app-by-app level as well,” although for obvious space and time constraint reasons, he did not do that in his report. Hot Tub Tr. at 86:10-15 (“We could allow for these percentage overcharges to vary by app. But what I've offered is a reliable and common methodology that could apply to ever member of the class.”). The salient point is that the same methodology can be used by every class member to establish antitrust impact.
Google contrasts Dr. Singer's pass-through analysis with Dr. Burtis's conclusion that, “when Google reduced service fees for many transactions in 2018, 2021, and 2022, developers in the data set only reduced prices for about 2% of products subject to the service fee reductions.” Dkt. No. 273 at 9. This observation is again of minimal value. At the hot tub, Dr. Singer raised several serious questions about Dr. Burtis's analysis, such as whether she misinterpreted product SKUs in the transactional database by missing the fact that some SKUs were different listings for the same product. See Hot Tub Tr. at 64:3-75:10. Consequently, the Court cannot say that Dr. Burtis's 2% conclusion is enough to deny certification at this time.
Google's suggestion that Dr. Singer did not do any empirical analyses is wholly unfounded. The record, namely his report and hot tub testimony, amply establish that his opinions were solidly grounded in the transactional data and other evidence in the case. See id. at 9:18; Singer Report. He also ran multiple tests using that real-world data to confirm the accuracy of his predictions about such things as “the relationship between an app's share within its category and its price.” Hot Tub Tr. at 81:18-20.
Google says that Dr. Singer's methods cannot identify “uninjured” class members. Dkt. No. 273 at 9. In effect, Google demands that each class member individually prove an injury before certification may be granted. The law provides otherwise. See Olean, 31 F.4th at 668-69 (“a district court is not precluded from certifying a class even if plaintiffs may have to prove individualized damages at trial, a conclusion implicitly based on the determination that such individualized issues do not predominate over common ones.”). It is true that a class may not be certified when it would be so overinclusive that substantial numbers of uninjured people would populate it. See id. at 669. Google has not shown this is a concern here. In addition, plaintiffs' case is that Google's monopolistic practices inflated the “headline rate” that was used as the basis for all developers' negotiations with Google, which affected all of the prices set by the developers and paid by consumers to Google. This is not unlike a price-fixing case, where the “price-fixing affects all market participants, creating an inference of class-wide impact even when prices are individually negotiated.” Id. at 671 (quoting In re Urethane Antitrust Litigation, 768 F.3d 1245, 1254 (10th Cir. 2014)). “Setting the certification bar at the extreme height defendants propose would almost certainly kill off most antitrust class actions well before an adjudication of the merits of the case.” In re Capacitors Antitrust Litigation (No. III), Case No. 17-md-02801-JD, 2018 WL 5980139, at *7 (N.D. Cal. Nov. 14, 2018).
To be clear, the Court is not deciding the question of certification here on the basis of price fixing cases. What matters is that a defendant's common practices are often a good basis for establishing classwide proof of liability and injury. For the same reason, the fact that Google's service fee rates have in fact differed at the margins across developers is not fatal to certification. Plaintiffs may still argue, using Dr. Singer's analysis, that everyone was injured because the headline rate, which was the starting point for all negotiations, was affected.
Google says that Dr. Singer's alternative Play Points model is not common proof of impact. Dkt. No. 273 at 18-19. Google's Play Points program is a consumer reward program, akin to airline frequent flyer miles and the like. For this model, Dr. Singer used the Rochet-Tirole model, but he held the “but-for take rate . . . fixed at its observed average value of 29.3 percent.” Singer Report ¶ 251. For this part of his analysis, he used the Rochet-Tirole model for both the Android App Distribution Market and the In-App Aftermarket. He concluded that, “[a]ccording to this model, the Play Points program would be expanded to be worth an average of $0.77 per transaction, or approximately 8.7 percent of consumer spend (in the competitive but-for world). Because the expanded Play Points program is a direct subsidy to consumers, there is no need to estimate a pass-through model to establish antitrust impact.” Id. ¶ 253.
Dr. Singer's analysis stumbles a bit on this point. It is hard to square the Play Points model with plaintiffs' injury claims. The SAC focuses heavily on the overcharge that consumers paid because of Google's monopolistic practices. See, e.g., SAC ¶ 18 (“As a result, consumers pay more for applications and in-app purchases than they would in a competitive market. For every in-app purchase, just as for an initial app purchase, consumers pay Google ....Google then taxes the transaction at the exorbitant rate of up to 30%”). The SAC never mentions fewer Play Points as another injury, which gives pause about the applicability of the Play Points model. See Comcast, 569 U.S. at 35 (“at the class-certification stage (as at trial), any model supporting a ‘plaintiff's damages case must be consistent with its liability case, particularly with respect to the alleged anticompetitive effect of the violation'”) (citations omitted). For present purposes, the Court concludes that Dr. Singer's main overcharge models satisfy the commonality and predominance requirements of certification. The Court has not relied on the Play Points model for certification, and leaves for another day the question of whether it might be presented to a jury.
As a closing point, Google says that class members may be worse off in plaintiffs' but-for world because Google may have to change its current practices to stay competitive by cutting back on services it currently offers for free. In Google's view, “in a world without existing Android security standards, security-conscious consumers would be worse off because they would face costs to keep their data and devices secure.” Dkt. No. 273 at 20. Concerns like these are far too speculative and conditional to be a serious barrier to certification. As is the case for all of Google's attacks on Dr. Singer, it may argue the point at trial, see Olean, 31 F.4th at 682 n.31, but it does not erode plaintiffs' showing of common evidence to prove antitrust impact.
c. Damages
Google's damages objections are readily dispatched. Google says that plaintiffs lack a common method of calculating class-wide damages “for the same reasons they lack common proof of impact.” Dkt. No. 273 at 21-22. Plaintiffs have prevailed on that issue, and Google's damages attacks are overruled on the same grounds. The Court adds that it is well-established circuit law that “damage calculations alone cannot defeat certification,” Yokoyama v. Midland Nat'l Life Ins. Co., 594 F.3d 1087, 1094 (9th Cir. 2010), and “the presence of individualized damages cannot, by itself, defeat class certification under Rule 23(b)(3).” Leyva, 716 F.3d at 514. Dr. Singer has stated how his common methodology can be used to drill down further and make calculations at an app-by-app level. This is enough for now.
2. Cartwright Act and UCL Claims
Google did not separately challenge plaintiffs' Rule 23 showing for the Cartwright Act and UCL claims brought under California law. Plaintiffs also did not separately discuss their claims in the class certification briefing. Even so, the Court has independently examined the state law claims to determine the propriety of certification under Rule 23.
To start, plaintiffs say that the proposed multi-state class may bring claims under California law because “Google has consistently included a choice-of-law provision in its user agreements designating California law as controlling in litigation brought by users.” Dkt. No. 251 at 3. Google does not disagree with this. See Dkt. No. 273. In effect, both sides agree that the California statutes should be applied beyond the state's borders.
Although the parties did not brief the question in any detail, the Court will provisionally apply California state law to the multi-state class. “Subject to constitutional limitations and the forum state's choice-of-law rules, a court adjudicating a multistate class action is free to apply the substantive law of a single state to the entire class.” In re Hyundai, 926 F.3d at 561 (citing, among other cases, Phillips Petroleum Co. v. Shutts, 472 U.S. 797, 823 (1985)). For the Cartwright Act specifically, “the perpetration of anticompetitive activities within California ‘creat[es] state interests' in applying California law to that conduct.” AT & T Mobility LLC v. AU Optronics Corp., 707 F.3d 1106, 1112 (9th Cir. 2013) (citation omitted). The UCL also applies to claims by the multi-state class for anticompetitive conduct by Google that occurred within California. See Norwest Mortgage, Inc. v. Superior Court, 72 Cal.App.4th 214, 224-25 (1999).
Consequently, the Court will apply California law to the multi-state class, as the parties propose. This application is provisional in the sense that, as in In re Hyundai, “no party argued that California's choice-of-law rules should not apply to this class . . . arising from an MDL in a California court,” but an objector or other interested person may do so later in the case. In re Hyundai, 926 F.3d at 561. If so, further discussion of the issue may be warranted. Id.; see also Patel v. Facebook, Inc., 932 F.3d 1264, 1276 (9th Cir. 2019) (“of course, if future decisions or circumstances lead to the conclusion that extraterritoriality must be evaluated on an individual basis, the district court can decertify the class.”).
With the choice of law decided for now, the Court has no trouble concluding that plaintiffs have demonstrated commonality and predominance for the Cartwright Act and UCL claims. This conclusion flows readily from the discussion of these elements with respect to the Sherman Act claims.
C. Typicality and Adequacy (23(a)(3), (4))
Google has made a rather rare challenge to the adequacy of class counsel. Dkt. No. 273 at 22-23. The challenge is made because the Joint Prosecution Agreement divvied up the representation of the various states between the consumer plaintiffs and the State Attorneys General. Google says that class counsel “continue to have a financial interest in the claims of consumers they no longer seek to represent,” and so the JPA “creates improper incentives and dual loyalties that conflict with counsel's duty to the proposed class.” Id.
Why this might be so is not clear. Even assuming some continued “loyalty” to former class members would arise, which is hardly a self-evident proposition, the nature of the ostensible conflict with counsel's duty to the classes is not explained. If anything, the government and private actors are marching arm-in-arm toward a common victory over Google. Their claims are just about the same, the evidence will be the same, and all indications in the record are that their interests are well-aligned. The possibility that “under the JPA, counsel's ability to keep earning fees depends on the absence of a settlement providing a recovery for the proposed class,” id., may be true, but it is irrelevant. Under the lodestar approach to awarding class counsel's fees, counsel will always be able to ask for a higher lodestar the longer a case is litigated prior to settlement. This alone does not make class counsel inadequate under Rule 23(a). In addition, the Court has broad discretion over fee awards, and ample authority to investigate and respond to any concerns.
For the four named plaintiffs who no longer reside in a state within the adjusted class definition, plaintiffs have a problem. A class representative must be a member of the class. See Dukes, 564 U.S. at 348. For that reason, the Court declines plaintiffs' “proposed corrected order,” Dkt. No. 304, which attempts to make these plaintiffs a part of the class by expressly naming them, and terminates the parties' related filings. Dkt. Nos. 306, 311. Plaintiffs Mary Carr (Washington resident), Daniel Egerter (California resident), Zack Palmer (Massachusetts resident), and Serina Moglia (New York resident) are not included in the proposed (b)(3) class as it is currently defined, and they are consequently not adequate representatives under Rule 23(a)(4).
By all appearances, interim class counsel have been litigating this case diligently and competently. The Court finds them to be adequate and will confirm their appointment. Plaintiffs Matthew Atkinson and Alex Iwamoto are typical and adequate, and will be appointed class representatives.
D. Superiority of Class Adjudication (23)(b)(3))
The last remaining factor for class certification is superiority under Rule 23(b)(3). Google does not contest this factor. “In adding ‘predominance' and ‘superiority' to the qualification-for-certification list, the Advisory Committee sought to cover cases ‘in which a class action would achieve economies of time, effort, and expense, and promote . . . uniformity of decision as to persons similarly situated, without sacrificing procedural fairness or bringing about other undesirable results.'” Amchem, 521 U.S. at 615.
As discussed, there is a substantial body of common evidence that could be presented by the proposed class on the elements of their antitrust and consumer protection claims. Class members' individual damages are likely to be very small, and certainly not large enough to make it worth their while to bring individual lawsuits in the absence of this class action. Although some attention will need to be given to fine-tune how the individualized questions on impact and damages can be managed at trial, a class action is superior to individual proceedings in this consolidated action.
IV. THE RULE 23(B)(2) CLASS
Plaintiffs have also asked for certification of an injunctive relief class under Rule 23(b)(2). For this proposed class, Google says that certification cannot be granted because the proposed injunction is insufficiently described and supported, and plaintiffs have not shown that their request for injunctive relief predominates over the monetary relief they are seeking. Dkt. No. 273 at 23-25.
“Class certification under Rule 23(b)(2) is appropriate only where the primary relief sought is declaratory or injunctive.” Ellis, 657 F.3d at 986 (quoting Zinser, 253 F.3d at 1195). That is not the case with consumer plaintiffs' requests for relief here -- the primary relief they seek is without a doubt monetary. Consequently, certification of a (b)(2) class is denied. There is the possibility of an injunction ancillary to an award of damages, which may be considered later in the case as warranted by developments.
CONCLUSION
The following class is certified under FRCP Rule 23(b)(3):
All persons in the following U.S. states and territories:
Alabama, Georgia, Hawaii, Illinois, Kansas, Maine, Michigan, Ohio, Pennsylvania, South Carolina, Wisconsin, Wyoming, American Samoa, Guam, Northern Mariana Islands, Puerto Rico, and the U.S. Virgin Islands
who paid for an app through the Google Play Store or paid for in-app digital content (including subscriptions or ad-free versions of apps) through Google Play Billing on or after August 16, 2016, to the present.
Plaintiffs Matthew Atkinson and Alex Iwamoto are appointed named representatives, and interim class counsel, see Consumer Dkt. No. 78, are confirmed as class counsel under Rule 23(g).
The parties are directed to jointly file by January 20, 2023, a proposed plan to give notice to the certified class and an opportunity to opt out.
Certification under Rule 23(b)(2) is denied, as is the requested exclusion of Dr. Singer's testimony under Federal Rule of Evidence 702.
IT IS SO ORDERED.