ICLE Amicus in Carr v Google
The International Center for Law & Economics (“ICLE”) is a nonprofit, nonpartisan, global research and policy center committed to developing the intellectual foundations for sensible, economically grounded policy. ICLE promotes the use of law and economics methodologies, and economic findings, to inform public policy, and has longstanding expertise in antitrust law.
ICLE has an interest in ensuring that antitrust law promotes the public interest and consumer welfare by remaining grounded in sensible rules informed by sound economic analysis. This includes ensuring that courts and agencies correctly apply the standards for class certification in antitrust cases involving two-sided transaction platforms.
Amicus is authorized to file this brief by Fed. R. App. P. 29(a)(2) because all parties have consented to its filing.
Amicus hereby states that no party’s counsel authored this brief in whole or in part; that no party or party’s counsel contributed money that was intended to fund the preparation or submission of the brief; and that no person other than amicus or its counsel contributed money that was intended to fund the preparation or submission of the brief.
The aim of the federal antitrust laws is to protect competition and the innovation and value-creation that it fosters. To that end, a wide range of entities are authorized to prosecute antitrust claims provided they meet the standards that the legislature and courts have articulated to prevent over-enforcement, which itself can mute competitive incentives and dampen innovation. Those standards include the prerequisites of standing and, in the case of class actions, commonality and predominance.
In this case, plaintiffs have asked to pursue antitrust claims against Google, a 26-year-old company—younger than the average PhD student—that has transformed the way consumers interact with the internet, increased efficiency, and created immeasurable value. Plaintiffs assert that “Google illegally monopolized the Android app distribution market”—which the District Court characterized as a two-sided platform—“with anticompetitive practices in the Google Play Store.” 1-ER-3, 10. They have further asked to proceed as a class, representing 21 million consumers who entered into distinct transactions with various developers related to (in aggregate) 300,000 apps. Defs.-Appellants’ Br. (“Br.”) at 20.
The value of a two-sided platform to both consumers and sellers depends on the platform successfully balancing their relative interests, demands, and capacity. That distinctive aspect of two-sided platforms makes questions about standing (such as injury) and the predominance of common issues particularly complex.
On November 28, 2022, the District Court granted class certification, relying on Illinois Brick v. Illinois, 431 US. 720 (1977), and Apple v. Pepper, 139 S. Ct. 1514 (2019), for the proposition that all 21 million consumers were “direct purchasers” and therefore a purported overcharge to the developer must necessarily also injure them. 1-ER-18–19. It further found a “pass-through” formula was sufficient “class-wide proof of antitrust impact and injury.” 1-ER-19. In reaching this conclusion, the District Court failed to engage meaningfully and rigorously with the economic realities of two-sided platforms.
In its landmark Apple v. Pepper decision, the Supreme Court held that consumers could be considered “direct purchasers” of a two-sided platform even though developers set the retail prices for apps. 139 S. Ct. at 1520 (“It is undisputed that the iPhone owners bought the apps directly from Apple. Therefore, under Illinois Brick, the iPhone owners were direct purchasers who may sue Apple for alleged monopolization.”). In other words, such consumers were not so remote from the platform as to be barred from bringing suit for lack of standing. But status as a “direct purchaser” does not categorically mean that plaintiffs have met the requirements of Article III standing. “Direct purchasers” from two-sided platforms must still show that, as compared to a but-for-world platform using a different price structure, they have been adversely affected.
The Supreme Court in Ohio v. American Express emphasized this point. Noting first that “[l]egal presumptions that rest on formalistic distinctions rather than actual market realities,” are “generally disfavored in antitrust law” (Ohio v. Am. Express Co., 138 S. Ct. 2274, 2285 (2018) (quoting Eastman Kodak Co. v. Image Technical Servs., Inc., 504 U.S. 451, 466–67 (1992)), it found that “[P]rice increases on one side of the platform . . . do not suggest anticompetitive effects without some evidence that they have increased the overall cost of the platform’s services.”
Id. at 2287. In other words, given the unique dynamics of two-sided platforms, the existence and quanta of injury to consumers cannot be assessed without considering whether a “price” change—including quality parameters such as convenience and security—feeds through the indirect network effects running from one side of the platform to another to alter the relative value of the platform to users and then, assuming that it does, determining which (if any) users are adversely affected.
Plaintiffs skipped over this complexity, instead relying on inapplicable supply-and-demand assumptions and focusing on the pass-through of a single price term. A suit by consumers against a platform based on products priced by developers necessarily invokes a much more complicated interaction:
Apple’s overall pricing structure includes a component whereby Apple allows app developers to set app prices, but it also includes relevant prices and terms set by Apple, including: the price of iOS devices; the commission charged to app developers; the price of its own iOS apps; the (unavoidable) ability for app developers to charge for services outside of the iOS ecosystem (without paying the thirty percent commission); and the structure, price, and availability of app marketing in the App Store.
It is not enough to show that developers are charged higher fees; or even that developers could charge higher prices to app consumers. Instead, Plaintiffs must assess both the price effects and non-price quality dimensions.
As things stand, however, plaintiffs focus solely on the costs of Google’s behavior—i.e., the purportedly higher Play Store fees—ignoring how it contributes to making the Android ecosystem safer and of higher quality (both of which boost participation on both sides of the market). Put differently, plaintiffs’ mistake is to narrowly focus on the effect that Google’s behavior has on Play Store fees, while ignoring how it benefits the broader Android ecosystem and how it may affect individual participants.
In short, plaintiffs have not met their burden of showing that common issues of injury will predominate over individualized inquiries; among other things, they have not offered an analysis rigorous enough to determine whether any members of the putative class were injured, let alone the majority of them. Accordingly, the District Court’s class certification order should be reversed.
A plaintiff must establish standing to bring a lawsuit in federal court. Spokeo, Inc. v. Robins, 578 U.S. 330 (2016). The minimum of Article III standing requires that the plaintiff has “(1) suffered an injury in fact, (2) that is fairly traceable to the challenged conduct of the defendant, and (3) that is likely to be redressed by a favorable judicial decision.” Id. The standing requirement applies just as equally to a “direct purchaser” plaintiff as it does to any other plaintiff. Where the “direct purchaser” is engaging with a two-sided platform—which is how the District Court treated plaintiffs and the purported Android App Distribution market in this matter—questions about injury and the traceability of such can be particularly complex.
Here, however, the District Court failed to rigorously engage with that complexity, suggesting instead that because it had classified consumers as “direct purchasers”, they per force suffered an injury from the service fees that Google charged to developers:
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.
1-ER-23. Reasoning that “[t]he overcharge has not been passed on by anyone to anyone” and “consumer plaintiffs paid the alleged overcharge directly to Google,” the District Court concluded that “there is no barrier to certification on this score.” 1-ER-18–19.
But, as discussed further in Section II below, just because a consumer is a “direct purchaser” on one side of a two-sided platform does not mean that the consumer paid any share of a purported overcharge. Rather, the relevant question is whether the retail prices that plaintiffs paid for the apps and associated services—prices that are set by the developers—were higher than they would have been in the but-for world, accounting for quality parameters such as safety and convenience. The Supreme Court’s decision in Apple v. Pepper is instructive on this point. There, the Supreme Court found that the absence of an intermediary between Apple’s app store and its consumers was sufficient to find the latter to be “direct purchasers” in the “retail market for the sale of apps.” 139 S. Ct. at 1521. In other words, under the framework articulated by the Supreme Court in Illinois Brick Co. v. Illinois, 431 US. 720 (1977), such consumers were not so remote from the purported antitrust violation as to be barred from suit. In reaching this decision, the Court rejected the notion that the directness of the purchasing relationship should be defined by who sets the price, i.e., consumers can be the “direct purchasers” of a firm that does not set the retail price of a good or service. Id. at 1522.
However, the Court was also crystal clear that the directness of the payment relationship did not mean that the consumer had in fact paid an overcharge, explicitly distinguishing between app store commissions (or fees) and the app prices paid by consumers:
To be sure, if the monopolistic retailer’s conduct has not caused the consumer to pay a higher-than-competitive price, then the plaintiff’s damages will be zero. Here, for example, if the competitive commission rate were 10 percent rather than 30 percent but Apple could prove that app developers in a 10 percent commission system would always set a higher price such that consumers would pay the same retail price regardless of whether Apple’s commission was 10 percent or 30 percent, then the consumers’ damages would presumably be zero.
139 S. Ct. at 1523. That is, the Court declined to assume that higher fees charged to app developers necessarily were passed on to consumers through the retail price set by the developer. The Court in Apple v. Pepper understood that the effects of a hypothetical anticompetitive fee increase may be shared between developers and consumers, with each group needing to show how it was affected to prove their claims. In the case of consumers, this means showing that the increased fees charged to developers resulted in a worse outcome for consumers.
Apple v. Pepper thus invites a “pass on” inquiry (though fee incidence may actually be the more appropriate terminology), including whether a fee increase has negatively affected a plaintiff, which in turn depends on whether developers have passed on the increased fees to users and whether consumer benefits derived from platform fees (e.g., increased security) outweigh those costs. Indeed, that is the direct and anticipated consequence of the Court’s decision in Apple v. Pepper. In its ruling, the Court intentionally opened the door to such allocation questions:
It is true that Apple’s alleged anticompetitive conduct may leave Apple subject to multiple suits by different plaintiffs. . . .The consumers seek damages based on the difference between the price they paid and the competitive price. The app developers would seek lost profits that they could have earned in a competitive retail market.
139 S. Ct. at 1525. See also id. at 1526 (Gorsuch, J. dissenting) (“Illinois Brick held that these convoluted ‘pass on’ theories of damages violate traditional principles of proximate causation and that the right plaintiff to bring suit is the one on whom the overcharge immediately and surely fell. Yet today the Court lets a pass-on case proceed.”).
In short, Apple v. Pepper does not create a presumption that potentially anticompetitive fee increases always injure direct purchasers in the platform context. Consequently, it was incumbent on the District Court to analyze rigorously whether plaintiffs could demonstrate that purported supracompetitive fees charged to developers caused changes to the value of the platform that injured consumers and that common questions regarding such injury predominate over individualized inquiries. See Bowerman v. Field Asset Services, Inc., 60 F.4th 459, 469 (9th Cir. 2023) (explaining that “class certification is inappropriate ‘when individualized questions . . . will overwhelm common ones,” and decertifying class based on predominance of individual questions over common ones.). As discussed below, the District Court erred by failing to grapple with the economic reality of two-sided platforms.
II. Two-sided Platform “Direct Purchasers” Cannot Demonstrate Injury, and Therefore Standing, Without Adequately Accounting for Indirect Network Effects
It is well recognized that “antitrust law should look at the ‘the economic reality of the relevant transactions’ rather than ‘formal conceptions of contract law.’” ((Apple, Inc. v. Pepper, 139 S. Ct. at 1529–30 (Gorsuch, J. dissenting) (cleaned up)).  Here, that means that plaintiffs cannot demonstrate injury without accounting for the dynamics of a two-sided platform, i.e., how altering the fees charged to developers may trigger feedback effects between the two sides of the platform to change the balance of features and ultimately affect consumers:
[T]here is no meaningful economic relationship between benefits and costs on each side of the market considered alone . . ., any analysis of social welfare must account for the pricing level, the pricing structure, and the feasible alternatives for getting all sides on board.
Plaintiffs’ narrow focus on pass-through of a single price term (or as the District Court termed it, “share of the overcharge,” 1-ER-19) fell short of this analysis, particularly under the standard required by Rule 23.
A two-sided platform is a business model that creates value by reducing the transaction costs of direct interactions between two or more types of users in ways that mere resellers cannot replicate. A critical feature of multi-sided platforms is that the demand of platform participants is interdependent—the extent of participation by one set of users on a platform depends on the participation of one or more other sets of users. A multi-sided platform uses both pricing and design choices to achieve critical mass. Without critical mass on all sides, the positive feedback effect, which enables the platform’s unique matching abilities, cannot be achieved. Further, interdependent demand on platforms often leads to situations where efficient pricing may involve below marginal cost pricing on one side and above marginal cost pricing on another. As a result, inferences drawn from the traditional indicia of competition—price and output effects—may be inapposite, particularly when they are assessed on only one side of a multi-sided market or without consideration of the effects on the design of the platform itself.
It is a well-accepted proposition in the literature that vertical restraints on multi-sided platforms can be procompetitive, anticompetitive, or competitively neutral depending on a host of complicated interactions among the various groups of platform users and between users and the platform itself. Procompetitive vertical restraints on multi-sided platforms may fall into one or more of at least three broad categories: (1) achieving economies of scale that provide benefits to consumers overall; (2) helping platforms deal with coordination and expectation problems to the benefit of platform users; and (3) providing benefits to one side of the platform that increase consumer welfare overall. These procompetitive effects are a function of the particular structure of such two-sided markets and necessitate adjustments to antitrust doctrine to ensure that presumptions and evidentiary burdens properly reflect the more complicated economic relationships among the parties involved.
Plaintiffs and their experts failed to fully account for these dynamics in their analysis. For example, Dr. Singer mistakenly surmises that lower fees would result in more consumer demand because:
A foundational principle in economics is that “demand curves” are downward sloping—meaning that, all else equal, consumers will demand more of a product or service the lower its price. How much more will be demanded depends on the consumer elasticity of the demand response to lower prices for Apps and In-App Content.
Dkt. 252-3 at 125. Similarly, he concludes that the supply of apps and in-app content would increase if Google’s service fees were reduced and the developers received more revenue: “Absent the Challenged Conduct, developers would realize larger proceeds, which would bring forward more App and In-App Content development, commensurate with a shifting out of the supply curve.” Id. And he makes these claims while explicitly excluding considerations of quality effects. Dkt. 252-3 at 74 n. 368 (“Although my primary impact focuses on price effects (over the take rate) it is possible that competition would occur on non-price quality dimensions as well.”). Moreover, when he does acknowledge that competition can occur on the quality dimension, he neglects entirely the possibility of quality (design) changes in the platform itself. Id. at 108.
This approach is wrong for two important reasons. The first is that lower service fees to the developer do not necessarily translate into lower retail prices to the consumer, as the real-world evidence demonstrated. Br. at 39–41. More fundamentally, the intuition that lower prices result in higher demand does not always hold in two-sided markets. For example, if the fee structure encourages participation from developers on the other side of the market, then that may lead to higher user demand. Likewise, if higher developer fees or a particular fee structure facilitate platform design choices that improve quality for users (and/or developers), that, too, may stimulate user demand. In short, the simple reallocation of costs and benefits across the sides of a two-sided market can be output increasing, output reducing, or output neutral. Looking solely at price effects simply cannot distinguish between these scenarios.
This is why the Supreme Court in Ohio v. American Express emphasized that the legal analysis of injury in a two-sided market requires consideration of both sides of the platform: “Price increases on one side of the platform likewise do not suggest anticompetitive effects without some evidence that they have increased the overall cost of the platform’s services.” 138 S. Ct. at 2286 (internal citation omitted). “Competition,” it declared, “cannot be accurately assessed by looking at only one side of the platform in isolation.” Id. at 2287.
The notion that prices charged on one side have an ambiguous effect on demand on the other is one of the central findings of the economic literature regarding two-sided markets, because the “quality” of the final product is intrinsically connected to the other side of the market. In their highly influential work, Jean-Charles Rochet and Nobel prize winner Jean Tirole observed that:
We define a two-sided market as one in which the volume of transactions between end-users depends on the structure and not only on the overall level of the fees charged by the platform.
They also highlight that the basic, “canonical” model they develop is an important starting place to understand the relationships between sides of a two-sided market, but it is not a description of reality, and a proper analysis must go beyond narrow price considerations. As a result, “policy interventions to alter the price structure (as opposed to the price level)” are not likely to be “solidly grounded.” Indeed, as Andre Hagiu has shown:
[I]n all of these articles, the volume of transactions . . . is not directly affected by platforms’ prices: Rochet and Tirole and Armstrong essentially assume that each member of one side interacts with an exogenously given proportion of members on the other side. . . . In my model the variable fees charged by the platform (royalties) play a central role, because they affect the prices and volumes of trade between sellers and buyers and therefore social welfare. On the other hand, the allocation of the royalties among the two sides is largely inconsequential in my model.
The “canonical” assumption of a fixed quantity and ratio of users is inapt where, as here, users decide whether to join the platform at different times and consume its services in varying amounts. Any assessment of the effects of a revised price structure would require a complete analysis across relevant metrics of the optimal balancing of demand on both sides of the market under the new structure.
But here, Plaintiffs and the District Court discounted non-fee metrics:
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.’ Concerns like these are far too speculative and conditional to be a serious barrier to certification.
1-ER-25 (internal citations omitted). This conclusion ignores the fact that, in two-sided platforms, what might superficially appear to be a fee increase on one side may in fact be a crucial component of the underlying ecosystem:
Where a single, two-sided product is at issue, the price may be spread across users on both sides of the market. Moreover, non-price product characteristics will necessarily differ between different sets of users. . . Given the differential incidence of price and quality across a platform, it is impossible to capture the competitive dynamics and to measure the competitive effects by viewing only the partial price on one side.
It is thus impossible to assess whether a particular participant on the platform has been injured without considering how the change affects the platform as a whole and, to the extent that there is an anticompetitive effect, which side (or sides) and which participants within that side (or sides) are worse off.
In sum, the District Court was wrong to conclude plaintiffs could assert injury based on a purported overcharge in service fees to one side of the platform without fully analyzing the price and non-price feedback between the two sides of the platform.
III. Injury Cannot Be Assumed from Anticompetitive Harm to a Subset of Heterogenous Two-Sided Platform “Direct Purchasers”
Plaintiffs, and ultimately the District Court, further err by assuming that the complicated effects of a change to the platform is directionally the same for all consumers.
The first problem is (as discussed above) that antitrust injury cannot be inferred from consideration of only a partial price change, particularly in the context of a platform relationship. Because users are heterogenous across many dimensions, the assumption of anticompetitive effect (let alone commonality across users) from a price change for a subset of consumers is particularly hard to maintain.
This is not just a consequence of the two-sided nature of the market at issue. Consider, for example, the basic principle that harm to a particular competitor—even the loss of a particular competitor—is not the same thing as harm to competition. See Brown Shoe Co. v. United States, 370 U.S. 294, 320 (1962). One necessary implication of this principle is that it is not enough to show harm to “inframarginal” consumers or to a particularly sensitive subset of consumers, because, for some of these consumers—say, those with strong brand loyalty to a particular firm—the loss of their preferred competitor would also harm them, even if competition itself were not affected. Holding otherwise would undermine this fundamental limitation on the scope of antitrust injury. Because this distinction between harm to some consumers and anticompetitive harm is so crucial to ferreting out conduct that creates or maintains monopoly power, antitrust law requires a demonstration that the conduct at issue has an actual anticompetitive effect—not merely that it results in some harm—before a case may be brought.
This is a fundamental tenet of antitrust law and economics, but it is particularly magnified in the platform setting. There, because operators are optimizing the platform for the benefit of multiple groups of users on multiple sides of a multi-sided market, superficial harm to one group may well be part of an overall optimization strategy, and injury cannot be assumed on the basis of harm to a subset of users. Most obviously, apps and app prices are embedded in a broader ecosystem and are instrumental to its value—as Plaintiffs’ expert even understands (without acknowledging its significance):
The functionality and user enjoyment derived from a mobile device is highly dependent upon the range and quality of apps available on it. In addition to producing a mobile operating system, Google has created a distribution channel for delivery of Android-compatible apps developed by third parties, and developed its own universe of Apps, for Google Android. Google itself has developed some of the most popular Android- and iOS-compatible apps, including Google Search, Google Maps, Chrome, YouTube, and Gmail.
Dkt. 252-3, at 10.
Indeed, restrictions in one dimension may not even constitute “harm” to that subset of consumers where it is accompanied by corresponding benefits, for instance where the restrictions serve to protect consumers’ privacy and data or to assure adequate monetization and distribution. Platform users can benefit from features that contribute to the overall success of the ecosystem. In other words, higher prices may reflect higher quality—stemming directly from the very conduct that Plaintiffs claim is the source of injury.
Second, even if one ignores the concept of virtuous feedback loops, Plaintiffs’ claim that all class members are negatively affected still falls flat. Consumers are not similarly situated, and it is inappropriate to assume a simple pro rata effect from a change in app store prices. Game apps account for an enormous percentage of app store purchases, and “spending on the consumer side is also primarily concentrated on a narrow subset of consumers: namely, exorbitantly high spending gamers. . . . 81.4% of all Apple accounts spent nothing and account for zero percent of the App Store billings for the quarter. . . , [and] 6% of App Store gaming customers in 2017 accounted for 88% of all App Store game billings. . . .” Epic Games, Inc. v. Apple Inc., 559 F. Supp. 3d 898, 953–54 (N.D. Cal. 2021). Without analyzing whether and how the reduction in fees charged to developers affects other costs associated with the platform—either direct costs or indirect costs through reduced quality—it cannot be assumed that even those consumers who make purchases in the Play Store are worse off. Moreover, because user groups vary along other dimensions that are likely to correlate with price—most notably in terms of their sensitivity to privacy and security risks and their tolerance of user-interface impediments like ads or “choice screens”—the effects of price reductions or increases accompanied by other qualitative changes are not captured by price, vary considerably across users, and are not likely to correlate with pro rata app usage. Any inquiry into injury will turn on these individualized questions.
For the reasons stated above, the District Court’s class certification order should be reversed.
 In a two-sided market a firm sells two different products or services to two different groups of consumers. See Filistrucchi et al., Market Definition in Two-Sided Markets: Theory and Practice, 10 J. Competition L. & Econ. 293 (2014) (“In particular, competition authorities have failed to recognize the crucial difference between two-sided transaction and non-transaction markets….”).
 Bruno Jullien, Alessandro Pavan, & Mark Rysman, Two-sided Markets, Pricing, and Network Effects, 4 Handbook of Industrial Organization 488 (2021) (“[I]ndirect network effects . . . emerge when the adoption and use of a product leads to increased provision of complementary products and services, with the value of adopting the original product increasing with the provision of such complementary goods . . . Indirect network effects thus lead to a feedback loop as more participants on each side of the platform find it more valuable to adopt and use the platform when they expect the other side to attract more users.”)
 Geoffrey A. Manne and Kristian Stout, The Evolution of Antitrust Doctrine After Ohio v. Amex and the Apple v. Pepper Decision That Should Have Been, 98 Neb. L. Rev. 425, 458 (2019).
 We understand that there are summary judgment motions pending regarding whether plaintiffs have properly defined a market involving the sale of in-app purchases (IAPs) and subscriptions to consumers because: (1) “Plaintiffs’ experts have conceded that their alleged Android app distribution market does not involve IAPs or subscriptions at all”; and (2) “according to plaintiffs’ experts,” the other putative relevant market involves Google selling in-app billing services to developers. No. 3:21-md-02981-JD, Dkt. 480, at 18. We do not take any position on these issues.
 To draw an analogy with physical retail, the harm to consumers that buy from a retail cartel is the difference between the competitive and cartel price of the goods they purchase, not the increase to those retailers’ margins under monopoly. This distinction has important ramifications for two-sided markets and the case at hand.
 Manne & Stout, supra note 3, at 458 (referencing both Ohio v. Amex and Apple v. Pepper and emphasizing “the need for the plaintiffs . . . to allege injury and present their prima facie case consistently with the economic realities of the two-sided market at issue”).
 David S. Evans, The Antitrust Economics of Multi-Sided Platform Markets, 20 Yale J. Reg. 325, 355–56 (2003).
 For a more thorough discussion of the nature of two-sided platforms, see generally David S. Evans & Richard Schmalensee, The Antitrust Analysis of Multisided Platform Businesses, Oxford Handbook On International Antitrust Economics (Roger Blair & Daniel Sokol eds., 2013).
 See, e.g., David S. Evans, Economics of Vertical Restraints for Multi-Sided Platforms (Univ. of Chi. Inst. for Law & Econ. Olin Res. Paper No. 626, Jan. 2, 2013), http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2195778 [https://perma.unl.edu/T5CP-QECW], at 4.
 See Evans & Schmalensee, supra note 8, at 6.
 See Evans, supra note 9, at 8–10.
 Unless otherwise noted, references to “Dkt” refer to No. 3:21-md-02981-JD.
 Jean-Charles Rochet & Jean Tirole, Two-sided Markets: A Progress Report, 37 The RAND J. of Econ. 646 (2010).
 Id. at 663.
 Jean-Charles Rochet & Jean Tirole, Platform Competition in Two-Sided Markets, 1 J. of the Eur. Econ. Ass’n 1009 (2003).
 Andre Hagiu, Pricing and Commitment by Two-Sided Platforms, 37 RAND J. of Econ. 720, 722 (2006)
 See, e.g., Benjamin Klein, et al., Competition in Two-Sided Markets: The Antitrust Economics of Payment Card Interchange Fees, 73 Antitrust L.J. 571, 598 (2006) (“The economic theory of two-sided markets indicates that relative prices on the two sides of the market are independent of the degree of competition faced by a supplier in such a market. While total prices will be influenced by competition, relative prices are determined by optimal balancing of demand on the two sides of the market.”).
 Geoffrey A. Manne, In Defence of the Supreme Court’s “Single Market” Definition in Ohio v. American Express, 7 J. Antitrust Enf. 104, 109 (2019).
 See Dirk Auer, Appropriability and the European Commission’s Android Investigation, 23 Colum. J. Eur. L. 647 (2017).
 As the court in Epic v. Apple notes, the Google Play Store appears to have similar characteristics. Epic Games, Inc. v. Apple Inc., 559 F. Supp. 3d at 954 n. 243.