Two Sides to Every Story: Growing Tensions Between Legal Rules and Economic Realities for Platform Industries

By Andrew J. Ewalt (Freshfields Bruckhaus Deringer US LLP)1


In Ohio v. American Express (Amex),2 the Supreme Court introduced into U.S. antitrust law a new concept that had never before appeared in any judicial decision: the two-sided transaction platform. Ordinarily, when two products are sold to different groups of customers and are not substitutes for each other, an antitrust plaintiff can define separate relevant markets for each of those products and prevail with proof that a defendant harmed competition in either of those two markets. But when the two products are linked by a two-sided transaction platform, Amex requires the relevant market to encompass both products and the dismissal of claims based on one-sided market definitions.

Though Amex makes clear the importance of identifying two-sided transaction platforms, the opinion leaves considerable scope for litigation over whether any particular business is properly characterized as such a platform. As courts have grappled with that issue, another question has also come into clearer focus: is the characterization decision one of fact or law? How courts resolve that question will impact not only how individual cases are litigated and resolved, but also how antitrust doctrine evolves in the future.

This article reviews the Amex decision, showing how the opinion lays out the relevant facts for courts to consider in determining whether a business qualifies as a two-sided transaction platform. Next, it summarizes the first cases that have applied Amex and explains that some of them — including most notably the Second Circuit’s decision in US Airways v. Sabre — have wrongly held that a court can decide as a matter of law whether a business is a two-sided transaction platform. The consequences of these legal errors became clear in Sabre/Farelogix, where the court found as a factual matter that the parties compete, but nevertheless held that they cannot compete as a matter of law. The article concludes with a few thoughts on the implications for antitrust litigation and doctrine that depend on whether characterizing a business as a two-sided transaction platform is a question of fact or law.



The Amex case began in October 2010, when the U.S. Department of Justice and a group of states brought a rule-of-reason challenge to the vertical “antisteering provisions” in American Express’s contracts with merchants.3 All parties agreed that credit card networks operate two-sided platforms connecting merchants and cardholders,4 but they disagreed about how antitrust analysis should account for that feature of the industry. The government argued that proof of anticompetitive effects on merchants alone sufficed to carry its initial burden under the rule of reason and shifted to American Express the burden of justifying the antisteering provisions. American Express, on the other hand, contended that the government must show that the provisions had an overall net anticompetitive effect on both merchants and cardholders. The district court sided with the government,5 but the Second Circuit and the Supreme Court concluded that the district court had erred by not defining a single relevant market that included both merchants and cardholders.6

The Supreme Court held that “only one market should be defined” because credit card networks operate two-sided transaction platforms.7 Before describing that particular type of platform, however, Justice Thomas’s majority opinion identified three characteristics shared by all two-sided platforms. First, “a two-sided platform offers different products or services to two different groups who both depend on the platform to intermediate between them.”8 Second, two-sided platforms exhibit “indirect network effects” — that is, “the value of the two-sided platform to one group of participants depends on how many members of a different group participate.”9 Third, “[s]triking the optimal balance of the prices charged on each side of the platform is essential for two-sided platforms to maximize the value of their services.”10

The Court recognized that “it is not always necessary to consider both sides of a two-sided platform” and required definition of a single relevant market encompassing both sides only for a “special type of two-sided platform known as a ‘transaction’ platform.”11 The “key feature of transaction platforms is that they cannot make a sale to one side of the platform without simultaneously making a sale to the other.”12 In addition, “[o]nly other two-sided platforms can compete with a two-sided platform for transactions.”13 By contrast, a “market should be treated as one sided when the impacts of indirect network effects and relative pricing in that market are minor.”14 For example, “[b]ecause of . . . weak indirect network effects, the market for newspaper advertising behaves much like a one-sided market and should be analyzed as such.”15

Without saying so explicitly, the Court’s reasoning implies that characterizing a business as an ordinary two-sided platform (where the two sides may be analyzed separately) or a two-sided transaction platform (where the two sides must be analyzed together) raises issues of fact. Amex held that accurate characterization depends on the strength of indirect network effects, but that issue cannot be resolved in a vacuum. Evidence of industry conditions is required to illuminate whether indirect network effects are relatively stronger or weaker in any particular context. If lower courts are to remain faithful to Amex, they should neither ignore that type of evidence nor resolve characterization disputes as a matter of law.

In dissent, Justice Breyer criticized the majority for failing to explain why the products sold to customers on each side of a two-sided transaction platform must be combined into a single relevant market.16 Based on the majority’s reasoning, Justice Breyer identified “four relevant features” that define two-sided transaction platforms: “they (1) offer different products or services, (2) to different groups of customers, (3) whom the ‘platform’ connects, (4) in simultaneous transactions.”17 He then argued that, under that definition, farmers’ markets, travel agents, and internet retailers would all qualify as two-sided transaction platforms.18 Justice Breyer’s clear implication was that Amex’s holding could extend from the credit card industry to many other parts of the economy, but his four-part test overlooked a critical feature that the majority relied upon to cabin its ruling — strong indirect network effects.


NCAA Grant-In-Aid Caps

The first post-Amex decision to grapple with whether another industry involved a two-sided transaction platform was In re NCAA Grant-In-Aid Cap Antitrust Litigation.19 There, a class of Division I football and basketball players challenged NCAA rules limiting the compensation that they could receive.20 Before Amex was decided, both the plaintiffs and the defendants had asked the district court to adopt a relevant “‘college education market’ . . . wherein colleges compete for the services of athletic recruits by offering them scholarships and various amenities,”21 which was the same one-sided market that had been defined in an earlier case challenging different NCAA rules.22 The court accepted the agreed market definition and granted summary judgment on that issue.23 With no live dispute about how to define the relevant market, the court also excluded the opinion of the defendants’ expert, Dr. Kenneth G. Elzinga, about colleges being multi-sided platforms because it found that those opinions were “not relevant to any of the issues remaining for trial.”24

When Amex came down two months later, the district court invited both sides to argue whether the Supreme Court’s decision ought to affect the earlier rulings.25 After considering those arguments, the district court still refused to admit Dr. Elzinga’s opinion that the relevant market was “‘a multi-sided market for college education in the United States’ in which colleges operate as multi-sided platforms that balance their pricing to numerous constituencies.”26 This time, however, the court grounded that conclusion in analysis of the factual support for Dr. Elzinga’s opinion. For instance, it found that, unlike in Amex, Dr. Elzinga did not describe a “simultaneous interaction or proportional consumption through a platform by different market participants of what essentially constitutes ‘only one product.’”27 The court also emphasized that Dr. Elzinga “d[id] not perform any economic analysis to support his assertions” about purported relationships between the “various sides of his multi-sided platform” and “d[id] not examine any economic data at all to quantify, test, evaluate, or confirm any of the economic relationships upon which his proposed multi-sided relevant market is predicated.”28

Despite ruling on market definition and the admissibility of expert testimony as a matter of law, the district court’s entire analysis turned on the facts in the record before it and whether those facts suggested that college sports were similar enough to credit card networks for Amex to apply. The holding was not that college athletics could never be characterized as a two-sided platform, only that it had not been shown to be one in that case. The NCAA Grant-In-Aid case thus illustrates how courts can approach the issue of whether a particular industry involves a two-sided transaction platform as a question of fact, rather than law.


Amex Merchant Litigation

The next case to address that issue involved a challenge brought by merchants to American Express’s same antisteering rules that were at issue in Amex. The merchants litigated for a time in parallel with the government case, but their case was stayed while the government case was tried and appealed. After the Supreme Court decided the government case, and the merchant case resumed in the district court, American Express sought summary judgment on the merchants’ theory that the relevant market was one-sided. According to American Express, the Supreme Court had decided that both sides of credit card networks must be analyzed as part of a single relevant market as a matter of law, and the merchants should not be permitted to proceed on a theory that depended on a relevant market that included only merchants and excluded cardholders.29

The merchants, on the other hand, contended that two types of evidence showed that American Express’s network did not actually exhibit the features that the Supreme Court identified as characterizing two-sided transaction platforms. First, they claimed that “Amex is a ‘mature’ market, thus calling into question whether it exhibits indirect network effects.”
Second, the merchants pointed to “abundant evidence that Amex does not ‘balance’ the prices on the two sides of its platform and that those prices are not ‘relative’ to each other or ‘interconnected.’”

The district judge (whose earlier ruling against American Express in the government case had been reversed by the Supreme Court) considered himself bound by Amex’s “holding” that “‘[c]redit-card networks are two-sided platforms.’”32 The court believed that conclusion was “not a ‘determination of fact’” that could be altered by differences in the record evidence between the government case and the merchant case.33 Indeed, the district court even recognized that the Supreme Court did not “critically engage” with the factual dispute raised by the merchant plaintiffs and had based its holding on an “unchallenged” assumption,34 presumably referring to the government’s concession that credit card networks are two-sided platforms.35 Nevertheless, “[b]ecause the Supreme Court has already answered the exact question presented . . . with respect to the same defendant,”36 the court considered itself “b[ound] . . . as a matter of stare decisis”37 to reject the merchants’ “one-sided market arguments as a matter of law.”38

After the Supreme Court had ruled that American Express operated a two-sided transaction platform, it would have been surprising for any district court to allow a jury to reach the opposite conclusion in the merchant case. It would have been even more extraordinary for the very judge whose decision had been reversed by the Supreme Court to take such a step. Without so directly confronting the Supreme Court, however, the district court suggested that judges retain the flexibility to determine as a matter of fact when other industries involve two-sided transaction platforms, recognizing that Amex’s rule is “still subject to case-by-case application.”39 Accordingly, although the decision in the merchant case clearly states that American Express is a two-sided transaction platform as a matter of law, the holding is limited to that specific context and leaves open the larger issue of whether determining if other industries involve two-sided transaction platforms would raise questions of law or fact.


US Airways v. Sabre

In US Airways, Inc. v. Sabre Holdings Corp.,40 the Second Circuit took a more categorical approach toward global distribution systems (“GDSs”). GDSs are computer systems on which airlines list tickets for sale and travel agents search for and book flights for their customers.41 US Airways alleged that Sabre — the largest GDS in the United States, with a market share over 50 percent — violated Section 1 by including in its contracts with airlines “full content” provisions requiring airlines to make all of their fares available on the Sabre GDS on the same terms as those fares were available elsewhere.42

The case proceeded through five and a half years of pre-trial litigation, until roughly one month before trial began, the Second Circuit issued its opinion in the government case against American Express.43 Relying on that decision, Sabre argued at trial that the relevant market was two-sided — that is, that “analysis . . . of the challenged restraints’ impact on competition in the GDS market was required to include the combination of its impact on both airlines and travel agents.”44 US Airways, however, contended that the Sabre platform was one-sided because it was a “mature market” with few indirect network effects.45

Confronted with that dispute, the district court instructed the jury to reach a primary verdict as to whether the market was one-sided or two-sided and to determine the amount of US Airways’s damages, if any, based on that finding. The court also charged the jury with reaching an alternative damages verdict based on the assumption that the market was two-sided (if its primary verdict was that it was one-sided) or one-sided (if its primary verdict was that it was two-sided).46 The jury found that the market was one-sided and that, regardless of whether it was one-sided or two-sided, US Airways would have incurred the same damages.47

The Second Circuit found error in both the primary verdict and the alternative verdict. According to the court of appeals, the primary verdict could not stand because Sabre exhibited the four features of two-sided transaction platforms that Justice Breyer described in his Amex dissent. As a result, the court concluded that it was “not a jury question” whether such a platform should be analyzed as one-sided or two-sided.48 The “jury must be instructed to consider both sides of the [transaction] platform being evaluated; the relevant market for such platforms must, as a matter of law, always include both sides.”49 The court likewise declined to rely on the alternative verdict because it would “have been impossible for the jury to have followed the district court’s instructions but to have concluded that the compensable damages if the platform were one-sided, as the jury found it to be, were identical in amount to compensable damages if the platform were two-sided, as the jury was required to assume” in reaching the alternative verdict.50

US Airways is inconsistent with Amex. The Second Circuit’s sole authority for striking down the primary verdict came from Justice Breyer’s Amex dissent, and while purporting to summarize the majority’s holding, the dissent failed to identify perhaps the most significant feature distinguishing two-sided transaction platforms from non-platform industries: the presence of strong indirect network effects. The Amex majority emphasized the importance of indirect network effects both in understanding the economics of all two-sided platforms and in distinguishing two-sided transaction platforms.51 But the Second Circuit refused to permit the jury’s verdict to rely on evidence that indirect network effects were weak in the GDS industry. It declared that US Airways’s theory was “wrong as a matter of law in light of Amex,”52 even though the Amex majority stated explicitly that a “market should be treated as one sided when the impacts of indirect network effects . . . are minor.”53



About a month before the Second Circuit heard arguments in US Airways, Sabre agreed to acquire Farelogix.54 Unlike Sabre — which operates a GDS that serves both airlines and travel agents — Farelogix’s only customers are airlines.55 DOJ investigated the merger for nine months and then challenged the transaction in court.

According to DOJ, the merger would substantially lessen competition in a relevant market for “booking services,” which it described as “IT solutions that enable airlines to deliver their offers to travel agencies and to process resulting orders.”56 That market included only some of the services that GDSs provided to airlines and none of the services that they provide to travel agents. In other words, it was a one-sided market. After a full trial on the merits, Judge Leonard P. Stark concluded that DOJ “failed to prove by a preponderance of the evidence that the Sabre-Farelogix transaction is reasonably probable to substantially lessen competition.”57

Although there were multiple bases for the decision, the primary reason, and the one that has attracted the most attention, was that “Sabre and Farelogix do not compete in a relevant market” as a matter of law.58 The court reached that conclusion “[d]ue to the combination” of US Airways and Amex.59 Finding “highly persuasive” US Airways’s holding that “‘the relevant market for [a GDS] platform must as a matter of law include both sides,’” Judge Stark declined to accept DOJ’s one-sided market for booking services.60 He also explained that DOJ could not prove the merger would be likely to reduce to competition in a two-sided market because DOJ conceded that “Farelogix is not a two-sided platform”61 and Amex stated that “‘only other two-sided platforms can compete with a two-sided platform for transactions.’”62

Judge Stark was “not . . . comfortable” resting his decision entirely on the “determination of law that Sabre and Farelogix cannot compete in a relevant market” because that holding conflicted with his factual findings that “as a matter of real-world economic reality . . . Sabre and Farelogix do compete to a certain extent.”63 There was indeed ample evidence showing that Sabre and Farelogix competed,64 and that evidence might have carried the day if DOJ had been able to persuade the court not to follow US Airways. But the distinctions that DOJ offered — US Airways “was directed to a different legal question, was based on a different factual record[, and was] not binding precedent”65 — essentially conceded that the Second Circuit’s analysis was correct. Judge Stark might not have found US Airways quite so persuasive if DOJ had offered the more fundamental critique that the Second Circuit erred in holding that GDSs are two-sided transaction platforms as a matter of law because Amex left such characterization decisions to the finder of fact. For that argument to have made a difference in Sabre/Farelogix, however, DOJ also would have had to prove as a factual matter that indirect network effects across GDS platforms are weak, as the plaintiff in US Airways had claimed to have done.

Alternatively, DOJ could have confronted directly Amex’s statement that “[o]nly other two-sided platforms can compete with a two-sided platform for transactions.”66 There were several grounds for doing so. First, the statement is pure dicta because the Amex decision did not require the Supreme Court to address whether two-sided transaction platforms compete with other businesses. Second, the parties in Amex did not urge the Court to hold that two-sided transaction platforms compete only with other two-sided platforms.67 Third, the Court misread the only support for its statement, an article by Lapo Filistrucchi and others.68 Although that article contended that two-sided transaction platforms should be distinguished from other two-sided platforms, it did not assert that two-sided transaction platforms only compete with other two-sided platforms. To the contrary, it recognized that “[c]andidate substitute products are not only other platforms that offer, to both sides, the possibility to transact, but also non-intermediated transactions, such as . . . a cash payment.”69 In other words, the article clearly indicates that antitrust analysis should take into account how cash competes with credit card networks like American Express, even though cash is not a two-sided platform.

Finally, the facts in Sabre/Farelogix demonstrate that Amex was wrong to say that two-sided transaction platforms only compete with other two-sided platforms. There was considerable evidence that Sabre competed not only with other GDSs, but also with Farelogix “as a matter of real-world economic reality.”70 In addition, even though airlines’ own websites (“”) are not two-sided platforms, Judge Stark found that they should have been included in the relevant market because they compete with GDSs for at least leisure travelers.71 Further, the defendants’ expert, Dr. Kevin Murphy, admitted that “‘GDSs can face competition from one-sided competitors.’”72 All of these facts left Judge Stark feeling “not . . . comfortable” with his “determination of law that Sabre and Farelogix cannot compete in a relevant market.”73

DOJ wasted no time in appealing Judge Stark’s decision to allow Sabre to acquire Farelogix the day after it was announced.74 The next day, however, the UK’s Competition and Markets Authority reached the opposite conclusion and blocked the transaction.75 A few weeks later, Sabre and Farelogix terminated their merger agreement.76

Ordinarily, the decision to abandon a merger would end the litigation. But rather than quietly allow its appeal to become moot, DOJ asked the Third Circuit to vacate Judge Stark’s decision because DOJ would “no longer have an opportunity to argue that the district court’s reading of Amex was mistaken.”77 DOJ was especially concerned that the Sabre/Farelogix “ruling—if not vacated—could have an outsized effect on cases involving competition in the digital economy, where it is not uncommon for multi-sided platforms to face competition from one-sided rivals.”78 Although it granted DOJ’s request to vacate, the Third Circuit pointedly emphasized that it was “express[ing] no opinion on the merits of the parties’ dispute before the District Court” and that its order “should not be construed as detracting from the persuasive force of the District Court’s decision, should courts and litigants find its reasoning persuasive.”79 That disposition fell short of completely undermining Judge Stark’s opinion, as DOJ had sought. In fact, as a technical matter, the Third Circuit’s order did not affect the precedential weight of Sabre/Farelogix at all because no district court decision ever has more than persuasive authority. Still, the vacatur could complicate future litigants’ attempts to rely on Judge Stark’s reasoning and, as a result, limit the opinion’s influence going forward. If a future case provides another opportunity to revisit the merits, it will be interesting to see whether DOJ is then willing to take the position that Amex itself — rather than the “district court’s reading” in Sabre/Farelogix — was wrong to claim that two-sided transaction platforms only compete with other two-sided platforms.


Implications for Future Cases

Ongoing federal and state investigations into large technology companies raise the prospect that courts will soon have to decide whether particular businesses qualify as two-sided transaction platforms.  Whether that characterization decision requires a legal or a factual determination will have several important implications for how any litigation proceeds. For instance, legal questions can be resolved relatively early on a motion to dismiss or a motion for summary judgment, while factual disputes cannot be resolved before trial. The types of evidence requiring development will also vary, as judges can resolve legal issues on a written record, while more engaging and dynamic testimony may be necessary to persuade a jury about how to analyze industry facts. Further, a prevailing party is less likely to preserve its victory on appeal if the lower court’s decision depends on legal conclusions that are reviewed de novo, rather than factual determinations, which usually receive deference.

Beyond its significance for individual cases, the distinction between law and fact will also influence the broader evolution of antitrust doctrine. Like other areas that develop through a common-law process, antitrust cases build on one another in path-dependent ways. US Airways and Sabre/Farelogix illustrate how, if the first case to apply Amex to a particular business holds as a matter of law that the business operates a two-sided transaction platform, other courts may be heavily influenced, or even bound, by that determination, however uncomfortable the result. The precedential value of these determinations raises the stakes in the first case, but there is no guarantee that the parties or the court that end up with the first case will be best positioned to litigate and resolve the issue for all future cases. While relitigating the facts of whether a firm operates a two-sided transaction platform in one case after another might seem less efficient, it also would preserve courts’ flexibility to reach just results in every case and ultimately lead to a body of antitrust decisions that better reflect the economic realities of an industry than would be possible if the first decision were to end all debate.

With so much at stake, parties will likely continue to contest whether characterizing a business as a two-sided transaction platform depends on a legal or factual determination. The next significant opinion may come in the FTC’s pending monopolization case against Surescripts, “a health information technology company operating in two complementary markets: electronic prescription routing . . . and eligibility, collectively known as ‘e-prescribing.’”80 And regardless of how that case is decided, litigants will continue to dispute whether particular businesses ought to be characterized as two-sided transaction platforms, and judges will have to decide whether that question is one of law to be resolved by the court, or one of fact often reserved for a jury.

Click here for a PDF version of the article

1 Andrew J. Ewalt is a partner in the Washington, DC office of Freshfields Bruckhaus Deringer US LLP. While working at the U.S. Department of Justice, he participated in the government’s case against American Express. The views expressed are not necessarily those of Freshfields or its clients.

2 138 S. Ct. 2274 (2018).

3 See Complaint, United States v. Am. Express Co., No. 10-cv-4496 (E.D.N.Y. Oct. 4, 2010), available at

4 See Brief for the United States as Respondent Supporting Petitioners at 3, Ohio v. Am. Express Co., No. 16-1454 (U.S. Dec. 7, 2017), available at (“The credit-card business involves a ‘two-sided’ platform, in which the value of a network to customers on each side depends in part on the number of customers on the other.”).

5 United States v. Am. Express Co., 88 F. Supp. 3d 143, 172-74 (E.D.N.Y. 2015).

6 See United States v. Am. Express Co., 838 F.3d 179, 196-200 (2d Cir. 2017), aff’d sub nom. Ohio v. Am. Express Co., 138 S. Ct. 2274 (2018).

7 Amex, 138 S. Ct. at 2287 (internal quotation marks omitted).

8 Id. at 2280.

9 Id.

10 Id. at 2281.

11 Id. at 2286, 2280.

12 Id. at 2280.

13 Id. at 2287.

14 Id. at 2286.

15 Id.

16 Id. at 2297-2301.

17 Id. at 2298.

18 Id. at 2299.

19 No. 14-MD-2541, 2018 WL 4241981 (N.D. Cal. Sept. 3, 2018).

20 See generally In re NCAA Grant-In-Aid Cap Antitrust Litig., 375 F. Supp. 3d 1058 (N.D. Cal. 2019).

21 O’Bannon v. NCAA, 802 F.3d 1049, 1070 (9th Cir. 2015); see also O’Bannon v. NCAA, 7 F. Supp. 3d 955, 986 (N.D. Cal. 2014) (finding a “national market in which NCAA Division I schools compete to sell unique bundles of goods and services to elite football and basketball recruits”).

22 See In re NCAA Grant-In-Aid Cap Antitrust Litig., No. 14-MD-2541, 2018 WL 1524005, at *8 (N.D. Cal. Mar. 28, 2018).

23 Id.

24 See In re NCAA Grant-In-Aid Cap Antitrust Litig., No. 14-MD-2541, 2018 WL 1948593, at *2-3 (N.D. Cal. Apr. 25, 2018).

25 In re NCAA Grant-In-Aid Cap Antitrust Litig., No. 14-MD-2541, 2018 WL 4241981, at *3 (N.D. Cal. Sept. 3, 2018).

26 Id. at *4.

27 Id.

28 Id. at *5.

29 In re Am. Express Anti-Steering Rules Antitrust Litig., 361 F. Supp. 3d 324, 338 (E.D.N.Y. 2019); see also Mem. Supp. Defs.’ Motion for Summ. J. at 4-8, In re Am. Express Anti-Steering Rules Antitrust Litig., No. 11-MD-2221 (E.D.N.Y. Oct. 12, 2018).

30 Anti-Steering Rules Antitrust Litig., 361 F. Supp. 3d at 339.

31 Id. (internal quotation marks omitted).

[xxxii] Id. at 338 (quoting Amex, 138 S. Ct. at 2285).

33 Id. at 341 n.12.

34 Id. at 339.

35 See supra note 4.

36 Anti-Steering Rules Antitrust Litig., 361 F. Supp. 3d at 341.

37 Id. at 340.

38 Id. at 341.

39 Id. at 340.

40 938 F.3d 43 (2d Cir. 2019).

41 Id. at 49-50.

42 Id. at 50, 51.

43 Id. at 51-52; see also United States v. Am. Express Co., 838 F.3d 179 (2d Cir. 2016).

44 US Airways, 938 F.3d at 53.

45 Id. at 58.

46 Id. at 57.

47 Id.

48 Id. at 58.

49 Id.

50 Id. at 59.

51 Amex, 138 S. Ct. at 2280-81, 2286.

52 US Airways, 938 F.3d at 59.

53 Amex, 138 S. Ct. at 2286.

54 See United States v. Sabre Corp., No. 19-cv-1548, 2020 WL 1855433, at *3 (D. Del. Apr. 7, 2020).

55 Id. at *10.

56 See Complaint ¶¶ 1, 22, 44-46, United States v. Sabre Corp., No. 19-cv-1548 (D. Del. Aug. 20, 2020).

57 Sabre, 2020 WL 1855433, at *32.

58 See id. at *32-35.

59 Id. at *32.

60 See id. at *33 (quoting US Airways, 938 F.3d at 58).

61 See id. at *34.

62 See id. at *33 (quoting Amex, 138 S. Ct. at 2287) (alteration omitted).

63 See id. at *35 n.16.

64 See id. at *15, 26-28.

65 See id. at *34.

66 Amex, 138 S. Ct. at 2287.

67 See Brief for Respondents at 46, Ohio v. Am. Express Co., No. 16-1454 (S. Ct. Jan. 16, 2018), available at​27977/​20180116154741999_pdfa%20Cravath%20-%20Ohio%20v%20Amex%20-%20Brief%20on%20Merits%20for%20Respondent.pdf (distinguishing types of two-sided platforms without discussing competition involving each type).

68 See Lapo Filistrucchi et al., Market Definition in Two-Sided Markets: Theory and Practice, 10 J. Comp. L. & Econ. 293 (2014).

69 Id. at 303 (emphasis added); see also id. at 301 (“[A]nalysis of a merger between two payment card platforms should . . . consider whether cash . . . exert[s] competitive pressure on payment card companies on both sides of the market.”).

70 See Sabre, 2020 WL 1855433, at *35 n.16; see also id. at *15.

71 See id. at *23-24, *37-38.

72 See id. at *33; see also Pls.’ Proposed Findings of Fact ¶ 147, United States v. Sabre Corp., No. 19-cv-1548 (D. Del. Feb. 20, 2020) (ECF No. 244).

73 See Sabre, 2020 WL 1855433,. at *35 n.16.

74 Notice of Appeal, United States v. Sabre Corp., No. 19-cv-1548 (D. Del. Apr. 8, 2020) (ECF No. 278).

75 Comp. & Makts. Auth., Anticipated Acquisition by Sabre Corp. of Farelogix Inc., Final Report (2020), available at https://assets.publishing.​service.​​media/5e8f17e4d3bf7f4120cb1881/Final_Report_-_Sabre_Farelogix.pdf.

76 Press Release, Sabre Corp., Sabre Corporation Issues Statement on its Merger Agreement with Farelogix (May 1, 2020), available at

77 The United States’ Suggestion of Mootness and Motion To Vacate the District Court’s Decision and Order Granting Judgment to Defendants at 6, United States v. Sabre Corp., No. 20-1767 (3d Cir. May 12, 2020), available at​file/​1277201/​download.

78 Id.

79 Order at 2, United States v. Sabre Corp., No. 20-1767 (3d Cir. July 20, 2020) (ECF No. 23).

80 See FTC v. Surescripts, LLC, No. 19-cv-1080, 2020 WL 264147 at *1 (D.D.C. Jan. 17, 2020).