Federal Trade Commission v. Facebook Inc.

Opinion

UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA

FEDERAL TRADE COMMISSION,

Plaintiff, v. Civil Action No. 20-3590 (JEB)

FACEBOOK, INC., Defendant.

MEMORANDUM OPINION

At the time of the last great antitrust battle in our courthouse — between the United States and Microsoft — Mark Zuckerberg was still in high school. Only after his arrival at Harvard did he launch “The Facebook” from his dorm room. Nearly twenty years later, both federal and state regulators contend, in two separate actions before this Court, that Facebook is now the one violating the antitrust laws. The company, they allege, has long had a monopoly in the market for what they call “Personal Social Networking Services.” And it has allegedly maintained that monopoly, in violation of Section 2 of the Sherman Act, through two different kinds of actions: first, by acquiring firms that it believed were well positioned to erode its monopoly — most notably, Instagram and WhatsApp; and second, by adopting policies preventing interoperability between Facebook and certain other apps that it saw as threats, thereby impeding their growth into viable competitors. Both suits seek equitable relief from this conduct, including forced “divestiture or reconstruction of businesses” as well as orders not to undertake similar conduct in the future. See ECF No. 3 (Redacted Compl.) at 51–52. (The

2

Court here cites a copy of the FTC’s Complaint that has minor redactions to protect confidential business information, and it mentions certain redacted facts only with the parties’ permission.)

Facebook now separately moves to dismiss both the State action and the FTC action.

This Opinion resolves its Motion as to the FTC’s Complaint, and the Court analyzes the States’ largely parallel claims in its separate Opinion in No. 20-3589. Although the Court does not agree with all of Facebook’s contentions here, it ultimately concurs that the agency’s Complaint is legally insufficient and must therefore be dismissed. The FTC has failed to plead enough facts to plausibly establish a necessary element of all of its Section 2 claims — namely, that Facebook has monopoly power in the market for Personal Social Networking (PSN) Services. The Complaint contains nothing on that score save the naked allegation that the company has had and still has a “dominant share of th[at] market (in excess of 60%).” Redacted Compl., ¶ 64. Such an unsupported assertion might (barely) suffice in a Section 2 case involving a more traditional goods market, in which the Court could reasonably infer that market share was measured by revenue, units sold, or some other typical metric. But this case involves no ordinary or intuitive market. Rather, PSN services are free to use, and the exact metes and bounds of what even constitutes a PSN service — i.e., which features of a company’s mobile app or website are included in that definition and which are excluded — are hardly crystal clear. In this unusual context, the FTC’s inability to offer any indication of the metric(s) or method(s) it used to calculate Facebook’s market share renders its vague “60%-plus” assertion too speculative and conclusory to go forward. Because this defect could conceivably be overcome by re-pleading, however, the Court will dismiss only the Complaint, not the case, and will do so without prejudice to allow Plaintiff to file an amended Complaint. See Ciralsky v. CIA., 355 F.3d 661, 666–67 (D.C. Cir. 2004).

3

To guide the parties in the event amendment occurs, this Opinion also explains two further conclusions of law. First, even if the FTC had sufficiently pleaded market power, its challenge to Facebook’s policy of refusing interoperability permissions with competing apps fails to state a claim for injunctive relief. As explained herein (and in the Court’s separate Opinion in the States’ case), there is nothing unlawful about having such a policy in general. While it is possible that Facebook’s implementation of that policy as to certain specific competitor apps may have violated Section 2, such finding would not change the outcome here: all such revocations of access occurred in 2013, seven years before this suit was filed, and the FTC lacks statutory authority to seek an injunction “based on [such] long-past conduct.” FTC v. Shire ViroPharma, Inc., 917 F.3d 147, 156 (3d Cir. 2019). Regardless of whether the FTC can amend its Complaint to plausibly allege market power and advance this litigation, then, the conduct it has alleged regarding Facebook’s interoperability policies cannot form the basis for Section 2 liability. Second, the agency is on firmer ground in scrutinizing the acquisitions of Instagram and WhatsApp, as the Court rejects Facebook’s argument that the FTC lacks authority to seek injunctive relief against those purchases. Whether other issues arise in a subsequent phase of litigation is dependent on how the Government wishes to proceed.

4

Table of Contents

I. Background ............................................................................................................................. 5 A. Social Networking .............................................................................................................. 5 B. Facebook Blue .................................................................................................................... 6 C. Alleged Monopoly Maintenance......................................................................................... 7 1. Instagram ........................................................................................................................ 8 2. WhatsApp ..................................................................................................................... 10 3. Interoperability Permissions ......................................................................................... 11 a. Facebook Platform .................................................................................................... 11 b. Conditioning Access ................................................................................................. 13 D. Procedural History ............................................................................................................ 15 II. Legal Standard ...................................................................................................................... 16 III. Analysis ................................................................................................................................. 17 A. Monopoly Power ............................................................................................................... 18 1. Market Definition ......................................................................................................... 19 a. Legal Framework ...................................................................................................... 19 b. Market-Definition Allegations .................................................................................. 21 c. Analysis..................................................................................................................... 23 2. Market Share ................................................................................................................. 27 B. Platform Policies ............................................................................................................... 33 1. Refusal to Deal.............................................................................................................. 34 a. Legal Framework ...................................................................................................... 34 b. Application ................................................................................................................ 39 i. Facebook Policies ................................................................................................. 39 ii. Specific Refusals ................................................................................................... 41 2. Conditional Dealing ...................................................................................................... 46 C. Challenging Acquisitions under Section 13(b) ................................................................. 50 IV. Conclusion ............................................................................................................................ 53

5

I. Background A. Social Networking At the dawn of our century, in the much earlier days of the internet, a number of websites began to offer what came to be known as “social networking” services. See Redacted Compl., ¶ 38. Friendster and Myspace, both launched in 2002, were among the earliest. Id. Although the precise definition of a “Personal Social Networking Service” is disputed (as that is the market in which Facebook has its alleged monopoly), it can be summarized here as one that enables users to virtually connect with others in their network and to digitally share their views and experiences by posting about them in a shared, virtual social space. Id., ¶ 40. For example, users might view and interact with a letter-to-the-editor-style post on politics by a neighbor, pictures from a friend’s recent party, or a birth announcement for a newborn cousin. Id.

Perhaps because humans are naturally social, this new way of interacting became hugely popular. Although Myspace and Friendster had an early lead, by 2009 they had been surpassed by a new competitor. Id., ¶¶ 38, 41. Created at Harvard in 2004, “The Facebook,” as it was initially called, was a social-networking service initially limited to college students. Id., ¶ 41. Within a few years, it had expanded to the general public (and dropped “The” from its name). Id. By at least 2011, it was the dominant player in personal social networking. Id., ¶ 62. Today, the FTC alleges, its flagship product, Facebook Blue, has hundreds of millions of users in the United States. Id., ¶ 3. The following details of Facebook’s conduct are drawn from the FTC’s Complaint, as the Court must consider its allegations true at this stage. The allegations are quite similar, though not identical, to those made by the States in the parallel case and recounted in the Court’s companion Opinion.

6

B. Facebook Blue Facebook Blue is what its millions of users think of when they think of “Facebook.”

Generally speaking, using Facebook Blue entails interacting with user-created content — i.e., content created or shared by one’s Facebook “friends,” id., ¶¶ 40, 89 — or creating content oneself by posting. That is not all that users see or do, however. They may also, for instance, encounter “publisher-created content like news articles . . . and advertisements” in their “news feed.” Id., ¶ 54; see also id., ¶¶ 44, 134. Such content can come in text, photo, or video form. Id., ¶ 54. In addition, Facebook users can play games or use other applications built either by Facebook or by third parties. Id., ¶¶ 97, 129. Facebook also offers other services beyond Facebook Blue to its users, such as Facebook Messenger, a free mobile-messaging service. Id., ¶¶ 37, 115.

Unlike most businesses, Facebook charges users no fee; instead, it makes money by selling advertising. Id., ¶¶ 43–51. By leveraging “the vast quantity of user data [it] collects,” the company “allows advertisers to target different campaigns and messages to different groups of users.” Id., ¶ 44; see also id., ¶ 4. Under this business model, as the Complaint puts it, Facebook “refrain[s] from charging a monetary price . . . to users, relying instead on monetizing user data and engagement through advertising.” Id., ¶ 42. Put differently, users exchange their time, attention, and personal data, rather than money, for access to Facebook. That approach has been highly profitable: in 2019, for instance, global advertisers paid Facebook nearly $70 billion, and it made profits of more than $18 billion. Id., ¶¶ 4, 44. To be clear, although Facebook’s data- collection and -use practices have been subject to increasing scrutiny, they are not the subject of this action.

7

C. Alleged Monopoly Maintenance Instead, this suit alleges that Facebook has violated and is violating the antitrust laws, the focus of which, generally speaking, is to promote and ensure competition. After rising to become the “dominant personal social networking provider in the United States” around 2011, id., ¶ 62, Facebook allegedly made a fateful strategic pivot: rather than competing to provide the best product, it would instead protect its monopoly by leveraging its power to foreclose and forestall the rise of new competitors. Id., ¶¶ 5, 9. In particular, the company’s executives saw a substantial threat to Facebook’s dominance in the advent of mobile devices — first and foremost, smartphones — capable of accessing the internet. Id., ¶ 70. Although Facebook had mobile functionality, it had been built with websites and desktop or laptop computers in mind and thus “offered a relatively poor experience for mobile users” compared to newer competitors. Id.; see also id., ¶¶ 78–79. Zuckerberg and other Facebook executives fretted over the possibility that other apps might create attractive mobile-native features and then leverage those features into exponential user growth, end-running Facebook’s established position. Id., ¶¶ 107–112. Even if such an app was not already providing social-network-like functionality, once it had a big enough base of users, it would still pose a potential threat to Facebook Blue. Id. Facebook executives feared fast-growing mobile-messaging services in particular, nervous that such apps could easily morph into direct competitors by adding social features.

In response to these perceived threats, the company allegedly used its monopoly power to eliminate or destroy competitors in order to maintain its market dominance. Id., ¶¶ 5–9. The FTC claims that this exclusionary conduct had “three main elements.” Id., ¶¶ 9, 71. First (and second), Facebook reached deep into its very deep pockets to acquire Instagram and WhatsApp, two promising potential competitors, thereby preventing their emergence as serious rivals. Id.,

8

¶ 71. (Attempts to purchase other competitors such as Snapchat and Twitter were rebuffed. Id., ¶ 73.) Third, it adopted and then enforced policies that blocked rival apps from interconnecting their product with Facebook Blue, thereby both (i) blunting the growth of potential competitors that might have used that interoperability to attract new users, and (ii) deterring other developers from building new apps or features or functionalities that might compete with Facebook, lest they lose access as well. Id., ¶¶ 23–26.

1. Instagram

Begin with Insta, as those in the know — viz., our children — refer to it. Launched in late 2010, Instagram was an innovative photo-editing and -sharing app designed for the era of smartphones with built-in cameras. Id., ¶¶ 79–80. Plaintiff alleges that Instagram’s photo- sharing app also qualifies as a PSN service, meaning that it was a direct competitor to Facebook Blue. Id., ¶ 63. From the get-go, Instagram’s user base grew explosively, eventually attracting the attention of Facebook executives who feared that their own photo-sharing features paled in comparison. Id., ¶¶ 81–85. That disparity gave Instagram a chance to reach a large enough scale to be threatening as a new, mobile- and photo-first social network — whether the firm got there on its own or if, as worried Facebook, it were purchased by a large company like Google or Apple. Id., ¶ 86. After about eighteen months of watching Instagram’s rise, Zuckerberg and his team eventually shifted from trying (and failing) to compete to instead trying to buy. Aiming to both neutralize Instagram as a competitor and “integrate” the “mechanics” of its popular photo- sharing features with Facebook Blue in order to forestall the growth of future Instagrams, id., ¶ 91, Zuckerberg offered to purchase the company for $1 billion in April 2012. Id., ¶ 95. Instagram’s founders agreed. Id.

9

As required by the Hart-Scott-Rodino Act, 15 U.S.C. § 18a, the FTC reviewed the acquisition prior to closing to assess whether it posed anticompetitive concerns. Whereas most mergers are cleared quickly, in this instance the review took over four months. During that scrutiny, the agency took the rare step of “requir[ing] the submission [by the parties] of additional information or documentary material relevant to the proposed acquisition.” 15 U.S.C. § 18a(e)(1)(A). Eventually, however, Facebook and Instagram satisfied the agency’s concerns, and in August (over four months after the merger was announced), the Commission voted 5–0 to allow it to proceed without any challenge or conditions. See FTC, FTC Closes its Investigation into Facebook’s Proposed Acquisition of Instagram Photo Sharing Program (Aug. 22, 2012), https://bit.ly/3bDa2mp. Although the FTC conveniently omits any mention of this review in its Complaint, the Court may take judicial notice of that public agency action. See Pharm. Rsch. & Manufacturers of Am. v. U.S. Dep’t of Health & Hum. Servs., 43 F. Supp. 3d 28, 33 (D.D.C. 2014); Herron v. Fannie Mae, No. 10-943, 2012 WL 13042852, at *1 (D.D.C. Mar. 28, 2012).

With Instagram safely in the fold, Facebook scaled back and eventually shut down its own mobile photo-sharing app. See Redacted Compl., ¶ 98. Internal emails cited by the Complaint reveal that it also fretted less about competition from other similar apps, since its ownership of Instagram meant it now “effectively dominate[d] photo sharing.” Id., ¶ 99. As time went on, Facebook also limited Facebook Blue’s promotion of the technically separate Instagram app and website, allegedly to avoid Instagram’s “cannibalizing” user engagement on its flagship service. Id., ¶¶ 102–04. All this post-acquisition conduct, the FTC claims, confirms that Facebook’s executives saw, and continue to see, Instagram as a significant competitive threat in the social-networking arena. Id., ¶ 102.

10

2. WhatsApp

The other high-profile acquisition Plaintiff focuses on here involves not a competitor in the PSN market, like Instagram, but a company that might quickly become one. As noted above, Facebook’s executives saw mobile-native apps in general as a threat. They were particularly concerned with internet-based, so-called “over-the-top mobile messaging services” such as WhatsApp. Id., ¶ 107. Since 2011, OTT messaging services have grown astronomically in use while SMS or MMS messaging (the kind of classic texting that relies on cellular networks rather than internet) has stagnated. Id. Even though mobile messaging services did not directly compete with Facebook Blue (as they are not PSN services), Facebook feared that such apps might well become competitors in the future; given the ubiquity of text messaging in modern life, a widely adopted messaging app could leverage its network effects to transition into a “mobile-first social network” by adding functions such as “gaming platforms, profiles, and news feeds.” Id., ¶ 111; see id., ¶¶ 108–112.

Facebook executives saw WhatsApp as the most potent threat among mobile-messaging services. Id., ¶ 113. Launched in 2009, it had approximately 450 million active users worldwide five years later and was growing exponentially thanks to its superior product. Id., ¶¶ 113–18. Zuckerberg and his team hoped that their Facebook Messenger app, released in 2011, would compete. Id., ¶¶ 115–16. But as WhatsApp continued to thrive and expand, Facebook instead resolved to try to buy it. Id., ¶ 120. After being initially rebuffed in late 2012, id., ¶ 121, that tactic found success in February 2014, when the two companies agreed on a purchase price of $19 billion. Id. The transaction was also subject to Hart-Scott-Rodino Act pre-merger review, see 18 U.S.C. § 18b, but the FTC, once again, did not block it.

11

Since acquiring WhatsApp, the agency alleges, Facebook has “kept [it] cabined to providing mobile messaging services rather than allowing” it to grow into a standalone PSN service. See Redacted Compl., ¶ 126. As with Instagram, Facebook has also limited its promotion of WhatsApp on its other services in the United States. Id. It follows, Plaintiff further claims, that “Facebook’s monopolization” via both its WhatsApp and Instagram acquisitions “is ongoing,” as it both “continues to hold and operate [the two companies], which neutralizes their direct competitive threats to Facebook,” and “continues to keep them positioned to provide a protective ‘moat’ around its [PSN] monopoly.” Id., ¶ 76.

3. Interoperability Permissions a. Facebook Platform

Not long after it expanded to the general public, Facebook released “Facebook Platform,”

a set of tools that allowed software developers to create interoperability between their products and Facebook Blue. Id., ¶ 129. As initially launched, Platform “encouraged software developers to build an entire ecosystem of apps and tools” that would be displayed and used within the Facebook website itself. Id. Such apps “rang[ed] games and page design tools to video-sharing tools and e-marketing apps.” Id. (The States’ Complaint in the parallel case refers to these apps as “canvas” apps. See No. 20-3589, ECF No. 4 (State Redacted Compl.), ¶ 190.) Such apps would make money by allowing users to purchase virtual goods or items within the app on a “freemium” model or via ad sales.

Three years later, in 2010, Facebook added new functionalities to Platform that expanded its reach off the Facebook site itself. These tools — called application programming interfaces or APIs — created mechanisms for sharing data between Facebook and other, freestanding third- party apps. See Redacted Compl., ¶ 130. One important API that Facebook offered to

12

developers was the “Find Friends” API, id., which enabled third-party apps to allow Facebook account holders to find and connect with Facebook friends within their separate apps, or to invite Facebook friends to join that app. Id. For instance, when first starting to use an independent chess app — i.e., an app used separately as opposed to on the Facebook site itself — a user with a Facebook account could nonetheless search within the app for other Facebook friends already using it, or invite them to join via Facebook, all without leaving the app. Another API allowed Facebook users to sign into third-party websites or apps using their Facebook log-in credentials. Id., ¶¶ 144, 154.

Facebook went even further in that direction later in the year when it launched its Open Graph API. Id., ¶ 131. Open Graph allowed third-party apps and websites to essentially integrate pieces of Facebook within their own service; for instance, apps could install the famous “Like” button, which, if clicked, would share a user’s “like” on the user’s Facebook profile. Id. Users could do this without even navigating away from the third-party service. Id., ¶¶ 131, 134. A user reading an article on WashingtonPost.com, for instance, could now like an article directly on-site and further choose to post a link of the article to the user’s Facebook profile. This sort of integration was, unsurprisingly, massively popular among app developers. “By July 2012, Open Graph was being used to share nearly one billion pieces of social data each day to Facebook Blue, giving Facebook substantially greater and richer information about its users and their online activities.” Id., ¶ 132.

According to Plaintiff, Facebook benefited significantly from its Platform program and open APIs. The company garnered goodwill and continued to increase its growth and user engagement. Id., ¶¶ 133–34. It also obtained access to a massive new trove of off-site user data. Id., ¶ 134. Third-party app developers likewise gained, improving the quality of users’

13

experience by integrating social functionality and benefiting from Facebook’s sizeable network of highly engaged users. Id., ¶¶ 132–33. Users, too, presumably enjoyed the increased efficiency and convenience.

b. Conditioning Access

Nonetheless, Facebook eventually began to use the power of its Platform tools over the growth trajectory of nascent apps to “deter and suppress competitive threats to its personal social networking monopoly.” Id., ¶ 136. Specifically, the FTC alleges, the company adopted policies under which its APIs would be “available to developers only on the condition that their apps” did not compete with Facebook Blue (or Facebook Messenger). Id. The company then enforced those policies against “apps that violated the[] conditions by cutting off their use of commercially significant APIs.” Id.

Facebook announced the first iteration of these policies in July 2011, alerting developers that going forward, “Apps on Facebook [could] not integrate, link to, promote, distribute, or redirect to any app on any other competing social platform.” Id., ¶ 139. This policy, by its terms, applied only to “[a]pps on Facebook” — i.e, the “canvas” apps described above that could only be accessed and used on the Facebook website itself. Id. Put differently, this initial policy did not affect the sort of freestanding, independent apps discussed above, such as our chess app or the Washington Post app. It was not until later that Facebook “imposed several other policies restricting” freestanding apps’ “use of Facebook Platform, including [the] APIs” just discussed. Id., ¶ 141. The first of those additional policies, announced in 2012, prohibited developers from “us[ing] Facebook Platform to export [Facebook] user data into a competing social network without our permission.” Id., ¶ 142. The next year, Facebook went further by instructing developers that their apps could “not use Facebook Platform to promote, or to export user data

14

to, a product or service that replicates a core Facebook product or service without our permission.” Id., ¶ 143.

Armed with these policies, Facebook then enforced them by cutting off API access to certain apps. As Plaintiff describes it, those cutoffs were “generally directed against apps in three groups.” Id., ¶ 152. First, Facebook terminated the API access of promising apps that were directly competing with Facebook Blue by providing Personal Social Networking Services, such as Path, a feed-based sharing app that limited the number of friends a user could have to encourage more intimate sharing. Id., ¶ 153. Second, Facebook targeted “promising apps with some social functionality” but which were not yet full-fledged competitors to Facebook Blue. Id., ¶ 154. As examples, the Complaint provides Vine, a video-sharing app owned by Twitter to which Facebook shut down API access in January 2013, and Circle, a “local social network” that had its permissions revoked in December of that year. Id., ¶¶ 154–55. Last, “Facebook blocked mobile messaging apps from using commercially significant APIs”; at one point, in August 2013, it “undertook an enforcement strike against a number” of such apps “simultaneously.” Id., ¶ 156.

Each of these revocations of access, the FTC alleges, significantly “hindered the ability of [the targeted] businesses to grow and threaten Facebook’s personal social networking monopoly.” Id., ¶ 157. During the period in which Circle had Facebook API access, for example, it was growing at a rate of 600,000-800,000 users per day; after losing its Facebook interconnections (particularly the Find Friends tool), however, its “daily new users dropped . . . to nearly zero.” Id., ¶ 154. Facebook’s actions also allegedly “alerted other apps that they would lose access . . . if they, too, posed a threat to Facebook’s . . . monopoly,” thereby deterring other apps from adding features or functionality that would attract the company’s ire. Id., ¶ 158.

15

There is an important coda to this story, however: Facebook “removed its ‘core functionality’ restrictions” in December 2018. Id., ¶ 148 (emphasis added). Although the company has not reinstated the policies (or, according to the Complaint, revoked any apps’ API access) since that time, the FTC alleges that Facebook “is likely to reinstitute such policies if [public] scrutiny passes.” Id., ¶¶ 149, 172.

D. Procedural History The FTC filed this action on December 9, 2020, asserting that the above conduct amounts to one count of monopoly maintenance under Section 2 of the Sherman Act. Id., ¶¶ 169–73. Unlike the States, the agency does not also allege a violation of Section 7 of the Clayton Act, which prohibits acquisitions that will substantially decrease competition. Plaintiff brings this suit under Section 13(b) of the FTC Act, which authorizes it to seek an injunction against an entity that “is violating” or “is about to violate” the antitrust laws, including Section 2. See 15 U.S.C. § 53(b). It hopes to procure an injunction aimed at preventing such conduct in the future as well as an order mandating “divestiture of assets, divestiture or reconstruction of businesses (including, but not limited to, Instagram and/or WhatsApp), and such other relief sufficient to restore the competition that would exist absent the conduct alleged in the Complaint.” Redacted Compl. at 51.

This case was initially assigned to Judge Christopher R. Cooper of this district. As noted above, however, a number of State Plaintiffs filed a very similar suit against Facebook just before this action was filed. That case, No. 20-3589, was assigned to this Court. Pursuant to Local Rule 40.5(c)(2), which governs related cases, Judge Cooper was required to reassign this action to this Court because it presides over the earlier-filed State case. See No. 20-3590, Minute Order of Jan. 12, 2021.

16

Facebook has now moved to dismiss both actions. See ECF No. 56 (MTD FTC); No. 20-

3589, ECF No. 114 (MTD States). While the cases could be consolidated, the Court believes that clarity will be enhanced by resolving the two Motions to Dismiss in separate, contemporaneously issued Opinions. As explained in its separate Opinion, it will grant the Motion to Dismiss the States’ entire case. See Mem. Op., No. 20-3589. By contrast, the Court here will dismiss only the Complaint, not the case, leaving the agency the chance to replead if it believes it can successfully remedy the infirmities described below. II. Legal Standard Facebook moves to dismiss this action under Federal Rule of Civil Procedure 12(b)(6) for failure to state a claim upon which relief can be granted. See MTD FTC at 1. In evaluating such Motion to Dismiss, the Court must “treat the complaint’s factual allegations as true . . . and must grant plaintiff ‘the benefit of all inferences that can be derived from the facts alleged.’” Sparrow v. United Air Lines, Inc., 216 F.3d 1111, 1113 (D.C. Cir. 2000) (quoting Schuler v. United States, 617 F.2d 605, 608 (D.C. Cir. 1979)). Although “detailed factual allegations” are not necessary to withstand a Rule 12(b)(6) motion, Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 555 (2007), “a complaint must contain sufficient factual matter, accepted as true, to ‘state a claim to relief that is plausible on its face,” Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009) (quoting Twombly, 550 U.S. at 570) — that is, the facts alleged in the complaint “must be enough to raise a right to relief above the speculative level.” Twombly, 550 U.S. at 555. The Court need not accept as true, then, “a legal conclusion couched as a factual allegation,” Trudeau v. FTC, 456 F.3d 178, 193 (D.C. Cir. 2006) (quoting Papasan v. Allain, 478 U.S. 265, 286 (1986)), nor “inferences . . . unsupported by the facts set out in the complaint.” Id. (quoting Kowal v. MCI Commc’ns Corp., 16 F.3d 1271, 1276 (D.C. Cir. 1994)). And it may consider not only “the facts

17

alleged in the complaint,” but also “any documents either attached to or incorporated in the complaint[,] and matters of which [courts] may take judicial notice.” Equal Emp’t Opportunity Comm’n v. St. Francis Xavier Parochial Sch., 117 F.3d 621, 624 (D.C. Cir. 1997). III. Analysis The offense of monopoly maintenance under Section 2 of the Sherman Act “has two elements: ‘(1) the possession of monopoly power in the relevant market and (2) the willful . . . maintenance of that power as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident.’” United States v. Microsoft Corp., 253 F.3d 34, 50 (D.C. Cir. 2001) (quoting United States v. Grinnell Corp., 384 U.S. 563, 570–71 (1966)). This second element is usually referred to by the shorthand of “anticompetitive” or “exclusionary conduct.” Id. at 58. Facebook, in seeking dismissal, contends that the Complaint does not allege facts establishing either element. The Court agrees that the first — the possession of monopoly power in the market for Personal Social Networking Services (as defined by the agency) — is not adequately pleaded here. No more is needed to conclude that the Complaint must be dismissed.

To promote clarity and efficiency going forward in the event Plaintiff amends its Complaint, however, the Court will also address two other issues in this Opinion. Both pertain to Section 13(b) of the FTC Act, which the agency relies on to seek an injunction here, and which authorizes such relief only where the defendant “is violating, or is about to violate,” the antitrust laws. See 15 U.S.C. § 53(b). First, to the extent that Facebook’s Platform-related conduct is actionable, it occurred nearly eight years ago, rendering an injunction under Section 13(b) unavailable as a matter of law. Second, and on the other hand, Section 13(b) does allow

18

the FTC to challenge Facebook’s acquisitions of Instagram and WhatsApp, contrary to the company’s contention otherwise.

A. Monopoly Power Begin with the linchpin of this Opinion: whether the FTC has plausibly alleged, as it must, that Facebook exercises monopoly power. As explained by the Circuit in Microsoft, monopoly power is the “the power to control prices or exclude competition,” such that a firm is a monopolist “if it can profitably raise prices substantially above the competitive level.” 253 F.3d at 51 (citations omitted). Where a plaintiff can provide direct proof that a “firm has in fact profitably done so, the existence of monopoly power is clear.” Id. Because such proof is rare, however, plaintiffs and courts usually search for indirect or “circumstantial evidence” of monopoly power by inferring it from “a firm’s possession of a dominant share of a relevant market.” Id.; see also Toys “R” Us, Inc. v. FTC, 221 F.3d 928, 937 (7th Cir. 2000) (noting that market power can be proven “through direct evidence of anticompetitive effects” or, “more conventional[ly],” “by proving relevant product and geographic markets and by showing that the defendant’s share exceeds [some] threshold”); S. Pac. Commc’ns Co. v. Am. Tel. & Tel. Co., 740 F.2d 980, 1000 (D.C. Cir. 1984) (“[C]ourts frequently approach the problem of measuring market power by defining the relevant product and geographic market and computing the defendant’s market share. Monopoly power is then ordinarily inferred from a predominant share of the market.”). Because “[m]arket power is meaningful only if it is durable,” a plaintiff proceeding by the indirect method of providing a relevant market and share thereof must also show that there are “barriers to entry” into that market. Lenox MacLaren Surgical Corp. v. Medtronic, Inc., 762 F.3d 1114, 1123–25 (10th Cir. 2014); Microsoft, 253 F.3d at 51 (explaining that defendant’s “share of a relevant market” must be “protected by entry barriers,” defined as

19

“factors . . . that prevent new rivals from timely responding to an increase in price above the competitive level”).

Although the FTC briefly suggests in its Opposition that it can offer direct proof of market power, see ECF No. 59 (FTC Opp.) at 8, it spends nearly its entire brief arguing why it has sufficiently pleaded indirect proof — viz., that Facebook has a dominant share of a relevant product and geographic market (the United States market for Personal Social Networking Services) protected by entry barriers. Id. at 8–19. Because the agency thus makes no real direct- proof argument, the Court will analyze the Complaint’s market-power allegations using the indirect framework. Again, that framework first requires the plaintiff to “establish[] the relevant market” in which the defendant firm allegedly has monopoly power. Sky Angel U.S., LLC v. Nat’l Cable Satellite Corp., 947 F. Supp. 2d 88, 102 (D.D.C. 2013) (quoting Neumann v. Reinforced Earth Co., 786 F.2d 424, 429 (D.C. Cir. 1986)). It then demands that a plaintiff establish that the defendant has a dominant share of that market protected by entry barriers. Id.; see, e.g., FTC v. AbbVie Inc., 976 F.3d 327, 373–74 (3d Cir. 2020) (above 60% market share sufficient); Image Tech. Servs. v. Eastman Kodak Co., 125 F.3d 1195, 1206 (9th Cir. 1997) (“Courts generally require a 65% market share to establish a prima facie case of market power.”). As the Court explains below, it is the market-share step that trips up the FTC here.

1. Market Definition

a. Legal Framework

Although the definition of a relevant antitrust market is typically a “factual” rather than a “legal” inquiry, certain “legal principles” nevertheless govern. Newcal Indus., Inc. v. Ikon Off. Sol., 513 F.3d 1038, 1045 (9th Cir. 2008). An antitrust market includes “two components: the product market and the geographic market.” Sky Angel, 947 F. Supp. 2d at 102. Because the

20

parties do not dispute here that the geographic market — that is, “the terrain in which competition takes place,” Novell, Inc. v. Microsoft Corp., 731 F.3d 1064, 1071 (10th Cir. 2013) — is the United States, see FTC Opp. at 10, the only issue is the “[t]he outer boundaries of [the relevant] product market” in which Facebook operates. Sky Angel, 947 F. Supp. 2d at 102 (quoting Brown Shoe Co. v. United States, 370 U.S. 294, 325 (1962)).

“A ‘relevant product market’ is a term of art in antitrust analysis,” United States v. H&R Block, Inc., 833 F. Supp. 2d 36, 50 (D.D.C. 2011), and is defined as “all products reasonably interchangeable by consumers for the same purposes.” Microsoft, 253 F.3d at 52. “Because the ability of consumers to turn to other suppliers restrains a firm from raising prices above the competitive level,” id. at 51, the analysis of market power (which simply means the power to do just that) must use, as its denominator, all products “roughly equivalent to another for the use to which [they are] put.” Queen City Pizza, Inc. v. Domino’s Pizza, Inc., 124 F.3d 430, 437 (3d Cir. 1997) (citation omitted). “In other words, courts look at whether two products can be used for the same purpose, and, if so, whether and to what extent purchasers are willing to substitute one for the other.” H&R Block, 833 F. Supp. 2d at 51 (citation omitted). At bottom, products that are sufficiently interchangeable compete with each other in the relevant legal sense. See Hicks v. PGA Tour, Inc., 897 F.3d 1109, 1120 (9th Cir. 2018).

Unsurprisingly, many “case[s] hing[e]” on this fight over what is included in this denominator. See, e.g., FTC v. Staples, Inc., 970 F. Supp. 1066, 1073 (D.D.C. 1997) (considering whether product market includes only “consumable office supplies” such as paper, pens, and post-it notes, or instead all office products, including computers, office furniture, and fax machines); H&R Block, 833 F. Supp. 2d at 50 (considering whether product market includes only “digital do-it-yourself tax preparation products” or instead includes “all tax preparation

21

methods,” including professional assistance and do-it-yourself using pen and paper); FTC v. Swedish Match, 131 F. Supp. 2d 151, 156 (D.D.C. 2000) (considering whether product market includes only loose-leaf chewing tobacco or instead also includes moist snuff).

b. Market-Definition Allegations The market-definition inquiry in this case is somewhat unusual because, unlike familiar consumer goods like tobacco or office supplies, there is no obvious or universally agreed-upon definition of just what a personal social networking service is. As a result, to discharge its “burden to define the relevant market,” Gross v. Wright, 185 F. Supp. 3d 39, 50 (D.D.C. 2016) (citation omitted), the FTC must do two things here. First, it must provide a definition of PSN services (which, obviously, would include at least Facebook Blue). Second, it must further explain whether and why other, non-PSN services available to the public either are or are not reasonably interchangeable substitutes with PSN services. Ultimately, that analysis should demonstrate that Facebook holds a dominant share of a market that includes such substitutes, if any.

On the first point, the agency explains that PSN services are “online services that enable and are used by people to maintain personal relationships and share experiences with friends, family, and other personal connections in a shared social space.” Redacted Compl., ¶ 52. Such services are allegedly defined, and distinguished, by their having “[t]hree key elements.” Id. “First, [they] are built on a social graph that maps the connections between users and their friends, family, and other personal connections.” Id., ¶ 53. “Second, [they] include features that many users regularly employ to interact with personal connections and share their personal experiences in a shared [virtual] social space, including in a one-to-many ‘broadcast’ format.” Id., ¶ 54. And “[t]hird, [they] include features that allow users to find and connect with other

22

users, to make it easier for each user to build and expand their set of personal connections. The social graph also supports this feature by informing [the user] which [new] connections” might be available based on her existing network. Id., ¶ 55.

Having defined PSN services, Plaintiff then alleges that there are in fact no “other types of internet services” that are “adequate substitutes.” Id., ¶ 57. It buttresses that conclusion by explaining why four different kinds of arguably comparable online services are not “reasonably interchangeable” with PSN services. Id., ¶ 58. First, “specialized social networking services” that “focus on professional . . . connections” (e.g., LinkedIn) are not substitutes because they are designed for and used primarily by professionals for sharing professional content. Id., ¶ 58. They therefore would not be used, as PSN services are, to “maintain personal relationships and share experiences with friends, family, and other personal connections.” Id., ¶ 52. The same is true, alleges the FTC, for “interest-based” social-networking services such as Strava (which relates to physical exercise). Id., ¶ 58. The agency also pleads that PSN services are not reasonably interchangeable with services that allow for consuming and sharing video or audio content, such as YouTube, Spotify, Netflix, or Hulu. Id., ¶ 59. That is because users of such services mostly consume such content passively or share content created by others (rather than content they have created), and such sharing, where it occurs, is not to the user’s network of personal connections but rather to a general and wide audience of unknown users. Id. In such a setting, users do not usually “communicate with friends, family, and other personal connections,” which is the hallmark of a PSN service. Id. Finally, Plaintiff explains that “mobile messaging services” cannot be substituted for PSN services because the former (i) lack a “shared social space” for interaction and (ii) do not employ a social graph to facilitate users’ finding and “friending” other users they may know. Id., ¶ 60. Zuckerberg himself has colorfully explained

23

one key difference in use that allegedly flows from these disparate features: a PSN service is “the digital equivalent of a town square,” whereas a mobile messaging service is “the digital equivalent of [a] living room.” Id.

According to the FTC, then, the relevant market here thus includes PSN services — such as Facebook Blue, Instagram, and Path, id., ¶¶ 63, 153 — and no other kinds of services.

c. Analysis

On this issue, Facebook contends that, for a number of reasons, Plaintiff’s allegations regarding market definition fall short of the pleading-stage requirement that such “alleged product market . . . be plausible.” Sky Angel, 947 F. Supp. 2d at 103 (quoting Todd v. Exxon Corp., 275 F.3d 191, 200 (2d Cir. 2001)). While there are certainly bones that one could pick with the FTC’s market-definition allegations, the Court does not find them fatally devoid of meat.

First, Facebook contends that the FTC’s Complaint contains an internal contradiction: its market definition appears to exclude services like Circle and Vine, yet one of the core allegations in this case is that Facebook’s revocation of API permissions from those apps was anticompetitive. See MTD FTC at 14–15. Both cannot be true, Defendant insists. As the FTC explains in response, though, the D.C. Circuit rejected exactly this argument in Microsoft. There, Microsoft argued that the district court had been wrong to exclude “middleware” software from the relevant product market for computer operating systems because much of the Government’s Section 2 case turned on “Microsoft’s attempts to suppress middleware’s threat to its operating system monopoly.” 253 F.3d at 54. The Circuit explained that “no contradiction exist[ed],” as middleware was still a “nascent” threat even though it was not yet competing directly in Microsoft’s operating-system market, and “[n]othing in § 2 of the Sherman Act limits

24

its prohibition to actions taken against threats that are already well-developed enough to serve as present substitutes.” Id. So too here: actions taken against Vine and Circle may have been anticompetitive even though those firms were not Facebook Blue’s competitors in a properly drawn product market.

Second, Defendant maintains, the FTC has neglected to allege any facts regarding the “cross-elasticity of demand between [PSN services] and [potential] substitutes for it.” Sky Angel, 947 F. Supp. 2d at 103 (citation omitted); see MTD FTC at 11–12. Cross-elasticity of demand is a measure of the degree to which “the rise in the price of [one] good] . . . would tend to create a greater demand for other like goods.” Queen City Pizza, 124 F.3d at 437–38 & n.6. It is thus one “measure of reasonable interchangeability.” Id. at 438 n.6. There is no authority, however, supporting Facebook’s argument that Plaintiff must plead specific facts regarding the price or non-price terms under which PSN-service users would switch (if ever) to alternatives. Instead, at this stage the FTC may permissibly plead that certain “factors” of both the service at issue and its potential substitutes — e.g., their “price, use[,] and qualities” — render them not “reasonably interchangeable” in the eyes of users. Tunis Bros. Co., Inc. v. Ford Motor Co., 952 F.2d 715, 722 (3d Cir. 1991); see, e.g., RealPage, Inc. v. Yardi Sys., Inc., 852 F. Supp. 2d 1215, 1225 (C.D. Cal. 2012) (denying motion to dismiss where complaint explained in qualitative terms why certain “conceivably interchangeable substitutes” were not “specialized to the needs” of defendant’s customers); CollegeNet, Inc. v. Common Application, Inc., 355 F. Supp. 3d 926, 958 (D. Or. 2018) (same where complaint alleged that product’s “distinct features” made it such that customers “would not turn to [alternatives] in response to” price increase); see also Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law vol. IIB, ¶ 531f, at 252 & n.24 (4th ed. 2014) (suggesting only that “some detail about lack of interchangeability should be given” in

25

complaint). That is what the agency has done here, albeit in a somewhat lean fashion. See Redacted Compl., ¶¶ 58–60.

Defendant next directly takes aim at the FTC’s allegation that users of PSN services would not switch, if prodded by a price increase or quality decrease in a PSN service, to other means of communicating and sharing with their personal connections that lack a “connection- finder” built on the user’s social graph (the third leg of the agency’s definition of a PSN). See MTD FTC at 17–19; see also ECF No. 62 (Reply FTC) at 7. That is implausible, Facebook contends, because it is “obvious that people” also know how to “connect and share with family and friends . . . via many [other] technologies,” such as “email, messaging, photo-sharing, and video-chats.” MTD FTC at 18 (internal quotation marks omitted). This argument asks the Court to engage in the sort of “deeply fact-intensive inquiry” that is improper at this stage. E.I. du Pont de Nemours & Co. v. Kolon Indus., Inc., 637 F.3d 435, 443–44 (4th Cir. 2011) (quoting Todd, 275 F.3d at 200). Although open to dispute, the agency’s allegation that users view services with and without a social-graph-based connection-finder as fundamentally different and non- interchangeable is “at least . . . theoretically rational,” Bayer Schering Pharma AG v. Sandoz, Inc., 813 F. Supp. 2d 569, 577 (S.D.N.Y. 2011) (citation omitted), and thus hardly “facially unsustainable,” RealPage, 852 F. Supp. 2d at 1224, or “untenable on its face.” Suture Exp., Inc. v. Cardinal Health 200, LLC, 963 F. Supp. 2d 1212, 1222 (D. Kan. 2013). This is therefore not one of the “relatively rare” cases of a “glaring deficienc[y]” in the market-definition pleadings that renders dismissal at the 12(b)(6) stage appropriate. Kolon Indus., 537 F.3d at 444.

Facebook’s final fruitless market-definition argument is that the Complaint impermissibly distinguishes PSN services from other possible substitutes based on their “primar[y] uses.” MTD FTC at 19. The company asserts that the question is whether other

26

services “can perform the same function[s]” as PSN services, not whether they are primarily used that way. Id. (quoting and adding emphasis to Cupp v. Alberto-Culver USA, Inc., 310 F. Supp. 2d 963, 970 (W.D. Tenn. 2004)). That misstates the law. The analysis looks to both “whether two products can be used for the same purpose, and, if so, whether and to what extent purchasers are willing to substitute one for the other.” H&R Block, 833 F. Supp. 2d at 51 (emphasis added); see also United States v. Aetna Inc., 240 F. Supp. 3d 1, 19 (D.D.C. 2017) (asking “whether there are other products offered to consumers which are similar in character or use, and how far buyers will go to substitute one commodity for another”) (emphasis added). It is for this reason that pen-and-paper do-it-yourself tax prep and assisted tax prep could be outside the market for digital do-it-yourself tax-prep services despite both providing the same basic function, H&R Block, 833 F. Supp. 2d at 54, and that Medicare Advantage plans and Original Medicare plans could constitute distinct product markets despite both being capable of providing healthcare insurance to seniors. Aetna, 240 F. Supp. 3d at 19, 41. All Plaintiff must do at this stage is provide a “plausible explanation as to why” users would not switch, even if they technically could, from PSN services to other services if prompted by a price hike. Todd, 275 F.3d at 200; cf. Concord Assocs., L.P. v. Ent. Properties Tr., 817 F.3d 46, 54 (2d Cir. 2016) (requiring “a plausible basis for explaining why [a proffered distinction] makes the difference” to consumers). While the agency certainly could have provided more on that front, the fact that other services are not primarily used for the sort of personal sharing that is the hallmark of a PSN service seems a plausible reason why little switching would occur. Whether due to network effects or the norms around what sort of content is generally posted on different platforms, it is not a stretch to imagine that users are reluctant to share a highly personal milestone on LinkedIn or post a video of their child’s first steps to YouTube.

27

2. Market Share

Although the Court therefore finds that the Complaint’s allegations do enough to make out a plausible market for PSN services, that hardly ends the analysis. After all, while courts typically evaluate market definition and market share separately, the two inquiries ultimately come together to produce the conclusion that matters: the defendant’s market power. See Cupp, 310 F. Supp. 2d at 971, 975. That point can be lost if courts “compartmentalize the two issues” and examine the sufficiency of a plaintiff’s market-share showings without looking back to how convincingly, or “how tenuously,” the market has been defined. See Areeda & Hovenkamp, ¶ 531, at 257. Particularly where “the market is idiosyncratically drawn,” the treatise authors explain, findings or allegations as to a “particular ‘market share’ carr[y] much less meaning.” Id.

This case, at least as pleaded by the FTC, exemplifies that principle. Although the Court, as just explained, finds the contours of the asserted product market plausible, the Complaint is undoubtedly light on specific factual allegations regarding consumer-switching preferences. Given that thin showing, and the fact that the PSN-services product market is somewhat “idiosyncratically drawn” to begin with, the Court must demand something more robust from Plaintiff’s market-share allegations. As it happens, however, those allegations are even more tentative: the FTC alleges only that Facebook has “maintained a dominant share of the U.S. personal social networking market (in excess of 60%)” since 2011, see Redacted Compl., ¶ 64, and that “no other social network of comparable scale exists in the United States.” Id., ¶ 3. That is it. These allegations — which do not even provide an estimated actual figure or range for Facebook’s market share at any point over the past ten years — ultimately fall short of plausibly establishing that Facebook holds market power. Given that finding, the court need not address the issue of whether the FTC has sufficiently alleged entry barriers.

28

Off the bat, there is ample authority that the FTC’s bare assertions would be too conclusory to plausibly establish market power in any context. See Synthes, Inc. v. Emerge Med., Inc., No. 11-1566, 2012 WL 4473228, at *11 (E.D. Pa. Sept. 28, 2012) (allegation that defendant “is a monopoly . . . with over 50% market share” was a “threadbare recital unsupported by factual allegations [that] the Court need not accept . . . as true”); Syncsort Inc. v. Sequential Software, Inc., 50 F. Supp. 2d 318, 330 (D.N.J. 1999) (“Here, [plaintiff] recited in conclusory fashion that [defendant] ‘controls the majority of the [relevant] market.’ . . . [T]his single statement of market power in the pleadings . . . is an insufficient allegation of the possession of monopoly power.”); Korea Kumho Petrochemical v. Flexsys Am. LP, No. 07- 1057, 2008 WL 686834, at *9 (N.D. Cal. Mar. 11, 2008) (holding that “[a]lthough [p]laintiff need not necessarily quantify [defendant’s] market share with precision,” allegation that defendant “domina[ted] . . . the [relevant] market” fell short of requirement to “assert some facts in support of its assertions of market power”); EuroTec Vertical Flight Sols., LLC. v. Safran Helicopter Engines S.A.A., No. 15-3454, 2019 WL 3503240, at *3 (N.D. Tex. Aug. 1, 2019) (noting that bare allegation of “market share of over 50 percent” was “conclusory”); Sherwin- Williams Co. v. Dynamic Auto Images, Inc., No. 16-1792, 2017 WL 3081822, at *7 (C.D. Cal. Mar. 10, 2017) (“[Claimants’] allegation that [firm] maintains a ‘a stranglehold in the automotive paint industry’ is . . . conclusory” and thus “lacks sufficient detail for the Court to plausibly infer . . . sufficient market power.”). It is hard to imagine a market-share allegation that is much more conclusory than the FTC’s here.

Even accepting that merely alleging market share “in excess of 60%” might sometimes be acceptable, it cannot suffice in this context, where Plaintiff does not even allege what it is measuring. Indeed, in its Opposition the FTC expressly contends that it need not “specify which

29

. . . metrics . . . [or] ‘method’ [it] used to calculate Facebook’s [market] share.” FTC Opp. at 18. In a case involving a more typical goods market, perhaps the Court might be able to reasonably infer how Plaintiff arrived at its calculations — e.g., by proportion of total revenue or of units sold. See U.S. Dep’t of Justice & FTC, Horizontal Merger Guidelines § 5.2 (2010) (suggesting these to be the typical methods). As the above market-definition analysis underscores, however, the market at issue here is unusual in a number of ways, including that the products therein are not sold for a price, meaning that PSN services earn no direct revenue from users. The Court is thus unable to understand exactly what the agency’s “60%-plus” figure is even referring to, let alone able to infer the underlying facts that might substantiate it.

Rather than undergirding any inference of market power, Plaintiff’s allegations make it even less clear what the agency might be measuring. The overall revenues earned by PSN services cannot be the right metric for measuring market share here, as those revenues are all earned in a separate market — viz., the market for advertising. See Redacted Compl., ¶ 164; see also, e.g., id., ¶ 101 (noting that prior to its acquisition, in addition to competing in the PSN services market, “Instagram also planned and expected to be an important advertising competitor” to Facebook). Percent of “daily users [or] monthly users” of PSN services — metrics the Complaint mentions offhandedly, see Redacted Compl., ¶¶ 3, 97 — are not much better, as they might significantly overstate or understate any one firm’s market share depending on the various proportions of users who have accounts on multiple services, not to mention how often users visit each service and for how long.

What about the share of total time spent by users on PSN services? Plaintiff says nothing about that metric in its Complaint. And although it seems tenable at first glance, that metric may also be of limited utility. That is because at least some of the features offered by a Facebook or

30

Instagram or Path are not, seemingly, part of those firms’ PSN-services offerings as defined by the FTC; time spent on those apps or websites, accordingly, is not necessarily time spent on a PSN service. The Commission, for instance, expressly alleges that social-networking services based on “interest-based . . . connections” such as Strava are not, by its definition, PSN services. Id., ¶ 58. That definition of what is in the market, perhaps counterintuitively to Facebook users, would mean that time a user spends engaging with specific interest-based Facebook pages or groups may not qualify as time spent on a PSN service. The same problem arises when a user “passive[ly] consum[es]” “online video” on a PSN service. Id., ¶ 59. To the extent that, say, Instagram users spend their time on the site or app watching a comedy routine posted by the official page of a famous comedian, are they spending time on a PSN service? If not, as the Complaint suggests is the case, id., then time spent “on Facebook” or “on Instagram” bears an uncertain relationship to the actual metric that would be relevant: time spent using their PSN services in particular. Put another way, the uncertainty left open by the Complaint as to exactly which features of Facebook, Instagram, et al. do and do not constitute part of their PSN services, while not necessarily rendering the alleged PSN-services market implausible, compounds the trouble created by the FTC’s vaguer-still allegations regarding Facebook’s share of that market.

Nor do the difficulties stop there. Readers may well have noticed that the discussion to this point has consistently referred to Instagram and Facebook as examples of PSN services. That is because, outside of Path, Myspace, and Friendster, all of which seem to be long defunct or quite small, see id., ¶¶ 38, 41, 153, Plaintiff’s Complaint does not identify any other providers of PSN services. Yet the FTC is apparently unwilling to allege that Facebook has ever (pre- or post-Instagram acquisition) had something like 85% or even 75% market share; instead it hedges by offering only that the number is somewhere north of 60%. The question naturally arises:

31

which firms make up the remaining 30–40%? Cf. Cupp, 310 F. Supp. 2d at 971 (“Without a[n] . . . accounting of the brands and suppliers to be included in the relevant market, the Court cannot determine [its] boundaries . . . . [and] is thus unable to assess Defendants’ market power.”); Total Benefits Plan. Agency, Inc. v. Anthem Blue Cross & Blue Shield, 552 F.3d 430, 437 (6th Cir. 2008) (“Without an explanation of the other insurance companies involved, and their products and services, the court cannot determine the boundaries of the relevant product market and must dismiss the case for failure to state a claim.”). Although Plaintiff is correct that it is not required to identify every alleged competitor in its pleadings, its choice to identify essentially none is striking. Especially when combined with its refusal to offer any clue as to how it calculated its noncommittal market-share number, the Court cannot see how the Commission has “nudged [its market power] claims across the line from conceivable to plausible.” Twombly, 550 U.S. at 570. Its “complaint must [therefore] be dismissed.” Id.

The Court’s decision here does not rest on some pleading technicality or arcane feature of antitrust law. Rather, the existence of market power is at the heart of any monopolization claim. As the Supreme Court explained in Twombly, itself an antitrust case, “[A] district court must retain the power to insist upon some specificity in pleading before allowing a potentially massive factual controversy to proceed.” Id. at 558 (citations omitted). Here, this Court must exercise that power. The FTC’s Complaint says almost nothing concrete on the key question of how much power Facebook actually had, and still has, in a properly defined antitrust product market. It is almost as if the agency expects the Court to simply nod to the conventional wisdom that Facebook is a monopolist. After all, no one who hears the title of the 2010 film “The Social Network” wonders which company it is about. Yet, whatever it may mean to the public, “monopoly power” is a term of art under federal law with a precise economic meaning: the

32

power to profitably raise prices or exclude competition in a properly defined market. To merely allege that a defendant firm has somewhere over 60% share of an unusual, nonintuitive product market — the confines of which are only somewhat fleshed out and the players within which remain almost entirely unspecified — is not enough. The FTC has therefore fallen short of its pleading burden.

That said, because it believes that the agency may be able to “cure [these] deficiencies”

by repleading, Foman v. Davis, 371 U.S. 178, 182 (1962), the Court will dismiss without prejudice only the Complaint, not the entire case, leaving Plaintiff “free to amend [its] pleading and continue the litigation.” Ciralsky, 355 F.3d at 666 (citation omitted) (explaining that dismissal without prejudice of the complaint, as opposed to the case, is not final). Whether and how the agency chooses to do so is up to it.

***

While the court could end here, it believes it profitable to provide some guidance to the parties should the case proceed. It thus offers some further legal analysis on two issues raised by the FTC’s allegations. For ease of organization, that discussion will “analyze the various” components of the agency’s Section 2 claim — namely, the Platform-policies component and the acquisitions component — “individually,” even though Plaintiff technically frames its challenge as an attack on Facebook’s overall course of anticompetitive conduct. See City of Groton v. Connecticut Light & Power Co., 662 F.2d 921, 928–29 (2d Cir. 1981). It begins by explaining why, in accord with its analysis in the parallel State case, even assuming that the FTC can establish market power here, its allegations as to Facebook’s Platform-related conduct cannot be a basis for injunctive relief under Section 13(b). It then offers the agency some consolation:

33

assuming again that market power can be demonstrated, Section 13(b) does, pace Facebook, allow the FTC to seek injunctive relief based on the Instagram and WhatsApp acquisitions.

B. Platform Policies By way of refresher, Plaintiff alleges that Facebook adopted and enforced a number of anticompetitive policies governing the use of its APIs. Most prominently, in 2013 it announced a policy of refusing to allow third-party, freestanding apps (like the chess app or the Washington Post app discussed above) to access those APIs if they “replicate[d] [Facebook’s] core functionality” — i.e., if they competed with Facebook Blue. See Redacted Compl., ¶¶ 136, 143. Facebook then enforced that policy against a number of freestanding apps, revoking API permissions after having previously offering them access. Id., ¶¶ 152–56. The FTC, like the States in the parallel action, see Mem. Op., No. 20-3589 at 22, contends that these actions represent unlawful “conditional dealing” or unlawful “refusal[s] to deal” with apps that had their API access revoked or blocked. See FTC Opp. at 31, 33.

This Court, however, agrees with Facebook’s bottom-line contention that neither theory offers a viable basis for an injunction under Section 13(b) of the FTC Act. That result follows from three conclusions. First, under current antitrust doctrine, Facebook’s general policy of refusing to provide API access to its competitors does not itself violate Section 2. Second, although specific instances in which Facebook revoked a competitor’s API permissions (after previously providing it access) might have violated Section 2, the last alleged instance occurred in 2013, so there is no way in which Facebook “is violating” or “is about to violate” the antitrust laws, which is a necessary condition for an injunction under Section 13(b). See 15 U.S.C. § 53(b). Third, and more prosaically, Plaintiff has failed to plead facts to support a “conditional dealing” theory. The Court looks at the first two together and then moves to the third.

34

For those who will consider this Opinion alongside the Court’s companion Opinion in the State case, the following “legal framework” section is the same in both. The analysis that follows diverges somewhat given the different legal regimes governing suits by States and suits by the FTC, although the outcome is similar.

1. Refusal to Deal

a. Legal Framework

The central principle that governs refusal-to-deal claims is that, as a general matter, a monopolist has “the right to refuse to deal with other firms,” which includes the right to “refus[e] to cooperate with rivals.” Verizon Commc’ns Inc. v. L. Offs. of Curtis V. Trinko, LLP, 540 U.S. 398, 408 (2004) (citing Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585, 601 (1985)). That is because “[m]onopolists are both expected and permitted to compete like any other firm,” and “[p]art of competing like everyone else is the ability to make decisions about with whom and on what terms one will deal.” Viamedia, Inc. v. Comcast Corp., 951 F.3d 429, 454 (7th Cir. 2020) (citations omitted). This general no-duty-to-deal rule holds even where a monopolist refuses to deal with its competitor merely “in order to limit entry,” Trinko, 540 U.S. at 407 — in other words, because it wants to prevent that rival from competing with it. See, e.g., Olympia Equip. Leasing Co. v. W. Union Tel. Co., 797 F.2d 370, 375–76 (7th Cir. 1986) (“Today it is clear that a firm with lawful monopoly power has no general duty to help its competitors” and thus no duty “to extend a helping hand to new entrants [or] help [rivals] survive or expand . . . .”); Novell, Inc. v. Microsoft Corp., 731 F.3d 1064, 1074 (10th Cir. 2013) (“Even a monopolist generally has no duty to share . . . its intellectual or physical property with a rival.”); FTC v. Qualcomm Inc., 969 F.3d 974, 993 (9th Cir. 2020) (“As the Supreme Court has repeatedly emphasized, there is no duty to deal under the terms and conditions preferred by a

35

competitor’s rivals.”) (quoting Pac. Bell Tel. Co. v. Linkline Commc’ns, Inc., 555 U.S. 438, 457 (2009)); Aerotec Int’l, Inc. v. Honeywell Int’l, Inc., 836 F.3d 1171, 1184 (9th Cir. 2016) (rejecting argument that refusal to deal is unlawful because it was motivated by “intent to foreclose competition”).

These decisions, to be clear, “are not premised on the view that [monopolist refusals to deal] are incapable of harming competition”; obviously, “refusals to aid new entrants can indeed” have that effect. See Daniel A. Crane, Does Monopoly Broth Make Bad Soup?, 76 Antitrust L.J. 663, 669 (2010). Rather, the rule declaring unilateral refusals to deal essentially “per se lawful,” id. at 666, or “presumptive[ly] legal[],” Novell, 731 F.3d at 1073, rests on three overriding considerations of antitrust policy. First, and most importantly, “[f]irms may acquire monopoly power by establishing an infrastructure that renders them uniquely suited to serve their customers. Compelling such firms to share the source of their advantage is in some tension with the underlying purpose of antitrust law, since it may lessen the incentive for the monopolist, the rival, or both to invest in those economically beneficial facilities.” Trinko, 540 U.S. at 407–08. Put another way, already large and successful firms “might be deterred from investing, innovating, or expanding . . . with the knowledge [that] anything [they] creat[e] [they] could be forced to share,” while “smaller [competitors] might be [similarly] deterred, knowing [they] could just demand the right to piggyback on [their] larger rival.” Novell, 731 F.3d 1073. That equilibrium would hinder, rather than advance, consumer welfare. See Olympia Equip. Leasing, 797 F.2d at 379 (“Consumers would be worse off if a firm with monopoly power had a duty to extend positive assistance to new entrants, or having extended it voluntarily a duty to continue it indefinitely.”). Second, “compelled sharing puts federal courts in the role of central planners,” requiring them to pick and choose the applicable terms and conditions of the forced sharing they

36

would order “despite their being ill-equipped to assume this role.” Aerotec, 836 F.3d at 1183. Finally, “compelled sharing may actually provide opportunities for collusion” between the monopolist and its rival or rivals. Id. Collusion is “the supreme evil of antitrust,” Trinko, 540 U.S. at 408, and itself quite “injuri[ous] to consumers and the competitive process alike.” Novell, 731 F.3d at 1073.

Nevertheless, the general no-duty-to-deal rule does have a “narrow-eyed needle” of an exception, id., at 1074, traceable to the Supreme Court’s decision in Aspen Skiing. In that case, the defendant owned three of the four ski resorts in the Aspen, Colorado, market (making it a monopolist), and it had long operated a joint venture with the fourth (a rival mountain) that provided a ticket good for all four. See 472 U.S. at 588–92. That joint ticket was profitable for both parties and made customers happy. Id. at 610. The defendant Aspen Skiing, however, eventually terminated the joint ticket. Plaintiff Highlands “tried a variety of increasingly desperate measures to re-create the joint ticket,” including eventually “offering to buy the defendant’s tickets at retail price.” Trinko, 540 U.S. at 408–09. Aspen Skiing refused to even accept those tickets at its resorts. According to the Court, based on these facts, the jury (which had found liability) could have reasonably concluded that Aspen Skiing “was not motivated by efficiency concerns and . . . was [instead] willing to sacrifice short-run benefits and consumer goodwill in exchange for a perceived long-run impact on its smaller rival.” 472 U.S. at 610–11. That, the Court held, was sufficient for Section 2 liability.

As the case has come to be understood, Aspen Skiing ran afoul of the Sherman Act —

despite the general no-duty-to-deal rule — because it acted in a predatory fashion, which is to say it deliberately harmed itself (and consumers) in order to harm its competitor more. Consider the analogy of a predatory pricing scheme, in which a firm prices its goods below its costs and

37

below its rivals’ costs with the goal of driving those rivals out of the market and leaving it standing alone; such a scheme works where the predatory firm takes advantage of its greater ability to withstand the losses caused by its below-cost pricing. See Novell, 731 F.3d at 1075 (drawing this comparison); Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 224 (1993) (explaining predatory-pricing doctrine). Aspen Skiing was doing something very similar. By terminating a profitable and customer-pleasing joint venture, and refusing to even recreate it by accepting vouchers bought at retail price, it was enduring short-term pain with the exclusive goal of driving its rival (which could not stand that pain) out of the market. The defendant thus lost the usual protection of the general no-duty-to-deal rule, since it had used its general right to choose with whom it deals “as part of a larger anticompetitive enterprise” — i.e., an enterprise aimed at “harming competition,” and therefore consumers, “by entrenching a dominant firm and enabling it to extract monopoly rents once the competitor is killed off.” Viamedia, 951 F.3d at 462 (quoting Novell, 731 F.3d at 1075) (cleaned up); see also Areeda & Hovenkamp vol. IIIB, ¶ 772d3, at 235.

The “larger anticompetitive enterprise” that characterizes an Aspen Skiing violation, crucially, cannot simply be an intent to harm — or, the flip side of the same coin, to avoid helping — a rival or rivals. “In concentrated markets the intent to maintain or improve one’s own market position always entails the knowledge that rivals must suffer,” and if antitrust law were to condemn every action taken by a monopolist with the “effect or purpose of limiting competition with itself,” there would be little left to the rule that monopolists have no duty to aid their competitors. See Areeda & Hovenkamp, ¶ 772c2, at 220–21. That is why the Supreme Court has described Aspen Skiing as “at or near the outer boundary of § 2 liability,” Trinko, 540 U.S. at 399, and “refused to extend liability to various other refusal to deal scenarios.” Novell,

38

731 F.3d at 1074; but see Steward Health Care Sys., LLC v. Blue Cross & Blue Shield of Rhode Island, 311 F. Supp. 3d 468, 483 (D.R.I. 2018) (arguing that Trinko should not be read as “pulling back the reins on refusal-to-deal claims”). And that is why the touchstone for liability is the sort of predatory design described above: “where the only conceivable rationale or purpose” for the refusal to deal is “to sacrifice short-term benefits in order to obtain higher profits in the long run from the exclusion of competition.” Aerotec, 836 F.3d at 1184 (citation omitted).

Courts have accordingly coalesced around a three-part test for unlawful refusals to deal, drawn from Aspen Skiing’s particular facts and aimed at sniffing out predation while avoiding over-inclusiveness. First, “there must be a preexisting voluntary and presumably profitable course of dealing between the monopolist and rival” with which the monopolist later refuses to deal. Novell, 731 F.3d at 1074; Qualcomm, 969 F.3d at 993. It was the absence of this preexisting relationship that doomed the plaintiffs in Trinko, as the Court upheld the dismissal of their Aspen Skiing claim at the pleading stage on that basis, despite their allegations that Verizon’s refusal to deal with rivals was intended to, and did, blunt their growth. See 540 U.S. at 404–05, 407, 409. Second, “the refusal to deal [must] involve[] products that the defendant already sells in the existing market to other similarly situated customers.” Qualcomm, 969 F.3d at 994; see Areeda & Hovenkamp, ¶ 773d3, at 233 (defendant must be “selling [the] particular product or service to others but refu[sing] to sell that same product or service to” the rival) (citing Trinko, 540 U.S. at 410). Third, and most importantly, “the monopolist’s discontinuation of the preexisting course of dealing must suggest a willingness to forsake short-term profits to achieve an anti-competitive end,” Novell, 731 F.3d at 1075; see Qualcomm, 969 F.3d at 993, rather than to advance a valid business purpose. See also Covad Commc’ns Co. v. Bell Atl. Corp., 398 F.3d 666, 675 (D.C. Cir. 2005) (to prevail on refusal-to-deal claim, plaintiff “will

39

have to prove [that defendant’s] refusal to deal caused [defendant] short-term economic loss”). The cases, and the leading treatise, offer a demanding gloss on this prong of the test: Qualcomm requires the predatory motivation to be “the only conceivable rationale or purpose” for the otherwise inexplicable profit sacrifice, see 969 F.3d at 993; Novell asks whether the cessation of dealing was “irrational but for its tendency to harm competition,” 731 F.3d at 1076; and Areeda & Hovenkamp look for “an unexplained, apparently irrational change in an established course of dealing.” Areeda & Hovenkamp, ¶ 772d3, at 233.

b. Application

i. Facebook Policies

Applying these principles, it is clear off the bat that Facebook’s adoption of a policy of not offering API access to competitors did not, standing alone, violate Section 2. As set out above, a monopolist has no duty to deal with its competitors, and a refusal to do so is generally lawful even if it is motivated, as Verizon’s was in Trinko, by a desire “to limit entry” by new firms or impede the growth of existing ones. See 540 U.S. at 407. It follows that a firm’s merely announcing its choice not to deal with competitors, whatever the motivation for doing so, cannot violate Section 2. The FTC’s core argument for why the policies themselves are unlawful — that their promulgation was intended to, and did, “change[] the incentives of third-party apps that relied upon the Facebook ecosystem [and thus] deterr[ed] them from including features and functionalities that might compete with Facebook,” Redacted Compl., ¶ 25; see also id., ¶¶ 77, 137 — misses the boat. The central teaching of the cases discussed above is that Facebook had no antitrust duty to avoid creating that deterrent. See, e.g., Olympia Equipment Leasing, 797 F.2d at 375 (“Today it is clear that a firm with lawful monopoly power has no general duty to help its competitors . . . by . . . pulling its competitive punches.”).

40

Facebook’s general policy of withholding API access from competitors, moreover, was plainly lawful to the extent it covered rivals with which it had no previous, voluntary course of dealing. (Other rivals are addressed in Section B.1.b.ii., infra.) Such prior history of dealing, recall, is a necessary element of an Aspen Skiing claim. That is yet another reason why the mere act of announcing or maintaining a general no-dealing-with-competitors policy cannot, in and of itself, violate Section 2; rather, the analysis must focus on particular acts. See Areeda & Hovenkamp, ¶ 773e, at 273 (explaining that Aspen Skiing violations are “visible and idiosyncratic event[s]”). Consider an example from the Complaint: citing its policies, in early 2013 Facebook blocked the API access of Vine, a new app it viewed as a competitor, mere hours after its launch. See Redacted Compl., ¶ 155. That decision was plainly lawful, per Trinko, because it was prospective: Facebook had not previously allowed Vine to access its APIs. Although the company was enforcing its “replicating core functionality” general policy against Vine, that fact makes no difference to the Aspen Skiing analysis; rather, whether the specific refusal to deal with Vine contravened Section 2 depends on the details of the refusal itself.

To be clear, it is possible that were a monopolist to embark on a concerted scheme of serial refusals to deal with rivals, that scheme or “course of conduct” could amount to a separate and independent violation of the Sherman Act. See Microsoft., 253 F.3d at 78 (noting course-of- conduct theory without passing on its validity). Even if such a claim were cognizable, however, simply maintaining a policy of refusing to deal with rivals would not be enough. Rather, to be actionable, an unlawful refusal-to-deal scheme would have to be made up of refusals that were themselves independent violations of the Aspen Skiing test. See Simon & Simon, PC v. Align Tech., Inc., No. 19-506, 2020 WL 1975139, at *8 (D. Del. Apr. 24, 2020) (“Plaintiff’s characterization of [the defendant’s] otherwise non-actionable refusals to deal as a ‘scheme’

41

do[es] not save its claims.”); Eatoni Ergonomics, Inc. v. Rsch. In Motion Corp., 826 F. Supp. 2d 705, 709–10 (S.D.N.Y. 2011) (“[Plaintiff] does not, and cannot, cite any authority for the proposition that a series of unilateral acts that do not violate the antitrust laws may be aggregated into an unlawful ‘course of conduct.’”), aff’d, 486 F. App’x 186 (2d Cir. 2012); Masimo Corp. v. Tyco Health Care Grp., L.P., No. 02-4770, 2004 WL 5907538, at *5–6 (C.D. Cal. June 10, 2004) (explaining that a course-of-conduct liability theory cannot “allow for clearly legal acts to be thrown into the mix to bolster a plaintiff’s antitrust case”). Otherwise, as several scholars have persuasively explained, monopolists might face liability for refusals to deal that are categorically protected from scrutiny for the policy reasons explained above. See Crane, 76 Antitrust L.J. at 666–69; Douglas H. Ginsburg & Koren Wong-Ervin, Challenging Consummated Mergers Under Section 2 (George Mason Univ. L. & Econ. Paper No. 20-14, May 2020), https://bit.ly/3wPRpnx. The FTC, therefore, cannot get anywhere by reframing Facebook’s adoption of a policy of refusing to deal with all competitors as the execution of an unlawful scheme of monopoly maintenance. Rather, to be actionable, such a scheme must involve specific instances in which that policy was enforced (i) against a rival with which the monopolist had a previous course of dealing; (ii) while the monopolist kept dealing with others in the market; (iii) at a short-term profit loss, with no conceivable rationale other than driving a competitor out of business in the long run.

ii. Specific Refusals

The FTC, to be sure, has alleged several specific refusals to deal that in fact may meet those requirements. All such pleaded instances, however, took place in 2013. See Redacted Compl., ¶¶ 153–56. That fact is fatal to the agency’s claim for injunctive relief under Section

42

13(b) here, regardless of whether those specific refusals to deal ran afoul of Aspen Skiing, because it means that no actionable violation is either ongoing or about to occur.

Section 13(b) allows the FTC to “bring suit in a district court of the United States to enjoin” allegedly unlawful conduct only where it has “reason to believe . . . that any person, partnership, or corporation is violating, or is about to violate, any provision of law enforced by the [FTC].” 15 U.S.C. § 53(b). All agree that Section 2 of the Sherman Act is such a provision. As the Supreme Court has recently explained, given that provision’s requirement that the defendant “‘is violating’ or ‘is about to violate’ (not ‘has violated’)” the antitrust laws, it contemplates “relief that is prospective, not retrospective.” AMG Cap. Mgmt., LLC v. FTC, 141 S. Ct. 1341, 1348 (2021); see also FTC v. Credit Bureau Ctr., LLC, 937 F.3d 764, 774 (7th Cir. 2019) (“Section 13(b) serves a . . . forward-facing role: enjoining ongoing and imminent future violations.”); FTC v. Evans Prod. Co., 775 F.2d 1084, 1087 (9th Cir. 1985) (holding that Section 13(b) “contemplate[s] ongoing or future violations” and thus “an injunction will issue only if the wrongs are ongoing or likely to recur”). As one court has put it, Section 13(b) therefore “does not permit the FTC to bring a claim based on long-past conduct without some evidence that the defendant ‘is’ committing or ‘is about to’ commit another violation.” Shire ViroPharma, 917 F.3d at 156.

This constraint on Section 13(b)’s scope does not, it is worth noting, leave antitrust violations beyond the reach of the federal civil authorities. For one thing, the Department of Justice has broad authority to seek injunctive relief in antitrust cases. See 15 U.S.C. §§ 4, 25. And as to the FTC itself, “ever since [its] creation in 1914, it has been authorized to enforce the [law] through its own administrative proceedings” under Section 5(b) of the FTC Act. AMG Cap. Mgmt., 141 S. Ct. at 1346. Section 5(b) allows the agency to conduct such a proceeding

43

against “person [who] has been or is” violating the antitrust laws. See 15 U.S.C. § 45(b) (emphasis added). If it goes that route and prevails, the agency can then seek a court’s help in enforcing its orders, including via monetary penalties, 15 U.S.C. § 57b, “mandatory injunctions[,] and such other further and equitable relief as [the court] deem[s] appropriate.” Id. § 45(l). Thus, as Shire ViroPharma explained, “[i]f the FTC wants to [address] a past violation — where an entity ‘has been’ violating the law — it must use Section 5(b).” 917 F.3d at 159 (quoting 15 U.S.C. § 45(b)). Section 13(b), by contrast, mainly exists to “address[] a specific problem, namely, that of stopping seemingly unfair practices” (including antitrust violations) that are ongoing or about to occur “from taking place” until the Commission can complete an adjudication. AMG Cap. Mgmt., 141 S. Ct. at 1348.

Here, given the Court’s conclusions above, the FTC cannot use 13(b) to challenge Facebook’s Platform-related conduct because it has not pleaded that any actual Section 2 violation is ongoing or about to occur. The closest the agency comes is its allegation that, “[h]aving suspended its anticompetitive platform policies [in December 2018] in response to anticipated public scrutiny, Facebook is likely to reinstitute such policies if such scrutiny passes.” Redacted Compl., ¶ 149 (emphasis added). Even if the mere resumption of Facebook’s no-dealing-with-competitors policies were itself unlawful — which, as just explained at length, it is not — that conditional and conclusory allegation would be insufficient to establish the requisite imminence. See, e.g., Shire ViroPharma, 917 F.3d at 153, 160 (finding allegation that “[a]bsent an injunction, there is a cognizable danger that [defendant] will engage in similar conduct” because it had the “incentive and opportunity” to do so “woefully inadequate to state a claim under Section 13(b)”). There are no facts alleged, moreover, suggesting that the antitrust

44

“scrutiny” the company is facing is “about to” pass or indeed will pass at any time in the foreseeable future. Indeed, a quick glance at any newspaper yields the contrary conclusion.

The more fundamental problem, of course, is that Facebook would not only have to imminently reinstate its policies, but it would also have to imminently take the kind of further action that might come within Aspen Skiing: target its competitors with which it had a previous, voluntary course of dealing for API revocation in a manner that suggests predatory, short-term- profit-sacrificing behavior. The FTC has certainly not alleged that that specific sort of conduct is “about to” occur. Nor could it since, according to the Complaint, Facebook has not actually taken such an action in nearly eight years. See Shire ViroPharma, 917 F.3d at 160 (finding inadequate allegation that defendant would likely resume conduct where “complaint contain[ed] no allegations that [it] had engaged in [that behavior] in the five-year gap between the 2012 cessation . . . and the 2017 lawsuit”); cf. FTC v. AdvoCare Int’l, L.P., No. 19-715, 2020 WL 6741968, at *5–6 (E.D. Tex. Nov. 16, 2020) (finding it “implausible” that defendant’s twenty- year scheme of “past misconduct,” which ended one year earlier, would recur and thus justify Section 13(b) injunction because the “channel[s] of misconduct [were] either permanently defunct . . . or reformed, lawful, and monitored for compliance”).

The agency counters with an argument that is surprising in its breadth. Because the statute says that it may “bring suit” for an injunction “[w]henever [it] has reason to believe . . . that any person . . . is violating, or is about to violate” the antitrust laws, see 15 U.S.C. § 53(b), the FTC contends that the law “confers discretion” on it to decide if such “reason to believe” exists. See FTC Opp. at 41. The upshot, per the agency, is that its “‘reason to believe’ determination [cannot] be reviewed” on a motion to dismiss. Id. at 39.

45

The Court finds this position wanting. The agency’s argument focuses on the statutory trigger for when it may “bring suit . . . to enjoin” antitrust violations. The question, however, is whether Section 13(b) empowers the agency, as opposed to a court, to decide whether the conditions for an injunction are actually met such that a decree should indeed issue. Whatever the law tells the agency about what it must believe before it may file suit, for the agency to actually succeed, it must convince the court to agree with it that a violation is either ongoing or about to occur and that an injunction is warranted. See Evans Products, 775 F.2d at 1087 (“[A]n injunction will issue [under Section 13(b)] only if the wrongs are ongoing or likely to recur.”). It follows that the Rule 8 requirement that the FTC must “state a claim to [the relief that it seeks] that is plausible on its face” requires the agency to make a plausible showing that the defendant “is violating” or “is about to violate” the antitrust laws. Nothing in the text of the statute suggests otherwise. See Advocare, 2020 WL 6741968, at *4 (agreeing with this approach); Shire ViroPharma, 917 F.3d at 159 n.17 (declining to address argument that FTC has “unreviewable discretion to file suit” because it was not raised below, but expressing view that it is “unpersuasive”).

Even accepting the FTC’s argument, the two cases it cites — only one of which really embraces its “internal standard argument,” FTC v. Hornbeam Special Situations, LLC, 391 F. Supp. 3d 1218, 1223 (N.D. Ga. 2019) — still appear to have required the agency to “set[] forth at least some facts to support a reasonable inference that the [unlawful] behavior will reoccur in the future in the absence of . . . injunctive relief.” Id.; see also FTC v. Vyera Pharms., LLC, 479 F. Supp. 3d 31, 44 (S.D.N.Y. 2020) (noting several facts alleged that supported a “‘reason to believe’ that the defendants [were then] engaging in violations”). Here, for the reasons given

46

above, the Court would find that Plaintiff’s Complaint does not clear even that somewhat lower bar.

In sum, then, while it is possible that Facebook’s alleged scheme of revoking API access from competitor apps could form the basis of a plausible refusal-to-deal claim under Aspen Skiing, the Court need not address that question. Even assuming the answer is yes, injunctive relief is not available now for such a claim as a matter of law.

2. Conditional Dealing

Plaintiff gets no further by maintaining that Facebook’s policies also violated antitrust rules against what they call “conditional dealing.” FTC Opp. at 31. As an initial matter, the FTC is wrong to argue that a monopolist violates that so-called doctrine whenever it “induce[s] . . . trading partners or other firms not to compete with it . . . by conditioning access to some resource of the monopolist.” Id. The main case they cite for that proposition, Lorain Journal Co. v. United States, 342 U.S. 143, 149 (1951), says nothing of the sort. That is unsurprising, as such a broadly formulated rule would cover refusals to deal with competitors, thus contradicting the cases discussed above — such refusals can always be reframed as offers to deal only on the condition that the third party refrains from competing. See, e.g., Herbert Hovenkamp, FRAND and Antitrust, 105 Cornell L. Rev. 1683, 1697 (2020) (defining “simple refusal[s] to deal” covered by Trinko rule to include “refusal[s] . . . conditioned on a firm’s status that cannot readily be changed,” such as where a firm “agree[s] to sell to [non]competitors but not []competitors”).

To the extent any scholarly commentary uses the term “conditional dealing,” rather, the phrase generally refers to actions such as “tying” or “exclusive dealing.” Id. at 1697, 1708. The key fact distinguishing such conduct from a standard refusal to deal is that it is not “unilateral,”

47

but instead “involves some assay by the monopolist into the marketplace” that interferes with the relationship between rivals and third parties. Novell, 731 F.3d at 1072. Tying, for instance, occurs when a firm “require[s] third parties to purchase a bundle of goods rather than just the ones they really want,” id., thereby leveraging the monopolist’s power in the “tying” product market to harm its competitors (who lose access to customers) in the “tied” product market. See Microsoft, 253 F.3d at 84. “Exclusive dealing” is similar: it refers to a monopolist’s conditioning the sale of a product on the buyer’s agreement not to deal with its competitors. Id. at 70–71. Again, these “conditional dealing” schemes are thus categorically different from unilateral conduct that involves only the monopolist’s competitors, such as its refusal to deal with them. The distinction is critical, as antitrust law is far more tolerant of unilateral behavior. See Novell, 731 F.3d at 1072–73 (citing cases) (“Put simply if perhaps a little too simply, today a monopolist is much more likely to be held liable for failing to leave its rivals alone than for failing to come to their aid.”); Hovenkamp, 105 Cornell L. Rev. at 1697.

That brings us back to Lorain Journal, which, as it happens, involved a “a very special form of exclusive dealing, namely, a refusal to sell to end-user customers who purchase[d] from the monopolist’s competitor[].” Kenneth L. Glazer & Abbott B. Lipsky, Jr., Unilateral Refusals to Deal Under Section 2 of the Sherman Act, 63 Antitrust L.J. 749, 800 n.75 (1995). There, the Journal, a local newspaper, had a “commanding” position in the advertising market of the city of Lorain, Ohio. See 342 U.S. at 146. After a new radio station entered that market, the paper began to refuse to sell advertising space “in the Journal [to] any Lorain County advertiser who advertised or who [the paper] believed to be about to advertise” with the radio station. Id. at 148. The paper’s bet was that many advertisers could not afford to lose out on the chance to advertise with it, and it was right; the scheme caused most firms to avoid advertising on the radio station,

48

restoring the Journal to its local monopoly. Id. at 148–49. On those facts, the Court found a violation of Section 2 of the Sherman Act. Id. at 152–53. The key point for our purposes is that the Journal did not simply refuse to deal with the competitor radio station — e.g., it did not merely refuse to print advertisements for the station itself in its pages. The Journal instead refused to deal with any and all customers unless those customers agreed not to deal with the competitor station, thereby interfering with its rival’s ability to compete on the merits.

For “Facebook’s conditioning of access to its Platform [to be] unlawful under Lorain Journal” (or exclusive-dealing doctrine generally), as Plaintiff argues that it is, see FTC Opp. at 34, the agency would thus have to allege that Facebook conditioned access to its Platform APIs on app developers’ agreeing not to deal with other social-networking services. And, indeed, the FTC’s Opposition suggests that Facebook did just that. See FTC Opp. at 32 (“Facebook . . . grant[ed] third parties full access to its Platform only on the condition that they not threaten its monopoly by [inter alia] connecting with or promoting other PSN providers.”); id. at 31 (“Facebook conditioned access to its Platform on trading partners not competing with it or assisting competitors.”) (emphasis added). If Facebook’s policies had in fact interfered in that way with the ability of competing social-networking services to deal with app developers, that could plausibly have violated Section 2; such conduct might well have had a “significant effect in preserving [Facebook’s] monopoly” by keeping user engagement with competing social- networking services “below the critical level necessary for any rival to pose a real threat to [its] market share.” United States v. Dentsply Int’l, Inc., 399 F.3d 181, 191 (3d Cir. 2005) (finding Section 2 liability based on exclusive dealing); see also Microsoft, 253 F.3d at 69–70 (same).

Plaintiff, however, has not sufficiently pleaded the sort of conduct just described. It is true that Facebook’s 2011 policy prohibited “Apps on Facebook” from “integrat[ing], link[ing]

49

to, promot[ing], distribut[ing], or redirect[ing] to any app on any other competing social platform.” Redacted Compl., ¶ 139. But, as discussed above, that policy applied only to apps designed for use within the Facebook website, not freestanding, separate apps like our chess app or Washington Post app. The policy, moreover, does not appear to have prevented an app developer from building a version of its app that could be accessed and used within the website of another PSN service. For instance, the developer who built a personality-test app for Facebook could build the same app for Google+ without running afoul of the policy, cf. id., ¶ 140; its terms simply prohibited the Facebook version of the app from linking or redirecting to the Google+ version. The 2011 policy was thus a far cry from a policy that told app developers that they could only access Facebook’s platform if they promised to only build their app for Facebook. By rough analogy, it is as if the Lorain Journal, rather than refusing to carry advertisements from any business that also advertised with the competing radio station, instead merely required that advertisements appearing in its paper had to avoid mentioning the radio station. Such a focused prohibition on the use of a monopolist’s own facilities obviously could not have “significantly limited” the “opportunities for [competitors] to enter into or remain in [the] market” for personal-social-networking services, as is required for Section 2 liability. Microsoft, 253 F.3d at 69.

The 2012 and 2013 policies mentioned in the Complaint similarly did not much limit the opportunities for Facebook competitors to deal with third-party app developers. The 2012 policy allegedly prohibited freestanding apps (i.e., not “Apps on Facebook”) from “us[ing] Facebook Platform [APIs] to export [Facebook] user data into a competing social network.” Redacted Compl., ¶ 142. The 2013 policy went further, instructing that freestanding apps could not use Facebook’s APIs “to promote, or to export [Facebook] user data to, a product or service that

50

replicates a core Facebook product or service.” Id., ¶ 143. By their terms, neither of these policies imposed anything approaching a blanket prohibition on freestanding apps’ linking to or creating interoperability with Facebook competitors. It is not as if, say, Facebook instructed the Washington Post that its app could only allow a user to sign in with Facebook credentials (as opposed to, say, Google credentials), or only add a Facebook “Like” button and not a “share” button from Path or Circle. Nor did Facebook require the app to refrain from allowing users to search for their contacts from other social-networking services, as they did for their Facebook friends using the Find Friends API. Perhaps the FTC means to hang its hat on a very capacious reading of the policy’s prohibition on using Facebook Platform to “promote” competitors, but reading that term to prohibit the sort of dealings just described is implausible at best. That is especially true given that the Complaint alleges no specific instances of Facebook’s enforcing its policies to prevent dealing between rivals and other freestanding apps, even as it is replete with allegations regarding Defendant’s enforcement of its “no dealing with competitors” policies. See id., ¶¶ 153–58. Plaintiff’s “conditional dealing” theory of a Section 2 violation, accordingly, is not viable here given the conduct that it has alleged.

In summary, the Court concludes — and advises the parties going forward — that in its view, Facebook’s “imposition and enforcement of anticompetitive conditions on access to [its] APIs,” id., ¶ 71, as that conduct has been pleaded, either does not amount to exclusionary conduct violating Section 2 of the Sherman Act or, to the extent that it might, cannot be the basis for an injunction under Section 13(b) of the FTC Act.

C. Challenging Acquisitions under Section 13(b)

Before closing, the Court addresses one further issue raised by Facebook in its Motion to Dismiss: whether, even assuming market power can be established, Section 13(b) allows the

51

FTC to attack Facebook’s acquisitions of Instagram and WhatsApp. See MTD FTC at 39. Facebook argues that because those acquisitions happened all the way back in 2012 and 2014, respectively, Plaintiff’s challenges do not contend that the company “is violating” or “is about to violate” Section 2. See 15 U.S.C. § 53(b). In other words, Defendant maintains that the FTC’s assault on its mergers is aimed at the sort of “long-past conduct” that cannot form the basis of an injunction under Section 13(b). Shire ViroPharma, 917 F.3d at 156. (Following the lead of the parties and the cases, the Court uses the terms “acquisition” and “merger” interchangeably for purposes of this analysis.)

While Facebook’s argument has some intuitive appeal, it runs into contrary precedent.

The Supreme Court has clearly stated that the term “acquisition” as used in Section 7 of the Clayton Act — which prohibits purchases “the effect of [which] may be substantially to lessen competition, or to tend to create a monopoly,” 15 U.S.C. § 18 — does not refer to “a discrete transaction but [rather] a status which continues until the transaction is undone.” United States v. ITT Cont’l Baking Co., 420 U.S. 223, 241–42 (1975). In other words, “acquisition . . . mean[s] both the purchase of rights in another company and the retention of those rights,” such that Section 7’s “ban against [certain] acquisitions . . . include[s] a ban against holding certain assets,” not just “obtaining” them in the first place. Id. (emphasis added); see also Concord Boat Corp. v. Brunswick Corp., 207 F.3d 1039, 1050 (8th Cir. 2000) (“Clayton Act claims are not limited to challenging the initial acquisition of stocks or assets . . . since holding as well as obtaining assets is potentially violative of section 7.”) (citation omitted). As ITT Continental Baking explained, that understanding is the necessary upshot of the Court’s earlier decision in United States v. E.I. du Pont, 353 U.S. 586 (1957), which held that the United States could seek divestiture in 1949 of stock that du Pont had first acquired, and was still holding, in 1919.

52

The rule under Section 7 is thus that so long as an acquiring company continues to hold acquired assets, the Government may “at any time” argue that such company is violating Section 7. Du Pont, 353 U.S. at 597. In such a case, it would follow that the company “is violating . . . [a] provision of law enforced by the [Commission]” and thus comes within the ambit of Section 13(b). See 15 U.S.C. § 53(b).

Because the FTC’s challenges to Facebook’s acquisitions here arise under Section 2 of the Sherman Act, not Section 7 of the Clayton Act, the question becomes whether that same principle applies to claims brought under Section 2. The Court sees no reason why not, nor does Facebook offer one. Just as “the Government may proceed at any time that an acquisition may be said with reasonable probability to contain a threat that it may . . . tend to create a monopoly,” in violation of Section 7’s proscription against transactions having that effect, du Pont, 353 U.S. at 597, so too it should be able to proceed against an acquisition whenever “a threat of [Section 2’s] prohibited effects is evident,” id. at 598 — namely, that the acquisition is “tend[ing] to destroy competition itself” via “means other than competition on the merits.” Microsoft, 253 F.3d at 58, 62 (citation omitted). It is well established that mergers may constitute one such “means.” See Grinnell, 384 U.S. at 576 (finding Section 2 violation based in part on a number of acquisitions of competitors); BRFHH Shreveport, LLC v. Willis Knighton Med. Ctr., 176 F. Supp. 3d 606, 622 (W.D. La. 2016) (“[A]cquisitions of viable competitors alone may establish the anticompetitive conduct element of a section 2 claim.”); Behrend v. Comcast Corp., No. 03- 6604, 2012 WL 1231794, at *20 (E.D. Pa. Apr. 12, 2012) (same); see also Areeda & Hovenkamp vol. III, ¶ 701a, at 201 (noting “overlap between Sherman Act and Clayton Act treatment of mergers,” including in “case of mergers as unlawful monopolistic acts by firms with significant market power”).

53

The Court will not speculate further as to how, if at all, the Section 2 analysis of a claim involving a long-ago merger might differ from that regarding a more recent (or even forthcoming) purchase, including on the issue of remedy. Cf. Areeda & Hovenkamp vol. II, ¶ 320g, at 375 (suggesting that “the considerations underlying the laches doctrine might well be applied against the government” in long-delayed challenges “possibly by adjustment in certain burdens of proof or the remedy”); United States v. Pullman Co., 50 F. Supp. 123, 127 (E.D. Pa. 1943) (“While the doctrine of laches does not apply against the United States, we should approach the question of equitable relief with great hesitancy in a case where the last operative act was nearly half a century old.”). Facebook has made a number of arguments on this score that need not be addressed now. At this point, the Court simply holds that, contrary to the company’s main contention, an injunction under Section 13(b) is a theoretically available remedy in a Section 2 challenge to long-ago mergers so long as the defendant still holds the purchased assets or stock, as is the case here. IV. Conclusion For the foregoing reasons, the Court will grant Facebook’s Motion to Dismiss, but it will dismiss without prejudice only the Complaint, not the case. The Court will also grant leave to amend and order Plaintiff to file any amended Complaint within thirty days. A contemporaneous Order so stating shall issue this day.

/s/ James E. Boasberg

JAMES E. BOASBERG

United States District Judge Date: June 28, 2021