from the well-this-will-be-interesting dept
Everyone knew that this was coming eventually, but on Wednesday two separate antitrust lawsuits were filed against Facebook. First, the FTC filed a complaint, followed by 48 Attorneys General, representing 46 states, the District of Columbia and Guam (Guam!), similarly arguing that Facebook’s acquisitions of Instagram and Whatsapp were an antitrust violation. I will say, upfront, that both cases appear to have a lot more meat to them than the DOJ’s astoundingly weak case against Google. And yet… I’m still somewhat surprised at some of the claims made in both lawsuits that seem somewhat disconnected from reality.
As a quick summary, though, I’d say that both lawsuits make somewhat weak claims regarding acquisitions (mainly, but not limited to) Instagram and WhatsApp. Both lawsuits, though, do make much stronger claims about Facebook abusing its API to try to limit competition (though, doing so without the context of privacy questions that may have driven Facebook to close off more access to its API). I think the API questions are the most interesting to explore, and the ones where Facebook may face the most trouble in court.
Let’s go through each one separately. The key to the FTC’s case is twofold: (1) that Facebook buys up whatever upstart competitors it can find that represent a competitive threat (e.g., Instagram & Whatsapp) and (2) that it puts in place “restrictive policies” that hinder the upstart competitors it cannot acquire. I’m not sure that either claim is fully supported by the facts, but perhaps there’s more evidence to support them. Of course, a key aspect in any antitrust case is proving (1) that there’s a market in which the company is a monopoly and (2) that the company leverages that monopoly in a manner that is abusive to competition. The FTC case argues that the “market” here is “personal social networking,” which seems like a fairly narrowly defined market:
Facebook holds monopoly power in the market for personal social networking services (?personal social networking? or ?personal social networking services?) in the United States, which it enjoys primarily through its control of the largest and most profitable social network in the world, known internally at Facebook as ?Facebook Blue,? and to much of the world simply as ?Facebook.?
In the United States, Facebook Blue has more than [REDACTED] daily users and more than [REDACTED] monthly users. No other social network of comparable scale exists in the United States.
Of course no one denies that Facebook is the largest, but does that automatically make it a monopolist? In the space of “personal social networking,” you could easily argue that there are a number of significantly sized competitors, including Twitter, Snap, YouTube, and TikTok (which, notably, is a relatively new entrant that was able to build up a large audience, despite the presence of Facebook). The complaint says that buying Instagram and WhatsApp were attempts to retain this monopoly position:
Since toppling early rival Myspace and achieving monopoly power, Facebook has turned to playing defense through anticompetitive means. After identifying two significant competitive threats to its dominant position?Instagram and WhatsApp?Facebook moved to squelch those threats by buying the companies, reflecting CEO Mark Zuckerberg?s view, expressed in a 2008 email, that ?it is better to buy than compete.? To further entrench its position, Facebook has also imposed anticompetitive conditions that restricted access to its valuable platform?conditions that Facebook personnel recognized as ?anti user[,]? ?hypocritical? in light of Facebook?s purported mission of enabling sharing, and a signal that ?we?re scared that we can?t compete on our own merits.?
I can see Instagram fitting into that model, but am a little more perplexed by the inclusion of WhatsApp, which is more about messaging than a traditional “social network.” It really feels like they’re shoehorning WhatsApp into this bucket to make the argument seem stronger. The case even admits that WhatsApp is about messaging, not social networking, but then argues that Facebook feared that WhatsApp could eventually pivot to social media. Which, sure. But… it still seems like a stretch to argue that this automatically makes it anticompetitive to buy.
If anything, this lawsuit suggests that Facebook’s management has both read and internalized Clayton Christensen’s “The Innovator’s Dilemma” and recognized that competition could come from orthogonal directions.
As for Instagram, the company was quite small — famously only 13 employees — when Facebook acquired it. It is true that it was building up some level of success, but it seems to suggest some level of revisionist thinking to say that it was somehow going to take down Facebook, and that acquiring it somehow changed Facebook’s competitive landscape in any meaningful way.
The other issue is that both the Instagram and Whatsapp decisions were reviewed at the time — and approved by the FTC and the DOJ. If anything, this all feels a bit like revisionist history to go back many years later and say “well these were obviously anti-competitive” when they certainly didn’t appear to be at the time of acquisition. And, yes, the lawsuits have quotes from people inside Facebook noting that Instagram and WhatsApp could potentially represent a competitive threat, but merely buying up some potential competitors doesn’t automatically mean that it’s anti-competitive behavior.
The FTC’s case gets a lot stronger when it talks about certain practices the company put in place to discourage and stymie other competitors:
In order to communicate with Facebook (i.e., send data to Facebook Blue, or retrieve data from Facebook Blue) third-party apps must use Facebook APIs. For many years? and continuously until a recent suspension under the glare of international antitrust and regulatory scrutiny?Facebook has made key APIs available to third-party apps only on the condition that they refrain from providing the same core functions that Facebook offers, including through Facebook Blue and Facebook Messenger, and from connecting with or promoting other social networks.
As someone who believes strongly that, for competitive reasons, Facebook should open up its APIs, this argument seems more compelling to me. There’s also a much stronger case to be made here that selectively opening up its API based on competitive rationale is more likely to have resulted in consumer harm.
By suppressing, neutralizing, and deterring the emergence and growth of personal social networking rivals, Facebook also suppresses meaningful competition for the sale of advertising. Personal social networking providers typically monetize through the sale of advertising; thus, more competition in personal social networking is also likely to mean more competition in the provision of advertising. By monopolizing personal social networking, Facebook thereby also deprives advertisers of the benefits of competition, such as lower advertising prices and increased choice, quality, and innovation related to advertising.
It feels like that argument would need a lot more evidence. While I agree that the API arguments are the strongest, the entire complaint ignores the fact that there are competitors, and that some of those competitors have sprung up in recent years, despite what it claims are recent barriers to entry. Snap and TikTok both became prominent well after it is argued that Facebook was hellbent on suppressing competition. And both are doing decently well. The complaint doesn’t even mention TikTok, and only mentions Snap to say that Facebook failed in its attempt to purchase the company (which seems to cut against the argument that Facebook was just buying up any competitors). There are some other mentions of “Snap” but, amusingly, that’s only because it was the internal code-name for Facebook’s failed attempt to compete with Instagram, prior to purchasing Instagram.
Based on all of that, I think the claims about buying up Instagram and WhatsApp are pretty weak — but there is an interesting area to explore regarding whether or not Facebook used its API access to suppress competition.
Specifically, between 2011 and 2018, Facebook made Facebook Platform, including certain commercially significant APIs, available to developers only on the condition that their apps neither competed with Facebook (including, at relevant times, by ?replicating core functionality? of Facebook Blue or Facebook Messenger), nor promoted competitors. Facebook punished apps that violated these conditions by cutting off their use of commercially significant APIs, hindering their ability to develop into stronger competitive threats to Facebook Blue.
On July 27, 2011, Facebook introduced a new policy that ?Apps on Facebook may not integrate, link to, promote, distribute, or redirect to any app on any other competing social platform.?
This policy was intended to harm the prospects for?and deter the emergence of? competition, including personal social networking competitors. Indeed, the immediate impetus for the policy was Google?s launch of the Google+ personal social network. In a July 27, 2011, email, a Facebook manager explained: ?[W]e debated this one a lot. In the absence of knowing what and how google was going to launch, it was hard to get very specific, so we tended towards something broad with the option to tighten up as approach and magnitude of the threat became clear.? Later that same day, another Facebook employee protested the anticompetitive move to colleagues: ?I think its [sic] both anti user and sends a message to the world (and probably more importantly to our employees) that we?re scared that we can?t compete on our own merits.?
To me, this is the strongest argument in the entire filing, though it doesn’t seem to be the one most people are focusing on.
One other thing highlighted in the complaint — which we’ve discussed here before — is how Facebook bought an analytics company called Onavo, which it pretended as a VPN service, but which it actually used as a tool to snoop on what apps people were using, in order to quickly identify up and coming competitive threats:
By acquiring Onavo, Facebook obtained control of data that it used to track the growth and popularity of other apps, with an eye towards identifying competitive threats for acquisition or for targeting under its anticompetitive platform policies. As a December 2013 internal slide deck noted: ?With our acquisition of Onavo, we now have insight into the most popular apps. We should use that to also help us make strategic acquisitions.? Facebook also used Onavo data to generate internal ?Early Bird? reports for Facebook executives, which focused on ?apps that are gaining prominence in the mobile eco-system in a rate or manner which makes them stand out.? Facebook shut down Onavo in 2019 following public scrutiny; however, it continues to track and evaluate potential competitive threats using other data.
Of course, looking for useful data to identify competitors isn’t, by itself, anti-competitive. It’s kind of expected. I do think there’s a potential “false advertising” claim that could be made about Facebook claiming Onavo was a VPN or was somehow designed to protect user privacy — but that’s different from an antitrust claim.
Moving on to the AGs’ case. It replicates many of the same arguments, though in many more words. Once again, it focuses on Instagram and WhatsApp, but not in very convincing ways. The issue of Facebook’s API policies is more compelling:
As part of its strategy to thwart competitive threats, Facebook pursued an open first?closed later approach in which it first opened its platform to developers so that Facebook?s user base would grow and users would engage more deeply on Facebook by using third-party services. This strategy significantly boosted engagement on Facebook, enhanced the data it collected, and made the company?s advertising business even more profitable. Later, however, when some of those third-party services appeared to present competitive threats to Facebook?s monopoly, Facebook changed its practices and policies to close the application programming interfaces (?APIs?) on which those services relied, and it took additional actions to degrade and suppress the quality of their interconnections with Facebook.
Again, I am in agreement that this is the behavior by Facebook that is most concerning — but it also leaves out some important context. Namely, that Facebook was pressured to close down access to its APIs because many people (including some of the AGs who are part of this lawsuit) were making noises about the privacy concerns related to Facebook’s more open API access (as noted in the Cambridge Analytica scandal). One of the more frustrating things to me about all of this is how few people are willing to grapple with the tradeoffs here. If you complain about privacy issues it pushes Facebook to close off access to its APIs, which give it more anticompetitive power. But if you demand more open APIs, then you have more privacy questions.
The AG’s case, though, does make the (compelling) argument that Facebook used its more open API access setup to build growth, and then slowly cut back on it to block competitors:
One critical reason for Facebook?s accelerated growth trajectory was a series of initiatives that opened Facebook up to mutually beneficial partnerships with third parties.
In 2007, Facebook launched Facebook Platform?an innovative tool that set it apart from other firms. Facebook Platform had a set of open APIs?mechanisms for sharing data between independent services?that enabled developers to build applications that interoperated with the Facebook social networking site. Developers scrambled to create applications on the Facebook Platform, enhancing Facebook Blue?s functionality, driving more users to the Facebook site, and increasing the engagement of existing users. Facebook Platform created a symbiotic relationship between Facebook and developers that yielded significant value for both.
In 2008, Facebook introduced Facebook Connect?a tool that facilitated still greater interconnection with Facebook Blue. Through Facebook Connect, users could sign in to third-party websites using their Facebook credentials. By 2011, Facebook Connect had become one of the most popular ways to sign in to services across the internet as users took advantage of the efficiency afforded to them. As was the case with the Facebook Platform, third-party sites and Facebook itself found the relationship fostered by Facebook Connect to be a mutually valuable one. Facebook provided third parties with information about users and their friends and drove traffic to third-party sites by making it easier for users to sign in. In return, Facebook captured valuable data about users? off-Facebook activity to enhance its social graph and ability to target advertising.
In April 2010, Facebook invited even more interaction with third-party websites and apps. It launched the Open Graph API, enabling those sites to add plug-ins, such as the Facebook ?Like? button that allowed Facebook Blue users to become ?fans? of the third-party site. The sites were highly motivated to install the Like button and encourage its use, as a ?Like? would be shared on the user?s news feed and profile, thereby promoting the site to the user?s friends and family. One week after the introduction of Open Graph, 50,000 websites had installed Open Graph plug-ins. Those sites realized the immediate benefits of a massive new distribution channel, and Facebook?s growth increased accordingly.
But then when upstarts began using the openness to build up more competitive power, Facebook cut back on the access:
After the threat from Google+ had passed, and after years of promoting open access to Facebook Platform, Facebook increasingly turned to Platform as a tool to monitor, leverage, and harm (via rescinding API access) apps that Facebook viewed as actual or potential competitive threats.
In 2013, Facebook amended its Platform policy (described above) to forbid applications that ?replicat[e] [Facebook?s] core functionality,? with no explanation as to what Facebook considered its core functionality, or how such policies would apply when Facebook expanded its functionality to a new area.
Facebook began to selectively enforce its policies to cut off API access to companies Facebook worried might one day threaten its monopoly. Facebook itself described its Platform as a ?critical piece of infrastructure? for new apps being developed: this is particularly true for social apps which rely heavily on network effects. Facebook knew that an abrupt termination of established access to Facebook APIs could be devastating to an app?especially one still relatively new in the market. An app that suddenly lost access to Facebook?s APIs was hurt not only because its users would no longer be able to bring their friend list to the new app, but also because a sudden loss of functionality, which creates broken or buggy features, suggests to users that an app is unstable. Facebook?s actions therefore disincentivized developers from creating new features that might compete with Facebook: adding new social features to an existing app might come at the significant cost of access to Facebook?s APIs.
In 2014, Facebook announced significant changes to its Platform APIs with ?Graph API 2.0? (also referred to internally as Platform 3.0). In connection with Graph API 2.0, Facebook required prior review of all requests to access many Platform APIs, including the Find Friends API, resulting in those APIs being cut off for third-party apps that previously had enjoyed such access on Platform. Under the new approach, developers could not access Facebook?s APIs unless they submitted an application and received approval, which Facebook refused to grant to apps it classified as competitors or potential competitors. That allowed Facebook to proactively and categorically ensure that no app that might constitute a competitive threat would get API access in the first place, sparing Facebook the need to withdraw access after-the-fact.
One Facebook employee described in January 2014 the jarring impact of enticing developers to interconnect and then suddenly revoking their access:
We sold developers a bill of goods around [Open Graph] 2 years ago and have been telling them ever since that one of the best things they could do is to a/b test and optimize the content and creative. Now that we have successes . . . in 2013, we?re talking about taking it away . . . . Even if we were to give them more traffic on home page in some other way, it still nullifies all of their work to integrate [Open Graph] for the last 2 years.
As a further part of its scheme to maintain its monopoly position, Facebook has used its control over Facebook Platform to degrade the functionality and distribution of potential rivals? content when it perceived those firms as threats to Facebook?s monopoly power. This degradation suppressed the flow of user traffic to the rivals? services, reducing overall output, and harming users in the process.
Again, this entire narrative leaves out some of the privacy questions raised by the open platform access — but I do think that these are the strongest arguments in the complaint.
There are some mostly-redacted stories included in the complaint, arguing that certain companies, whose names are redacted, were directly harmed by Facebook’s API policies. It would be interesting to see who these companies are, because I’ve seen some somewhat scammy companies make the argument that Facebook’s API policies were designed to harm them, but then the details suggest that the reality was more that Facebook was trying to stop them from being scammy. The details matter here.
There are some examples with named companies, including Twitter’s Vine, Path, and Circle. Some of those stories seem somewhat compelling, though I’m not sure they move the needle enough.
In the end, I think these complaints are stronger than the DOJ’s complaint against Google. That’s not to say that they are particularly strong. I do think the most compelling arguments are entirely around Facebook’s use of the API, though the company will have arguments it can make in response regarding reasons why it made those decisions (with a focus on protecting user privacy). Whether or not that will be enough to hold up will likely be a key part of this case. The arguments about Instagram and WhatsApp, while they’ll get most of the public attention, and are the arguments that people frequently make about Facebook’s position, strike me as fairly weak. Both acquisitions were reviewed and approved at the time, and the presence of other competitors in the marketplace seems to undercut the arguments there.
The key question that I think the courts will look at is whether or not Facebook pulling back on API access trips an antitrust wire. Antitrust experts may have better suggestions for cases to look at, but the Aspen Skiing case appears to be roughly analogous — in which the question was whether or not a dominant firm terminating an agreement to deal with a competitor was an antitrust violation. However, subsequent cases have shown that the Supreme Court no longer seems to support that notion. Verizon v. Trinko likely cut back the Aspen Skiing rules significantly — and the courts would need to signal a pretty big shift to bring those rules back around for Facebook.
As for remedies, if Facebook were forced to spin off Instagram and WhatsApp, I honestly don’t think it would have that big of an impact on Facebook itself. I’m not against that being the final decision, but I don’t see how that would solve any of the issues at play. Either way, Facebook is going to be very busy in court for the next decade or so….