Net Neutrality: FCC Gatekeeper Theory Applies to Google, Apple, and Netflix

This is the second post in a series addressing fundamental questions presented by the prospect of applying per se net neutrality rules under Title II. The first post is available HERE.

The first post in this series concluded that the logic of the gatekeeper theory the FCC used to justify the imposition of per se net neutrality rules extends to any Internet intermediary that is capable of blocking, degrading, or favoring particular Internet services, applications, or content. This post presents a brief analysis of some intermediary services to which the gatekeeper theory would apply if the FCC relies on it to impose per se net neutrality rules under Title II.

The analysis demonstrates that Internet companies must also ‘ask permission’ to pass gates erected by Google, Apple, and Netflix (non-ISP gatekeepers) in order to obtain access to end users:

  • These non-ISP gatekeepers routinely use their gatekeeper control to block, degrade, or discriminate against upstream edge providers (far edge providers);
  • End users may incur significant costs in switching from one non-ISP gatekeeper to another; and
  • These non-ISP gatekeepers provide services in market segments that are highly concentrated.

It would thus be arbitrary and capricious for the FCC to impose per se net neutrality obligations only on ISPs under Title II.

To be clear, I’m not suggesting that reclassification of broadband Internet access services as ‘telecommunications services’ under Title II is necessary or that the FCC should otherwise regulate non-ISP gatekeepers. The analysis in the post is intended to illustrate that the FCC’s theory of gatekeeper control is radically over-broad under Title II and inconsistent with the competition theory of communications regulation set forth by Congress in the Communications Act.

It Would Be Arbitrary and Capricious to Limit the Gatekeeper Theory to ISPs Under Title II

The Open Internet Order’s decision limiting its gatekeeper theory to ISPs was arbitrary and capricious. Without analysis, the FCC concluded — in a footnote — that ISPs are “distinguishable from other participants in the Internet marketplace” because ISPs are “capable of blocking, degrading, or favoring any Internet traffic that flows to or from a particular subscriber.” This purported distinction is no distinction at all with respect to intermediary platforms that possess the same or similar capabilities as ISPs, including:

  • Mobile operating systems (OSs);
  • Search engines; and
  • Online video distributors.

Each of these services is a ‘non-ISP gatekeeper’ because it (1) operates as an intermediary service between end users and far edge providers, and (2) is “capable of blocking, degrading, or favoring” far edge services with respect to end users who choose that particular intermediary to access far edge services.

Some non-ISP gatekeepers are capable of exercising the same degree of control over access to end users as ISPs — e.g., mobile operating systems are capable of blocking, degrading, or favoring all Internet traffic that flows to or from a particular subscriber on the mobile Internet. Other non-ISP gatekeepers exercise control only with respect to particular types of Internet apps, content, or services — e.g., online video distributors, who are capable of blocking, degrading, or favoring video content.

From the perspective of a far edge provider that relies on non-ISP intermediary services to reach end users, however, this is a distinction without a difference. As explained in the first post in this series, the gatekeeper theory presumes that all edge providers must have access to all end users on all Internet platforms at all times in order to survive, innovate, grow, and compete (subject to ‘reasonable network management’).

Net neutrality proponents presume the same. They have traditionally opposed the notion that gatekeepers should be permitted to block, degrade, or discriminate against Internet traffic so long as their behavior is limited to a particular type of traffic or a particular market segment. For example, many net neutrality proponents oppose the FCC exemption for ‘specialized services’ and maintain that wireless and wired networks should be subject to the same net neutrality rules.

The fact that a particular intermediary — ISP or otherwise — can only block, degrade, or discriminate against a particular type of Internet traffic is irrelevant under the gatekeeper theory espoused by net neutrality advocates and adopted by the FCC in the Open Internet Order.

Google and Apple Use Their Gatekeeper Control Over Mobile Operating Systems to Block, Degrade, and Discriminate Against Far Edge Providers

It should be obvious that the providers of mobile operating systems are gatekeepers under the FCC’s theory. End users cannot access the Internet without devices (also known as ‘customer premises equipment’ or ‘CPE’), and devices cannot access the Internet without operating systems.

It should also be obvious that OS service providers routinely user their gatekeeper control over CPE to block, degrade, or discriminate against far edge providers in a multitude of ways, including, but not limited to:

  • Restricting access to application programming interfaces (APIs),
  • Restricting the placement or use of apps and other far edge services on mobile devices,
  • Restricting the manufacture of devices with competitive operating systems, and
  • Blocking far edge apps, services, and Internet traffic outright.

As Ars Technical recently reported, “The second you try to take Android and do something that Google doesn’t approve of, it will bring the world crashing down upon you.”’ Though the original source code for Android is still open source, Android is no longer ‘open’. Google began closing certain Android components in 2012, and once closed, all development on the open-source equivalent stops. ‘Google Play Services’ is a closed-source system-level component that provides APIs for Google apps and third-party developers that license access to them. Once an app is developed using Google’s proprietary APIs, however, it won’t run on the Kindle OS or any other non-Google version of Android — which makes it more challenging for third-party developers to create apps for the Kindle OS and other rival operating systems.

Google also uses its gatekeeper control over Android to discriminate against rival app stores. A competition complaint related to Android was filed against Google this summer in Europe, where Android has a 72% smartphone market share. According to Aptoide, an independent Android app store, Google is using multiple strategies to make it difficult or impossible for Android users to choose alternative app stores, including:

  • Blocking fully-functional third-party app stores from competing in Google’s Play Store,
  • Degrading the ability to install apps from third-party sources,
  • Bundling the Play Store with Google Mobile Services, and
  • Blocking access to install pages for third-party apps in Google’s Chrome browser.

Finally, Google uses its control over the licensing of its own apps for use with Android to discriminate against rival apps and operating systems. According to Ars Technica, “Google’s real power in mobile comes from control of the Google apps . . . [which] need to be licensed from Google.” Those licenses come with strings attached.

A class action lawsuit was recently filed against Google for tying its licensing agreements to a requirement that manufacturers restrict competing apps (e.g., Microsoft’s Bing) on Android devices. Google’s licensing deals with Samsung and HTC require them to set Google as the default mobile search engine and preload a suite of Google services, including YouTube and Google Maps, on “prime screen real estate” on their Android-equipped devices. Google also has a deal with Apple to be the default search engine on iPhones and iPads. No wonder Google has a 91% share of the mobile search engine market.

Google is even more aggressive in its efforts to restrict the development of rival operating systems using Android’s open-source code. The members of the so-called Open Handset Alliance “are contractually prohibited from building non-Google approved devices.” For example, Google does not permit any major equipment manufacturer to produce the Kindle Fire for Amazon (e.g., Acer, Asus, Dell, Foxconn, Fujitsu, HTC, Huawei, Kyocera, Lenovo, LG, Motorola, NEC, Samsung, Sharp, Sony, Toshiba, and ZTE).

Whatever its original intent, Google has turned its control over Android into the second-coming of Carterfone.

It is also common knowledge that Apple routinely engages in similar behavior with respect to iOS and its devices. In response to an FCC investigation into the reasons for Apple’s decision to block the Google Voice app, Apple was quite clear that it blocks applications for “a variety of reasons,” including apps that contain content Apple finds “objectionable.” Apple’s capability to block developer access to end users based on its views with respect to content is inconsistent with the view of many net neutrality proponents who believe Internet intermediaries should be prevented from exercising editorial control.

For example, Apple’s “Violence” guideline states, “‘Enemies’ within the context of a game cannot solely target a specific race, culture, a real government or corporation, or any other real entity.” As a result of this provision, Apple has routinely rejected or required developers to whitewash games based on historical events. After Apple rejected a simulation of the Syrian Civil War for depicting Syrians, the developer protested, “Our aim is to use games as a format to bring news to a new audience and submission processes such as this do make it a lot harder for us.” The developer of a WWII naval combat simulation was permitted to publish their game only after removing depictions of the Japanese flag. In Apple’s version of WWII, America didn’t fight Japan in the Pacific, it fought The Generic Enemy.

To use the words of a standard net neutrality trope, if an application developer wishes to offer an app to an end user who accesses the Internet through iOS, the application developer will literally be required to “ask Apple for permission.” If Apple says “no,” the developer is blocked by the first Internet gate it reaches, which makes the potential to be blocked by an ISP irrelevant.

If permission is granted, the developer has to pay 30% of their initial sales and ongoing subscription revenue to Apple for the privilege of accessing end users of the mobile Internet (revenue that garnered Apple at least $3 billion in 2013) through Apple’s proprietary platform. To use the words of the Open Internet Order, is a 30% cut of all developer revenue in perpetuity an “inefficiently high fee” enabled by Apple’s gatekeeper control over iOS users? The FCC avoided answering that question in 2010 through its arbitrary decision to exclude non-ISP gatekeepers from the scope of its net neutrality rules.

Google Uses Its Gatekeeper Control Over Internet Search to Block, Degrade, and Discriminate Against Its Rivals

Internet search engines are also gatekeepers under the net neutrality theory of communications regulation. The Federal Trade Commission (FTC), the European Commission, and many state antitrust regulators have investigated whether Google’s discriminatory practice of promoting its own content on its search results pages while selectively demoting the content of its competitors harms its potential competitors. The FTC’s investigation found that Google’s discriminatory practices “had the effect in some cases of pushing other results ‘below the fold’”, which “arguably weakened those websites as rivals to Google’s own shopping vertical.” Google’s discrimination thus has the same effect — degrading the access of its rivals to end users — that prompted the FCC to impose net neutrality regulations on ISPs under the gatekeeper theory.

The particular reason a far edge provider cannot access an end user on the Internet makes no difference to the end user or the provider. The outcome is the same whether the gate is closed by Google or an ISP: The far edge provider is effectively denied access to end users who rely on Google’s search engine to access products and services on the Internet.

Netflix Uses Its Gatekeeper Control Over Online Video Distribution to Block, Degrade, and Discriminate Against Video Content

Many consumers that do not subscribe to a multichannel video programming distribution (MVPD) service rely on online video distributors (OVDs) for access to pay-TV content. The video platform provided by an online video distributor, which acts as an intermediary between end users and the owners of pay-TV content, has the capability to block, degrade, or favor particular video content with respect to its end users. For example, Netflix blocked end users from accessing high definition (HD) and 3D movies in an effort to spread its costs across all broadband subscribers.

OVDs who produce their own content or negotiate exclusive programming distribution deals (Netflix does both) also have the ability to discriminate through (1) favorable placement of affiliated content on the distribution platform, (2) using fast lanes for the delivery of affiliated content, (3) entering into exclusive deals for the distribution of third-party content, or (4) refusing to carry unaffiliated content altogether (the equivalent to blocking by an ISP). For example, Netflix has negotiated exclusive rights to distribution of first-run Disney movies, which effectively blocks end users from accessing that content online unless they subscribe to Netflix.

End Users May Incur Significant Costs In Switching Non-ISP Gatekeepers

In the Open Internet Order, the FCC discredited the competition theory of communications regulation, at least in part, because “customers may incur significant costs in switching broadband providers.” The Verizon court appeared to rely primarily on this rationale to affirm the reasonableness of the FCC’s shift from competition theory to the gatekeeper theory:

To be sure, if end users could immediately respond to any given broadband provider’s attempt to impose restrictions on edge providers by switching broadband providers, this gatekeeper power might well disappear. . . . But we see no basis for questioning the Commission’s conclusion that end users are unlikely to react in this fashion.

The court concluded that a gatekeeper’s “ability to impose restrictions on edge providers simply depends on end users not being fully responsive to the imposition of such restrictions.”

A comparison of the costs end users may incur when switching ISPs and the costs they may incur when switching non-ISP gatekeepers indicates that the switching rationale is equally applicable to non-ISP gatekeepers.

According to the FCC, the costs of switching ISPs include:

  • early termination fees;
  • the inconvenience of ordering, installation, and set-up, and associated deposits or fees;
  • possible difficulty returning the earlier broadband provider’s equipment and the cost of replacing incompatible customer-owned equipment;
  • the risk of temporarily losing service;
  • the risk of problems learning how to use the new service; and
  • the possible loss of a provider-specific email address or website.

Consumers often incur the same types of costs when switching non-ISP gatekeepers in addition to other types of switching costs that are unique to certain types of non-ISP gatekeepers. Some of the switching costs that are unique to non-ISP gatekeepers — e.g., the incompatibility of software purchased for a particular OS with alternatives — are likely to be substantially higher than the costs incurred when switching ISPs.

Costs of Switching Mobile Operating Systems

When a user chooses a mobile OS, there are frequently (1) early termination fees, (2) inconveniences and fees related to ordering, installation, and set up, (3) the possibility of difficulties returning or recycling an earlier device and the cost of replacing it; (4) the risk of temporarily losing service and access to information stored on the previous device or in the cloud; (5) the risk of problems learning how to use the new operating system and its associated apps (a risk that is likely much higher than with ISPs); and (6) the possible loss of a provider-specific email address or cloud service.

The fact that end users incur these types of costs when they switch mobile operating systems is obvious with the possible exception of early termination fees (ETFs), which are usually attributed to mobile ISPs. In reality, however, ETFs associated with broadband-capable mobile devices are often the result of business decisions made by OS service providers, not mobile ISPs. For example, due to the popularity of its iPhone device, which is tied to iOS, Apple was able to force Sprint to buy at least 30.5 million iPhones at a price of $20 billion (about $650 each) over a four-year period regardless of whether Sprint can find customers to buy them. To meet its massive commitment to Apple, Sprint decided to subsidize up to $500 of the cost of each iPhone with the hope that it could recoup its losses over the course of service contracts that are subject to early termination fees. Sprint made the risky $20 billion bet because the “cellphone business is increasingly driven by hot smartphones.” The early termination fees associated with some smartphones are thus as attributable to the costs of switching operating systems as they are to the costs of switching mobile ISPs.

Switching OSs also involves costs that are typically inapplicable to ISPs. Because they are not interoperable, switching OSs includes the costs of repurchasing apps that are compatible with the new OS. Unlike early termination fees, which are prorated and decrease over time, this switching cost may increase over time as consumers buy more apps. The more apps a consumer buys for a particular OS, the greater the lock-in effect. To put this into perspective, iOS users spent over $1 billion in Apple’s App Store in December 2013 alone.

In comparison, the costs of paying early termination fees to switch mobile carriers are minuscule or non-existent: Competition in the mobile market has driven T-Mobile to pay up to $650 in early termination fees to customers who switch to its service, and many providers “lament the damage to their bottom line” caused by smartphone subsidies. As noted above, ETFs benefit the providers of OSs, as well as consumers and device manufacturers, as much or more than they benefit mobile ISPs.

There is also the possibility that some apps on which a consumer relies are not available on any alternative OS. Because operating systems are an intermediary in a two-sided network, (1) consumers tend to prefer operating systems that offer more apps, and (2) app developers tend to prefer operating systems that have more users. Some developers create apps only for one or two of the most popular OSs, which means that some apps on which end users rely might never become available on an alternative OS.

These facts indicate that the cost of switching mobile ISPs may be lower for many end users than the cost of switching mobile operating systems.

Costs of Switching Search Engines

The Open Internet Order’s analysis of an end user’s ability to switch providers in response to discrimination included lack of information. The FCC noted that “end users may not know whether charges or service levels their broadband provider is imposing on edge providers vary from those of alternative broadband providers.”

This informational deficit is just as likely to occur when search engines manipulate their results to favor their own services and disadvantage their rivals. Search engines do not disclose how their search algorithms work or the ways in which they display results, which leaves end users with no way of knowing whether a search engine is engaging in discrimination. In these circumstances, there is no reason to believe that consumers would be responsive to search engine discrimination at all, let alone “fully responsive.”

Costs of Switching Online Video Distributors

Many of the costs of switching ISPs are also applicable to the costs of switching online video distributors (OVDs). There are (1) inconveniences related to ordering, installation, and set up; (2) the possibility of replacing incompatible devices (i.e., if a consumer uses a Blu-Ray player or other device that isn’t capable of accessing an alternative OVD); (3) the risk of temporarily losing service and information related to the previous OVD (e.g., favorites, recommendations, and viewing lists); and (4) the risk of problems learning how to use the new OVD’s platform and its associated apps.

Similar to mobile OSs, OVDs also impose switching costs related to the exclusivity of their content. For example, an end user who is frustrated with the discriminatory practices of Netflix were to switch to an alternative OVD, the end user would lose access to House of Cards, Disney movies, and other video content online that is available exclusively on the Netflix platform.

Given these switching costs, end users are unlikely to be any more responsive to an OVD who blocks, degrades, or discriminates against far edge services than they would be to an ISP who engages in similar behavior.

The Markets for Non-ISP Gatekeeper Services Provided by Google, Apple, and Netflix Are Highly Concentrated

Although the FCC was clear that application of the gatekeeper theory does not depend on gatekeepers having market power with respect to end users, it noted that the threats posed by gatekeepers “would be exacerbated by such market power.” The FCC emphasized that, as of December 2009 (almost five years ago now), nearly 70 percent of households lived in census tracts where only one or two wireline or fixed wireless firms provided broadband service with download throughput of at least 3 Mbps and upload throughput of at least 768 Kbps.

According to the National Telecommunications and Information Administration, as of December 31, 2013, 93.4% of U.S. households have wireline Internet access meeting those throughput thresholds, and 98.6% of U.S. households have wireless Internet access meeting those throughput thresholds. The NTIA reports that nearly 88% of U.S. households have access to at least two wireline broadband providers, and that 99.4% of U.S. households have access to at least two wireless broadband providers. These more recent statistics indicate that competition among broadband ISPs increased dramatically in the four-year period since December 2009. (Note that the NTIA doesn’t provide data regarding the relative market shares of broadband ISPs.)

Concentration in the mobile OS, search engine, and OVD market segments is typically as high or higher than concentration among ISPs:

The data demonstrates that, to the extent the FCC concluded that market concentration among ISPs was relevant to its imposition of net neutrality rules under the gatekeeper theory, it is just as relevant to the imposition of such rules on these non-ISP gatekeepers.

The market dominance of some non-ISP gatekeepers would also make it more difficult for the FCC to forbear from treating them as common carriers under Title II. The next post in this series will discuss the challenges presented by the FCC’s recent implementation of its forbearance authority.