Last week’s Commission decision on Google’s Android operating system made for great headlines. On top of the eye-watering fine of €4.3 billion, it even had the dubious honour of eliciting a presidential tweet. It was the kind of decision that sparked extensive papers long before its adoption, and a barrage of commentary immediately afterwards. Now that the dust has settled, this blog post takes a closer look at what we know, unearthing – as has been suggested – ‘the exciting legal stuff beneath the noise’.
While some commentary may give you the opposite impression, we know fairly little about the decision. The Commission has only published a press release; the actual decision will only be made public at a later stage (in about half a year, if the previous Google decision offers any indication). It’s therefore important to remember we’re discussing a 5-page memo rather than the hundreds of pages of in-depth assessment that will follow.
Now, on to the ‘decision’. The Commission’s main finding is that ‘Google has imposed illegal restrictions on Android device manufacturers and mobile network operators to cement its dominant position in general internet search.’ To understand the intricate connection between Google’s mobile operation system (‘OS’) and its search engine, a historical note is appropriate.
Google Search started in 1997, but it was in 2000 that the company found its business model: the search engine would be financed by selling advertisers access to its users (even though Google’s founders earlier spoke out against advertising-funded search engines). Google then went on to become the world’s leading search engines – on desktops.
It is easy to see how the advent of advanced mobile devices in the mid-2000s could have worried Google: they had the potential to cut end users off from Google’s search engine. In 2005, Google acquired the developer of the Android mobile OS. In 2007 (the year of the first iPhone), Google started distributing Android as an open-source OS, allowing others to use it freely to run their mobile devices, and even to tweak it according to their wishes (resulting in so-called ‘Android forks’).
This set-up laid the groundwork for the Commission’s Google decision, which holds, inter alia, that Google tied its search engine to its app store on Android devices, and prevented the development of Android forks. As these practices constitute an abuse of dominance, a crucial part of the Commission’s assessment is focused on establishing said dominance.
The Commission establishes Google’s dominance with regard to the three products under examination, i.e. internet search, mobile OS and app stores. This exercise is easiest for the first market of ‘general internet search’, as the Commission already determined Google’s dominance on this market in its Google Search decision of June 2017. The criticism against this market definition therefore remains the same, namely the omission of specialized search services such as Amazon’s (for product search). This also means that, should the General Court/ECJ disagree with this market definition, such ruling would have consequences for both Commission decisions.
Secondly, the Commission defines a market for ‘smart mobile OS available for license’. With Android, Google is said to have a market share of over 95% on this market. Moreover, this market is characterized by high barriers to entry due to network effects: as more consumers use a mobile OS, more developers will write apps for the system, which in turn attracts more users – and so on. New or smaller OS cannot benefit (to the same extent) from this ‘positive feedback loop’.
But what about Apple (and its iOS)? Did the Commission completely disregard this strong player in the mobile OS market? Much of the commentary would have you believe so, but it did not. In fact, the Commission concluded – sensitively – that iOS and Android do compete, even though they are not in the same market. The reason they are not in the same market is simple: Apple’s iOS is not directly marketed at all. iOS is only integrated in iPhones, and not licensed to device manufacturers. Android, on the other hand, is available for license. Customers of mobile OS, i.e. device manufacturers, therefore cannot choose between iOS or Android.
In contrast, end users do have choice between iOS and Android. When they buy a phone, for example, they can choose between an iPhone (with iOS) or any other phone (which will, in almost 100% of the cases, run on Android). The Commission has long recognized indirect constraints coming from the retail market, especially in telecommunications markets. Here too, the Commission recognizes that Apple and Google compete for end users in the mobile sphere. However, its investigation showed that this competition did not sufficiently temper Google’s market power vis-à-vis device manufacturers – the relation where the abuse is situated.
It must be noted, however, that even if iOS were added to the market, Android would still give Google a market share of more than 50% in most European countries, which could also underpin a finding of dominance.
The final market is that of ‘app stores for the Android mobile OS’. The Commission finds that Google’s app store, the Play Store, accounts for more than 90% of apps downloaded on Android devices. This is no surprise, as installation of apps from any source other than Google Play is blocked on Android by default. Moreover, Google doesn’t allow other app stores to be downloaded from within Google Play itself. (To be fair, the (in)security of third-party apps may offer a valid reason to restrict such downloads.)
Google is thus dominant in a large part of the mobile value chain, which runs from devices to OS to app stores to the apps themselves. At the heart of the Commission decision is the way in which Google uses its grip on this value chain to cement its dominance in general internet search.
The Commission takes issue with three separate practices, all of which were aimed at protecting and extending Google’s search dominance. One of these practices is relatively straightforward: Google guaranteed exclusive pre-installation of Google Search on mobile devices from manufacturers and network operators. As it did so through financial incentives, Google’s conduct amounts to an exclusivity rebate, which the Commission appears to have assessed according to the standards set out in the ECJ’s recent Intel judgment.
There is more ground to dispute the anti-competitive nature of the other two practices.
Google has been offering its mobile apps to device manufacturers as a bundle. One of these apps – Google Play – is considered a ‘must-have’ app to device manufacturers, and Google has made licensing of this app store conditional upon pre-installation of the Google Search app and Google Chrome browser (another starting point for internet searches). We are thus in a fairly traditional tying scenario reminiscent of the Microsoft cases (Windows Media Player and Internet Explorer).
For tying to be illegal, there needs to be foreclosure. The Commission bases its foreclosure theory on status quo bias: users who finds certain apps pre-installed on their devices are likely to stick to them, which means that rival apps are denied a change to compete effectively. An often-heard counterargument is that users have gotten more tech-savvy since Microsoft and do download different apps when they’re unhappy with the default. Since they do not, Google Search must simply be the superior service. This is, however, called into question by the behaviour of Windows Mobile device users, the great majority of which stick to the pre-installed Bing search engine. Must we consider that search engine superior too then?
While the restrictive nature of the tying is quite accepted, most commentators argue that it is necessary. Google licenses Android to device manufacturers at no cost, so it allegedly needs to make money from search; if not, it would have to start charging for Android, which would make mobile devices more expensive – something no one wants. The problem with this argument is that search is far from Google’s only way to finance its OS.
Most obviously, Google makes money from the Play Store that is on virtually every Android device; a 30% commission fee on every (in-)app purchase. Additionally, it has emerged that Google receives Android users’ search and location data from the OS, even when specific Google services are not being used – conduct that the Australian Competition and Consumer Commission is examining. So contending that search is Google’s only (viable) way to get a return on its investment in Android appears incorrect.
The last abuse relates to Android’s open source model. Google publicly releases the required information and source code to allow developers to create custom variants of Android (‘forks’). Many Android forks currently exist, yet few successful ones (Amazon’s Fire OS is the best and perhaps only example). There’s a reason for this. As one article already noted in 2013: ‘Android is open, except for all the good parts’. While the OS itself is open source, Google has gradually made new versions of apps within the OS closed source – leaving an outdated version open source.
Device manufacturers consider an updated suite of Google apps (including Gmail, Maps, YouTube, Play Store, Search) essential. However, to be able to pre-install these proprietary apps, manufacturers have to commit not to develop or sell even a single device running on an Android fork. The Commission holds that this restriction has halted the development and marketing of Android forks, thereby limiting innovation. In doing so, Google also closed off an important channel for competitors to introduce apps, in particular search services, which could be pre-installed on those Android forks.
Google argued that the restrictions were necessary to prevent the ‘fragmentation’ of the Android ecosystem: forks could be affected by technical failures or fail to support the proprietary Google apps, thereby worsening the Android user experience. According to the Commission, however, Google did not provide any credible evidence in that regard.
Elsewhere, Google went even further. For example, it concluded anti-fragmentation agreements with South-Korean Samsung, prohibiting the phone maker from developing its own mobile operating based on Android – conduct that is being investigated by the South-Korean Fair Trade Commission. One could argue that these restrictions are similar to non-compete obligations in franchising agreements: ‘Is it not sensible for a company to prevent free-riding and to make sure that it does not create competition to itself when licensing its products and services?’
An additional rationale for Google’s conduct – not discussed in the press release – concerns the indirect network effects mentioned above. App developers have to build apps for a specific OS. If too many forks existed, they would have to divide their time and resources; apps would be available on one version of Android but not the other; etc. This would hurt consumers, as their access to innovative apps would be limited.
All in all, this third form of abuse may be the fuzziest one. While the conduct itself appears anti-competitive, there are undeniable pro-competitive rationales for the conduct. The decision itself will have to bring clarity on how the Commission weighed these different elements. In any case, Google’s strategy vis-à-vis forks makes one wonder why Android was made open source in the first place.
While fines make headlines, it’s the remedies that matter. At a minimum, Google has to stop the three types of practices discussed above. Some worry that the only solution is to make device manufacturers pay for the Android OS – something even Google CEO Sunder Pichai has hinted at. For reasons outlined above (revenue from Google Play and data collected through Android), such fears appear overblown. But while Google swallowed the fine quite easily, the required changes may have a bigger impact on its baseline.
There are more fundamental questions than the ones dealt with above, such as whether the EU is returning to a competition law approach based on market structure rather than consumer welfare – as some seem to believe (incorrectly, if you ask me). But for a solid answer to such questions – and many others – we’ll have to wait for the full decision.