Hybrid differentiation occurs when a platform discriminates among businesses in a related market in which it is not active itself with the aim to increase or maintain its competitive advantage elsewhere. An example is a platform blocking an app that interferes with its ability to gain revenues through advertising. Hybrid differentiation is particularly relevant where a platform builds an ecosystem or conglomerate around its main activity. As a competitive strategy beyond vertical integration that serves to protect the platform’s overall activities and business model, hybrid differentiation is different from the self-preferencing at stake in Google Shopping. This paper submits that the assessment of hybrid differentiation requires attention for the broader competitive strategies of dominant platforms going beyond traditional forms of leveraging and the relevant markets in which they operate.
By Inge Graef1
The Google Shopping saga has led to a fierce discussion about the theory of harm condemning self-preferencing by vertically integrated dominant firms as anticompetitive. This paper reflects on another type of discrimination, namely hybrid differentiation, whose effects are maybe even more complex and problematic than those of self-preferencing. Hybrid differentiation occurs when a platform discriminates among businesses in a related market in which it is not active itself in an effort to increase or maintain its competitive advantage within one of its other activities. An example is a platform blocking an app that interferes with its ability to gain revenues through advertising. Hybrid differentiation is particularly relevant where a platform builds an ecosystem or conglomerate around its main activity.
Digital platforms increasingly expand their activities by entering into markets adjacent to their original business, either by growing organically or by acquiring start-ups in neighboring markets. Think of the variety of services offered by Google ranging from online search, maps and mobile operating systems to self-driving vehicles and smart glasses. The acquisitions of Instagram and WhatsApp by Facebook illustrate how the social network provider has diversified its business. These developments give rise to conglomerates2 or ecosystems3 consisting of several related services offered by the same provider. While this growth can create efficiencies and improve the user experience, it also creates room for different services to be integrated with one another possibly to the detriment of outside services offered by rivals. The expansion of these conglomerates or ecosystems into more and more activities goes hand in hand with increased opportunity to discriminate against rivals as there will be more competitive interactions due to the many services offered. Beyond the pure self-preferencing that we already know from the Google Shopping case, there is also scope for hybrid differentiation to occur because of the interconnectedness between the various services.
Whereas self-preferencing as currently understood (i.e. the favoring of a platform’s own activities over those of non-affiliated rivals) occurs in scenarios of vertical integration, this is not the case for hybrid differentiation as considered in this paper. The key difference between hybrid differentiation and self-preferencing is that the exclusionary effects that the platform aims to achieve through hybrid differentiation are not situated in the same market as where the differentiation takes place. In the Google Shopping case, the impact of Google’s self-preferencing behavior was mainly felt in the market for comparison shopping services in which Google was vertically integrated. However, the main feature of hybrid differentiation is that the platform differentiates between non-affiliated businesses in a market in which it is not present itself in order to exclude competition somewhere else. Hybrid differentiation is a competitive strategy beyond vertical integration to ensure the continued relevance of the platform’s overall activities and business model.
This paper explores how to assess the anticompetitive nature of hybrid differentiation and submits that there is a need to look beyond competition in existing markets and traditional notions of leveraging. Section II discusses the notion of hybrid differentiation in comparison with two other forms of differentiation. Section III explores hybrid differentiation in the form of blocking of access to a platform. Section IV focuses on hybrid differentiation through tying of additional services. Section V concludes.
II. NOTION OF HYBRID DIFFERENTIATION
Discrimination or differentiation is at the heart of a number of ongoing competition investigations. The European Commission seems to be testing the relevance of its reasoning regarding self-preferencing in search rankings in Google Shopping to concerns about preferential access to data in the Amazon investigation opened in July 2019.4 Amazon is under the radar of the Italian competition authority as well, which is examining the allegation that Amazon discriminates on its platform in favor of third-party merchants who use Amazon’s logistics services.5 Beyond the investigation into the behavior of Amazon, the Commission is also looking into a complaint from Spotify against Apple’s conduct related to the App Store. This includes the possible discriminatory nature of the 30 percent fee Apple charges as a commission from rival apps like Spotify but not from its own apps like Apple Music.6 The Netherlands Authority for Consumers and Markets is looking specifically at the position of Dutch apps for news media in Apple’s App Store and possible preferential treatment by Apple.7
While all of these cases contain elements of discrimination or differentiation, the nature of the conduct and their potential to create competitive harm is different. This paper builds upon the categorization made elsewhere of differentiated treatment into three types, namely pure self-preferencing, pure secondary line differentiation and hybrid differentiation.8 This distinction serves as an analytical framework to assess the extent to which such practices are abusive under Article 102 of the Treaty on the Functioning of the European Union (TFEU). While this paper focuses on the third type of hybrid differentiation only, to understand its competitive impact it is helpful to illustrate how this notion relates to the other two forms of differentiation.
The three types of differentiation relevant to behavior of digital platforms can be defined as follows:
- “pure” self-preferencing: consisting of behavior whereby a vertically integrated platform treats its affiliated services more favorably than non-affiliated services;
- “pure” secondary line differentiation: occurring when a non-vertically integrated platform differentiates among non-affiliated services in a market in which it is not active itself;
- “hybrid” differentiation: conduct whereby a platform differentiates among non-affiliated services in an effort to favor its own business.9
The Google Shopping case is of course the key illustration of pure self-preferencing, where Google was fined for displaying its own comparison shopping service more prominently in its general search results to the detriment of rival comparison shopping services.10 As an example of pure secondary line differentiation, one can think of a hotel booking platform providing hotels that pay higher commission fees with a better ranking in the search results on its platform. This type of differentiation is labelled “pure secondary line differentiation” here by reference to the secondary line injury that it may cause. Secondary line injury occurs when a supplier distorts competition on a downstream market where it is not active by favoring and exploiting some customers over others.11 Hotel booking platforms like Booking.com or Expedia do not offer hotel rooms themselves and are thus not in competition with the hotels that rely on the platform to reach consumers. When a platform differentiates among non-affiliated customers in a market in which it is not operating itself, the harm is exploitative and thus qualifies as secondary line injury because it affects the platform’s downstream customers that are in competition with each other.
Primary line injury occurs when a supplier forecloses competitors from the market in which it operates itself. Pure self-preferencing thus leads to primary line injury, because the key objective of the platform is to exclude direct competitors such as the rival comparison shopping services in Google Shopping. Hybrid differentiation differs from pure self-preferencing because there is no favoring of a platform’s own services vis-à-vis non-affiliated services. Instead, the differentiation takes place among non-affiliated businesses but indirectly benefits the platform in a different market than the one in which the non-affiliated customers compete. While hybrid differentiation also involves exploitative elements by favoring some customers over others, an exclusionary motive prevails because the platform’s ultimate objective is to strengthen its own market position in relation to rivals competing with the platform’s activities elsewhere.12 Such competition elsewhere can constitute challenges against the platform’s existing business model more generally. An example is a platform blocking an app that interferes with its ability to gain revenues through advertising (see section III). Or the competition elsewhere can be a third market in which the platform is operating, in addition to the market of the main activity of the platform and the market in which the non-affiliated businesses compete. An example is a platform conditioning a business’s ranking on whether additional services are purchased (see section IV).
It is important to note that the three types of differentiation serve as a framework for assessing their anticompetitive nature. The fact that behavior falls within one of the categories therefore does not imply that it breaches Article 102 TFEU. The categorization mainly acts as a tool to select the appropriate lens to test the anticompetitiveness of differentiation. As the impact of hybrid differentiation goes beyond the relevant markets in which the platform is active, it raises particular challenges for competition analysis. To make these concerns more concrete and show how they require competition authorities to look beyond existing relevant markets, two real-world examples are discussed in the following sections.13
III. HYBRID DIFFERENTIATION AND BLOCKING OF PLATFORM ACCESS
The most far-reaching decision that a platform can take in its relationship with a business user is to block the latter’s access. Such a decision can result in hybrid differentiation. The situation of the app Unlockd vis-à-vis Google illustrates this.
Unlockd is an app that provides a different approach to mobile advertising by showing advertising or other content when a user unlocks her phone. In return for viewing ads, content, or offers upon unlocking their smartphone, users earn points they can exchange for mobile credit, data, entertainment or loyalty points.14 In 2018, Google announced its intention to remove Unlockd’s app from the Play Store because of an alleged violation of its terms and conditions that prohibit apps from interfering with the operation of the device or other apps. In May 2018, UK Judge Peter Roth granted Unlockd an interim injunction preventing Google from removing Unlockd’s apps made for the UK market.15 However, due to lack of funding to pursue the proceedings, Unlockd withdrew its claim in February 2019 and was in May 2019 even ordered by the UK Competition Appeal Tribunal to pay Google’s costs.16 The Australian Competition and Consumer Commission has been reported to start legal action against Google for its conduct vis-à-vis Unlockd in 2020.17
The removal of Unlockd from the Play Store can be qualified as a form of hybrid differentiation, because the reason for Google’s decision lies in its interest to protect its own business indirectly. Google is not a direct competitor of Unlockd, as it is not active in the markets in which the app competes. However, Unlockd may make it harder for Google to monetize its activities as advertising forms its main revenue stream. This means that there is an exclusionary motive for Google to block the app.
The blocking of access to a platform can be seen as a disruption of supply. This would mean that the notion of refusal to deal forms the relevant legal framework to assess the anticompetitive nature of such behavior. The difficulty of capturing hybrid differentiation under the concept of refusal to deal is the current interpretation of the requirement of exclusion of effective competition.18 The way in which this requirement has been applied in case law so far indicates that the relevant question is whether the dominant firm reserves a downstream market to itself by denying a competitor access to an input. In Magill, the Court of Justice noted that the Irish broadcasting stations “reserved to themselves the secondary market of weekly television guides by excluding all competition on that market.”19 Similarly, in IMS Health, the Court of Justice argued that the refusal was “such as to reserve to the owner of the intellectual property right the market for the supply of data on sales of pharmaceutical products in the Member State concerned by eliminating all competition on that market.”20
The Tiercé Ladbroke judgment illustrates the limits of this interpretation. The case dealt with a refusal by organizers of French horse races to provide Ladbroke, who was offering betting services in Belgium, with a transmission license for sound and pictures of the French horse races. Apart from the lack of indispensability, the General Court held that the condition of exclusion of effective competition was not met. The organizers of the French horse races were not competing with Ladbroke in the relevant market for the provision of betting services in Belgium. For that reason, they could not be seeking to reserve that related market for themselves.21 As such, the conduct that a refusal to deal abuse targets is the leveraging of market power from the upstream market, namely the input to which access is requested, to a downstream market in which the dominant firm competes with the access seeker. This would mean that scenarios in which a requesting undertaking needs access to an input in order to enter a market in which the dominant firm is not active (and not planning to be active in the foreseeable future) cannot be captured.22 Other types of abuse like tying, margin squeeze, and even the self-preferencing at stake in Google Shopping likewise focus on leveraging of market power to a market where the dominant firm already operates.
Such a narrow interpretation is becoming especially problematic as the scope for hybrid differentiation will continue to rise. Because platforms expand their activities to related markets, they also have incentives to block access of businesses that are not direct competitors but limit the platform’s ability to gain profits elsewhere in its ecosystem or pose a more long-term threat to a platform’s underlying business model. The notion of leveraging as developed in cases dealing with more static market settings does not capture this competitive reality of the platform economy. To monitor possible competitive harm, there is a need for competition authorities to look beyond the existing markets in which the dominant firm already competes. By blocking the access of a business, the platform can prevent or delay disruption by new entrants and strengthen its overall control over the room for innovation to occur beyond the markets in which it is present itself.23 Attention for dynamic effects beyond markets when assessing hybrid differentiation in the form of blocking of platform access is thus key to keep the platform economy competitive and innovative. The same also holds for hybrid differentiation that takes place through tying of additional services to which attention now turns.
IV. HYBRID DIFFERENTIATION AND TYING OF ADDITIONAL SERVICES
The investigation of the Italian Competition Authority is an example of how hybrid differentiation can occur by conditioning a business’s ranking on whether additional services are purchased. Amazon is allegedly providing improved visibility, higher search rankings and better access to consumers only to merchants that also use its logistics services.24 Even though some of these merchants may compete with Amazon in the context of its retail activities, the impact of this behavior mainly affects the market for logistics services. By giving less visibility to merchants that do not rely on its logistics services, Amazon is steering merchants to its own services and thereby reduces competition possibly to the detriment of rival providers of logistics services. Amazon thus exploits non-affiliated merchants by favoring some over others in order to obtain a benefit in the market for logistics services where the affected merchants do not compete, but where Amazon wants to strengthen its position. This exclusionary element elsewhere in Amazon’s activities is what qualifies its behavior as a form of hybrid differentiation.
Amazon’s conduct involves elements of tying or bundling as it is making a merchant’s ranking dependent on whether the merchant also purchases Amazon’s logistics services for the products and services sold through the platform. A difficulty in applying the notion of tying to Amazon’s behavior is that merchants are not required to make use of Amazon’s logistics services in order to be able to sell goods through Amazon’s platform. In other words, there is no contractual obligation or technical integration between the two services that forces businesses to rely on Amazon’s logistics services when using its marketplace services. This while tying normally deals with situations where the purchase of one product is made dependent on the purchase of another product. For instance, Microsoft was fined by the Commission in 2004 for tying Windows Media Player to the Windows operating system by pre-installing its media player on Windows PCs.25 Similarly, the commitments offered by Microsoft to the Commission in 2009 addressed the technical tying between Internet Explorer and the Windows operating system.26
In order to be abusive under Article 102 TFEU, tying requires a form of coercion precluding customers a choice to obtain the tying product without the tied product. Based on the 2007 Microsoft judgment of the General Court, the Commission has interpreted the notion of coercion broadly in its Google Android decision. Tying was used as the legal framework to assess the anticompetitive effects of Google’s requirement to make the licensing of the Google Play Store by Android device manufacturers conditional upon the pre-installation of the Google Search and Google Chrome apps.27 In this context, the Commission argued that coercion can still exist when the customer is not required to use the tied service or is entitled to use the same product supplied by a competitor of the dominant undertaking.28
At the same time, there seems to be a difference between the situation at stake in Google Android and the conditioning of the ranking of a business on whether it uses a dominant platform’s additional services. In the latter case, businesses can decide not to take a dominant platform’s tied service but still make use of its marketplace, while Android device manufacturers could not obtain a license for the Google Play Store if they decided not to pre-install the Google Search and Google Chrome apps. Similarly, device manufacturers were not given a choice with regard to the abuse relating to Android forks that was also analyzed through the notion of tying in Google Android. Google namely conditioned licensing of the Play Store and the Google Search app on device manufacturers agreeing not to sell mobile devices running on alternative, non-approved versions of Android.29 The extent of coercion is thus different here.
Alternatively, one can analyze the behavior as a form of mixed bundling, also referred as a multi-product rebate, where products are made available separately as well, but the sum of the separate prices is higher than the bundled price.30 In cases where the ranking of a business is made conditional upon its decision to purchase additional services from the platform, the mixed bundling does not lead to a lower bundled price. It is instead the higher quality of the bundled offer through a more favorable placement in the ranking that counts. This benefit can only be obtained by businesses if they rely on the additional services offered by the dominant platform.
According to the Commission in its Google Android decision, the capability of tying to restrict competition can among others be considered by looking at whether the tying “reduces the incentives of users to choose a product from among those of other suppliers than the dominant undertaking.”31 In the Microsoft case, the General Court argued that consumers were less likely to use an alternative media player due to the pre-installation of Windows Media Player on the Windows operating system.32 A similar effect seems present in the context of the Amazon investigation. Because merchants are punished for not taking Amazon’s logistics services through lower visibility in rankings to consumers, they are induced not to rely on logistics services provided by other market players.
Another issue is whether the tying is liable to foreclose competition in the market for the tied product. If in the Amazon case merchants despite the tying still rely on logistics services of other providers, this can offset the foreclosure effect created by Amazon’s behavior. Similarly, attention was paid to whether consumers used third party media players and web browsers in the Microsoft cases.33 According to the Commission’s Guidance on Article 102 TFEU Enforcement Priorities, a multi-product rebate is anticompetitive “if it is so large that equally efficient competitors offering only some of the components cannot compete against the discounted bundle.”34 The ability of rivals in the tied market to compete despite the tying or bundling is thus key.
In addition, the platform may argue that its behavior creates efficiencies for consumers. Amazon could for instance claim that the way it bundles its services creates an improved experience for consumers because it can better ensure the overall quality from sale until delivery. However, the acceptance of such arguments would risk opening the door for dominant firms to bundle all sorts of services to its main activity even if this has a foreclosure effect on rivals. As made clear by the Court of Justice in UK Generics, to show that efficiency gains offset negative effects on competition, the dominant firm “has to do more than put forward vague, general and theoretical arguments on that point or rely exclusively on its own commercial interests.”35
The effects created by what is called hybrid differentiation here are not entirely new and can also be assessed through other means beyond the notion of tying, in particular by reference to the benefits the behavior creates for the dominant firm. A relevant precedent is British Gypsum36 where it was found abusive for a dominant firm to engage in favorable treatment of customers in a horizontally related market where it was not dominant. Depending on whether customers were loyal to it in the main dominated market (the market for plasterboard) by not importing plasterboard from rivals, they qualified for priority deliveries in the related market (the market for plaster).37 The abusive conduct served to secure the firm’s dominant position in the main market. While there are similarities with the situation at stake in British Gypsum in particular as regards the way customer loyalty is used to obtain benefits across horizontally connected markets, the objective of the platform’s behavior in cases of hybrid differentiation at stake here is different. Most importantly, hybrid differentiation has a two-fold nature. On the one hand, the platform aims to obtain benefits in the non-dominated market where it wants to expand its activities. This makes the behavior more conventional as it constitutes a traditional form of leveraging, whereas the novelty of British Gypsum was that the abuse took place in the non-dominated market and the benefits occurred in the dominated market.38 On the other hand, the platform wishes to protect its competitive advantage over the entire ecosystem that it is building around the main market where it holds a dominant position. Even if the first effect does not turn out to be anticompetitive because of the limited impact in the non-dominated market, the second effect should not be ignored.
Situations can even be considered where both the abuse and the benefits of hybrid differentiation are situated in a part of the platform’s activities where it is not dominant. The Court of Justice found such behavior abusive in Tetra Pak II under the condition that there were “associative links” between the two markets under investigation, namely the market for aseptic packaging where Tetra Pak was dominant and the market for non-aseptic packaging where Tetra Pak was not dominant and where the abuse and benefits occurred.39 The restriction to situations with associative links between connected markets is becoming problematic where a platform enters markets that are not closely related to each other but where the platform can still engage in anticompetitive conduct through favorable treatment across activities or by recouping losses elsewhere within its ecosystem.40
In British Gypsum, the General Court argued that the behavior amounted to the “provision of equivalent services on unequal terms” and had to be regarded as anticompetitive in itself “by reason of the discriminatory purpose which it pursues and the exclusionary effect which may result from it.”41 This provides scope to hold hybrid differentiation abusive where it is based on a discriminatory strategy that is anticompetitive by its nature, although discussions will then likely focus on how to define the equivalence of transactions of a platform with businesses in slightly different positions.
Regarding the competitive assessment as whole, the key point is that the impact of arrangements where a business’s ranking is made dependent on whether it purchases additional services from the dominant platform goes beyond the relevant markets at stake. In particular, such arrangements can also influence the ability of potential competitors and new entrants to compete with parts of the ecosystem of the dominant player in the future.42 This market reality needs to be reflected in competition analysis to ensure that competition at the edges of the platform’s activities remains strong. To achieve this, attention should be paid in the competition analysis not only to the effects in the affected relevant markets but also to how the conduct enables a platform to enter an increasing number of related markets. The nature of the underlying strategies for hybrid differentiation should thus be analyzed. Limits to expansion by a platform through practices of differentiation are necessary to ensure that the overall platform economy remains competitive and that there is room for new entrants to challenge the platform in separate segments of its activities.
The increasing expansion in which digital platforms engage through the creation of conglomerates or ecosystems around their main activity creates room for practices that have been referred to as hybrid differentiation here. Hybrid differentiation occurs when a platform differentiates between non-affiliated customers in order to serve its own interests elsewhere in its ecosystem. The key challenge in terms of the ability to assess the anticompetitive nature of such conduct is that its impact goes beyond the markets in which platform itself is present. Foreclosure effects can have a longer-term impact on the future room for innovation by potential competitors and new entrants.
To adequately monitor competitive harm, there is a need to look at the competitive strategies of dominant platforms going beyond traditional forms of leveraging and the relevant markets in which they operate. Instead, the competitive harm relates to the ability of the platform to prevent or delay disruption, to ward off challenges against its existing business model and to strengthen its control over additional activities within the entire ecosystem.
This also means that regardless of whether the Google Shopping decision is confirmed on appeal, challenges remain. The Commission’s assessment of the self-preferencing in Google Shopping was still confined to existing relevant markets, while hybrid differentiation requires competition authorities to take a more forward-looking approach beyond the boundaries of competition in markets in which the platform already operates. The Google Android appeal may provide clarity regarding the application of the notion of tying in an ecosystem where different markets are connected, including search, mobile operating systems and app stores. If platforms no longer compete solely in narrowly defined relevant markets, competition analysis needs to reflect this competitive reality.
1 Inge Graef is assistant professor at Tilburg University, with affiliations to the Tilburg Law and Economics Center (“TILEC”) and the Tilburg Institute for Law, Technology, and Society (“TILT”). I would like to thank Francisco Costa-Cabral for his comments on an earlier version of this paper.
3 Jacques Crémer, Yves-Alexandre de Montjoye & Heike Schweitzer, “Competition policy for the digital era,” 2019, p. 30-38, available at http://ec.europa.eu/competition/publications/reports/kd0419345enn.pdf.
4 Press release European Commission, “Antitrust: Commission opens investigation into possible anticompetitive conduct of Amazon,” July 17, 2019, available at https://ec.europa.eu/commission/presscorner/detail/en/IP_19_4291.
5 Press release Italian Competition Authority, “A528—Amazon: investigation launched on possible abuse of a dominant position in online marketplaces and logistic services,” April 16, 2019, available at https://en.agcm.it/en/media/press-releases/2019/4/Amazon-investigation-launched-on-possible-abuse-of-a-dominant-position-in-online-marketplaces-and-logistic-services.
7 Press release Netherlands Authority for Consumers and Markets, “ACM launches investigation into abuse of dominance by Apple in its App Store,” April 11, 2019, available at https://www.acm.nl/en/publications/acm-launches-investigation-abuse-dominance-apple-its-app-store.
8 Inge Graef, “Differentiated Treatment in Platform-to-Business Relations: EU Competition Law and Economic Dependence,” Yearbook of European Law 2019, p. 448-499, available at https://doi.org/10.1093/yel/yez008.
9 Ibid, p. 452-453.
10 Case AT.39740 Google Shopping, June 27, 2017.
11 For the distinction between primary and secondary line injury in the context of abuse of dominance, see Pablo Ibanez Colomo, “Exclusionary discrimination under Article 102 TFEU,” Common Market Law Review 2014, p. 145.
12 For a further characterization, see Inge Graef, “Differentiated Treatment in Platform-to-Business Relations: EU Competition Law and Economic Dependence,” Yearbook of European Law 2019, p. 453.
13 The discussion in sections III and IV builds upon the analysis in Inge Graef, “Differentiated Treatment in Platform-to-Business Relations: EU Competition Law and Economic Dependence,” Yearbook of European Law 2019, p. 448-499.
15 Unlockd v. Google (2018) Ch D (Roth J), May 9, 2018. See R. English, “Win (for now) for app developer against Google,” UK Human Rights Blog, May 18, 2018, available at https://ukhumanrightsblog.com/2018/05/11/win-for-now-for-app-developer-against-google/.
16 Unlockd v. Google  CAT 17, May 21, 2019, available at https://www.catribunal.org.uk/sites/default/files/2019-05/1283T_Unlockd_CMC_CAT_17_210519.pdf.
17 Paul Smith, “ACCC to sue Google over Unlockd,” Australian Financial Review, November 5, 2019, available at https://www.afr.com/technology/accc-to-sue-google-overunlockd-20191030-p535u9.
18 The other conditions to hold a refusal to deal abusive are: the indispensability of the input; the prevention of the introduction a new product (relevant for intellectual property protected assets); and the absence of an objective justification. See Joined cases C-241/91 and C-242/91 Magill, ECLI:EU:C:1995:98; Case C-7/97 Bronner, ECLI:EU:C:1998:569; Case C-418/01 IMS Health GmbH, ECLI:EU:C:2004:257; Case T-201/04 Microsoft, ECLI:EU:T:2007:289.
19 Joined cases C-241/91 and C-242/91 Magill, ECLI:EU:C:1995:98, para 56.
20 Case C-418/01 IMS Health GmbH, ECLI:EU:C:2004:257, para 52.
21 Case T-504/93 Tiercé Ladbroke v. Commission, ECLI:EU:T:1997:84, par. 133.
22 Reaching the same conclusion in the context of access to data, see Josef Drexl, “Designing Competitive Markets for Industrial Data Between Propertisation and Access,” JIPITEC 2017, p. 282-283.
23 For a discussion on how competition law should assess competitive strategies aimed at preventing disruptive innovation, see Francisco Costa-Cabral, “Innovation in EU Competition Law: The Resource-Based View and Disruption,” Yearbook of European Law 2018, p. 305-343.
24 Press release Italian Competition Authority, “A528—Amazon: investigation launched on possible abuse of a dominant position in online marketplaces and logistic services,” April 16, 2019, available at https://en.agcm.it/en/media/press-releases/2019/4/Amazon-investigation-launched-on-possible-abuse-of-a-dominant-position-in-online-marketplaces-and-logistic-services.
25 Case COMP/C-3/37.792 – Microsoft, March 24, 2004 as upheld on appeal in Case T-201/04 Microsoft, ECLI:EU:T:2007:289.
26 Case COMP/C-3/39.530 – Microsoft (tying), December 16, 2009.
27 Case AT.40099 – Google Android, July 18, 2018, par. 740-753.
28 Case AT.40099 – Google Android, July 18, 2018, par. 748 referring to Case T-201/04 Microsoft, ECLI:EU:T:2007:289, par. 970.
29 Case AT.40099 – Google Android, July 18, 2018, par. 1011-1018.
30 Communication from the Commission — Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings  OJ C 45/7, par. 48.
31 Case AT.40099 – Google Android, July 18, 2018, par. 750 referring to Case T-201/04 Microsoft, ECLI:EU:T:2007:289, par. 1041.
32 Case T-201/04 Microsoft, ECLI:EU:T:2007:289, par. 1041: “in the absence of the bundling, consumers wishing to have a streaming media player would be induced to choose one from among those available on the market.”
33 Ibid, par. 1049-1077 and Case COMP/C-3/39.530 – Microsoft (tying), 16 December 2009, par. 39-54.
34 Communication from the Commission — Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings  OJ C 45/7, par. 59.
35 Case C-307/18 UK Generics, ECLI:EU:C:2020:52, par. 166.
36 With thanks to Francisco Costa-Cabral for referring me to this case.
37 Case T-65/89 British Gypsum, ECLI:EU:T:1993:31, par. 92-96 as upheld on appeal in C-310/93 P British Gypsum, ECLI:EU:C:1995:101.
38 For tables analyzing the links between dominance, abuse and effects in the different markets, see Giorgio Monti, EC Competition Law, Cambridge University Press 2007, p. 193 and Richard Whish and David Bailey, Competition Law, Oxford University Press 2018, p. 213.
39 Case C-333/94 P Tetra Pak II, ECLI:EU:C:1996:436, par. 28-30.
40 See Giorgio Monti, EC Competition Law, Cambridge University Press 2007, p. 193-194 who refers to the example of a firm dominant in the market for postal services that enters the market for toothpaste and engages in below-cost pricing for toothpaste by compensating its losses there with its high profits in the market for postal services. Note that the Commission has stated that it may also pursue predatory practices by dominant firms on markets on which they are not yet dominant. See Communication from the Commission — Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings  OJ C 45/7, footnote 39.
41 Case T-65/89 British Gypsum, ECLI:EU:T:1993:31, par. 94 as upheld on appeal in C-310/93 P British Gypsum, ECLI:EU:C:1995:101.
42 See Marc Bourreau & Alexandre De Streel, “Digital conglomerates and EU competition policy,” March 2019, p. 15 who quote Jean Tirole: “A start up that may become an efficient competitor to such firms generally enters within a market niche; it’s very hard to enter all segments at the same time. Therefore, bundling may prevent efficient entrants from entering market segments and collectively challenging the incumbent on the overall technology.”