Author: Tomáš Granec Student number: s12350109 Supervisor: Laurens Ankersmit Second reader: Jan Broulík
Master’s Thesis
European Competition Law and Regulation
Indispensability of data for business in data-driven markets:
a case for the essential facilities doctrine
Table of Contents
List of abbreviations ... 3
1. Introduction ... 4
1.1 Data: the new oil for the digital economy ... 5
1.2 Understanding the nature of data ... 6
1.3 The Advantage of data ... 7
1.4 The Merger of WhatsApp and Facebook ... 9
1.5 Conclusion ... 10
2. Data collection as a threat to competition ... 11
2.1 Why do we regulate competition? ... 11
2.2 Google’s data fortress ... 13
2.2.1 Scale-induced network effects ... 13
2.2.2 Scope-induced network effects ... 15
2.3 Threat to competitors, the competition process and consumer welfare ... 16
3. Google’s case from a competition law perspective ... 17
3.1 Essential Facilities Doctrine ... 18
3.2. Dominant position of Google ... 20
3.2.1 Upstream market for online search and downstream market for search advertising ... 20
3.2.2 Application of the indispensability test ... 25
3.2.3 Elimination of effective competition in the downstream market ... 29
3.3 The role of data protection in sharing of data mandated by the EFD ... 31
4. Conclusion ... 35
List of abbreviations
ECJ - European Court of Justice EFD - Essential Facilities Doctrine
EU - European Union
GDPR - General Data Protection Regulation
IoT - Internet of Things
1. Introduction
In the world of digital markets, data has an essential role in the development and continuity of a successful business. In fact, the business of some online platforms is so reliant on data collection and analysis that the markets they operate on became known as data-driven markets. A data-driven market can be described as a ‘market characterized by indirect network effects driven by machine-generated data about user preferences or characteristics’.1 One prominent
example of a data-driven market is the market for general online search. Indeed, Google’s business model revolves almost entirely around collection of user data and monetization thereof through state-of-the-art online advertising. Data collection and analysis is linked to network effects and a positive feedback loop which propel the successful online platform ahead of its competitors. While innovation in data-driven markets is in all respects desirable, it appears that it can lead to entrenchment of a dominant position. In the case of a data empire like Google, one could question whether a market entrant will ever have a chance to create effective competitive pressures. In that context, the general question this thesis aims to discuss is the following: should competition law interfere by making the data of a dominant company available to weaker competitors?
Furthermore, in order to consider this issue, the objective of this thesis is to determine whether data can qualify as an essential facility in the meaning of the essential facilities doctrine under European competition law. The first chapter contextualizes the general question of this thesis by exploring why data is so valuable to undertakings on the data-driven markets and how it contributes to development of their businesses. In this context, the merger between Facebook and WhatsApp will be discussed in order to showcase the advantage of data. In order to examine the practical relevance of the general question the second chapter then takes a closer look at the case of Google and discusses the crucial role data plays in preservation of its indisputably dominant position in the market for general online search. Furthermore, in order to address the general question in specific terms, the third chapter of this thesis will apply the essential facilities doctrine to a hypothetical scenario of a request for access to data of a dominant operator of a search engine. Finally, the duty of data protection will be briefly taken
1Jens Prüfer and C. Schottmuller, ‘Competing with Big Data’ (2017) Center for Economic Research Discussion
Paper Vol. 2017-007 <https://research.tilburguniversity.edu/en/publications/competing-with-big-data-2> accessed 3 October 2019, 5.
into account due to its close relevance to the application of the essential facilities doctrine in the context of data sharing.
1.1 Data: the new oil for the digital economy
The aim of this section is to establish that data plays a peculiar role in the modern-day digital markets. The collection of user data has become a widespread practice and a lot of companies invest considerable amounts of money in developing free services for users in order to be able to collect and analyse relevant information.2 The question is: what makes data so valuable that
companies would be willing to provide their services free of charge? After all, data-collection is a mere by-product of the user activity on a particular platform. To answer this question, the following part looks at three examples of current business models which utilize data to produce revenue and improve the quality of their products. Finally, the merger case of Facebook and WhatsApp is discussed as an example of a situation where seemingly harmless combination of datasets can nonetheless lead to a substantial increase in access to data due to network effects.
The former Consumer Commissioner M. Kuneva has stated that ‘Personal data is the new oil
of the internet and the new currency of the digital world’.3 The analogy between oil and data
lays in the idea that both, in their ‘raw’ form’, are mere commodities and require processing to become valuable assets. Just like oil must be turned into gas and plastics through chemical processes, data must be analysed before it can be put to use and be of great value to its holders.4
The extensive practice of data collection is, indeed, an indication of the pro-competitive value of data perceived by businesses.5 The information contained in this modern commodity can
provide its holder with valuable insights into consumer demographics, their behaviour and preferences or even the route of their daily commute. The idea behind extensive data collection is a simple one: a company with information about the behavioural patterns of its customers is
2 Allen P. Grunes and Maurice E. Stucke, ‘No Mistake About It: The Important Role of Antitrust’ (2015) The
University of Tennessee College of Law Research Paper #269
<https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2600051> accessed 12 October 2019, 7.
3 Commission, ‘Meglena Kuneva - European Consumer Commissioner - Keynote Speech - Roundtable on
Online Data Collection, Targeting and Profiling’ (EC Europa, 31 March 2009)
<https://ec.europa.eu/commission/presscorner/detail/en/SPEECH_09_156> accessed 1 October 2019.
4 Tim Berners-Lee and Nigel Shadbolt, ‘There’s gold to be mined from all our data’ (The Times, 31 December
2011) <https://www.thetimes.co.uk/article/theres-gold-to-be-mined-from-all-our-data-s3qffdkz7kq> accessed 2 October 2019.
5 Andres V. Lerner, ‘The Role of 'Big Data' in Online Platform Competition’ [2014] SSRN Electronic Journal 1,
likely to be able to provide better goods and services than its competitors. So normalized has the collection of consumer data become that it is now a widespread practice exercised by both online businesses and traditional brick-and-mortar shops.6 In the world of online platforms,
tracking customer activity has become the norm thanks to the ability of computers to perform this task automatically. The collected data varies from that which requires no direct human input, such as user’s geo-location or online behaviour, to voluntarily surrendered personal data including shopping preferences or taste in music. In addition to the providers of online services to whom customers come directly, data is increasingly being collected by third party firms such as data-brokers and advertising agencies for strictly commercial purposes.7
1.2 Understanding the nature of data
To understand the importance of data in the modern economy, it is paramount to first define what data is and how it contributes to business development. The term data, in the context of digital economy, encompasses a wide range of digital inputs recorded by companies which can be analysed and reveal useful information about the nature of the business. According to Drexl, a notable distinction should be made between ‘data’ and ‘information’. The term data refers to the raw records made by, for instance, a sensor of a self-driving car after it has spotted a pothole on the road. The record stored in form of digital data on the server of the car manufacturer does not yet constitute information. The information that can be extracted from this record is indeed that there is a pothole on the road. Before this information can be extracted, however, the data containing it first must be deciphered in a manner adequate to reveal it. The distinction between data and information appears to make a big difference from a competition-oriented perspective. While data can be relatively easily replicated due to its nature, extracting information from data requires sophisticated analytical tools. According to Gunnes and Maurice, the value of data for a company depends on the latter’s ability to analyse it and act quickly on the revealed information. In fact, data and data analytics have a mutually reinforcing relationship. The more data an algorithm analyses, the better its capacity to analyse and extract information from future incoming data.8 In this fashion, companies in possession of considerable amounts of data on
their customers find themselves climbing up a learning curve which allows them to improve the quality of their business and, consequently, gather even more relevant data.
6 Ibid 9. 7 Ibid 8.
For the purpose of this thesis, the term ‘data’ simply refers to ‘digital data’ which can be understood as records of various inputs collected by undertakings that contain ‘machine-readable encoded information’9. Moreover, due to the legal nature and limited capacity of the
thesis, no technical distinction between raw data and analysed data will be made. It should be noted, nonetheless, that in practice such distinction might be desirable for application of the existing competition law framework to data. For the purpose of simplicity of this thesis, however, the term data will refer to both the data itself and the information that can be retrieved from it.
1.3 The Advantage of data
The significance of data for the success of a business can be well observed when considering online platforms such as social networks and search engines. It has been claimed that the competitive strength of online platforms is being increasingly dictated by the amount and quality of data they are able to collect and analyse.10 Many online platforms operate their
business on a model of a double-sided network. A double-sided network is an economic platform which has two distinct user groups who provide each other with network benefits.11Social media platforms like Facebook, for example, generally attract two main user
groups: users (natural persons making use of the social platform to connect with others) and advertisers (legal or natural persons coming to the platform to advertise their product to the other users of the platform). One of the distinguishing features of this type of platform from other business models is that the users are able to use the social platform ‘free of charge’. Instead of charging money, the provider allows customers to use the platform in exchange for the data they will share and make available on it. It is through the use of this data, that the provider will be able to make the platform attractive for the other user group, advertisers. The more users and data the provider gathers, the better and more accurately targeted to specific users can the advertisements shown on the platform be. On the user side of the platform, the provider tries to deliver relevant services which capture the customers’ interest. In that regard,
9 Josef Drexl, ‘Designing Competitive Markets for Industrial Data – Between Propertisation and Access’ (2017)
Max Planck Institute for Innovation & Competition Research Paper No. 16-13
<https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2862975> accessed November 16 2019, 263.
10 Inge Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (PhD thesis, KU
Leuven 2016) 129.
11 Andrei Hagiu and Julian Wright, ‘Multi-Sided Platforms’ (2011) Harvard Business School Working Paper
12-024, <https://web.archive.org/web/20111112093216/http://www.hbs.edu/research/pdf/12-024.pdf> accessed 3 November 2019, 5.
the provider analyses existing data of a given user to optimize search results, suggested social network interactions or recommendations for interest groups and future purchases.12 Data,
therefore, plays a pivotal role in the development of the range of services the platform is able to provide to both categories of its users.
Similar to social media platforms, search engines provide their services to users free of charge and make money through contextual advertising. An advertiser will make a bid on different words which, when entered into the search query, will trigger the search engine to display the relevant ad.13 Search engines collect information about its users, such as time and date of the
search, issued queries and clicked results, in order to keep the search results and advertisements up-to-date and relevant.14
Furthermore, with the emergence of the Internet of Things, data-collection and analysis has experienced a major boost. On the IoT, various devices are connected to one another and their actions and inter-actions generate large amounts of data which can be of great value to businesses. At today’s stage of technological development, data generated through a product of one company for a specific purpose can become an important asset for another economic player for a very different purpose.15 For example, the software of a smart car collects a wide
range of data about, inter alia, the environment surrounding the vehicle, its geographical location and driving patterns of the driver. In addition to the value of this data for the car producer to improve its product, it can prove to be of great importance to other service providers. Geo-location of the car can inform map creators, such as Google Maps, about, for instance, a change of the direction of a one-way road; ensuring their service is up-to date with recent changes to road traffic rules. An insurance company might find it valuable to have access to an individual’s driving history when rendering an appropriate insurance package.16 When
addressing the value of data in the digital economy, therefore, the debate is not limited to data directly provided by individuals but extends to wide variety of data collected by all objects connected to the IoT.
12 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 129. 13 James Grimmelmann, ‘The Structure of Search Engine Law’ (2007) 93 Iowa Law Review 1.
14Fabrizio Silvestri, ‘Mining Query Logs: Turning Search Usage Data into Knowledge’ (2009) 4 Foundations
and Trends in Information Retrieval 1, 3.
15 Josef Drexl, ‘Designing Competitive Markets for Industrial Data – Between Propertisation and Access’
(2017) Max Planck Institute for Innovation & Competition Research Paper No. 16-13
<https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2862975> accessed November 16 2019, 263.
Though concrete cases of competition law related issues in the context of data-driven markets are still relatively sparse, the European Commission has already had a chance to present its view on the issue of a merger in a data-driven market in the Facebook/WhatsApp case. Thus, the following section aims to further showcase the advantage of data by discussing the merger of WhatsApp and Facebook.
1.4 The Merger of WhatsApp and Facebook
The merger of WhatsApp and Facebook is an example of a situation where despite the perceived lack of traditional entry barriers in the relevant market, the approved merger resulted in a significant increase of dataflow for the merged entity.
In its decision, the Commission, elaborating on its finding of the harmless nature of the merger, considered that in the market for communication apps, barriers to entry and expansion for competitors as well as switching costs for customers are very low.17 First, it argued, ‘consumer
communications customers have a broad range of choices when it comes to selecting and using consumer communications apps’.18 The Commission found that the market for communication
apps is dynamic, fast-growing and characterized by disruptive innovation. Therefore, it has been argued that there are no significant ‘traditional’ barriers for a new company to enter the market.19 A significant part of the argument was based on the idea of ‘multi-homing’ and the
Commission considered, in that context, that the merged entity will not become the exclusive communication app provider since ‘competitors will be able to gain users even though those
users don’t abandon the merged entity’s network’.20
The following facts have been raised by Stucke and Grunes in a response to the reasoning of the Commission.21 Despite multi-homing being a real phenomenon on the communication apps
market, the user base for both Facebook Messenger and WhatsApp raised rapidly after the merger went through. Messenger’s active users jumped from 600 million to 700 million in just three months and WhatsApp crossed the line of one billion users. Moreover, the fact that a user has accounts on multiple platforms does not mean he or she will use both equally. For instance,
17FACEBOOK/ WHATSAPP Commission Decision COMP/M.7217 [2014] OJ L-2985, para 94.
18 Ibid 87. 19 Ibid 116-117.
20 Stucke and Grunes (n 8) 167. 21 Ibid 167-169
in August 2015, in Ireland, many respondents of a survey had both Facebook Messenger (46 per cent) and Skype (43 per cent). Only 9 per cent of those who also had Messenger, however, used Skype daily. Usage of Messenger turned out to be far more frequent, with 49 per cent of users visiting the service daily.22 Larger user base invariably leads to production and harvesting
of more user-generated data.
In reality, none of the listed facts amounts to anti-competitive behaviour. It seems, however, that even though the combination of datasets of two large incumbents, Facebook and WhatsApp, did not raise competitive concerns, it was followed by a considerable growth of both platforms. One unique feature of online platforms is precisely the fact that they are easily accessible and undertakings that operate them will often try to increase the interconnectivity between their own platform and those of other companies. In doing so they can increase the number of ‘portals’ through which they can access customers. In the case of the Facebook/WhatsApp merger, Facebook has ensured access to WhatsApp’s staggering number of users; over 500 million people use WhatsApp on monthly basis.23 In the context of a business
model of online platforms like Facebook, which focuses on targeted advertising for its revenue, the acquisition of WhatsApp can be seen as a strategic move that, although not leading to a direct increase in market power, will ensure access to significant user dataset no longer belonging to a competitor. Just how relevant this dataset is for improvement of Facebook’s position in the market for communication platforms is yet to be seen. It is clear, however, that communication platforms are merely one of the multitude of services offered by undertakings on the digital markets. As it happens, a combination of multiple online services is often provided by a single undertaking. In any case, the ways in which collection of data on one platform can directly contribute to the attractiveness of another one will be explored further in the example of Google’s search engine.
1.5 Conclusion
The positive effects on the ability of an undertaking to strengthen its position in the market due to an increase in dataflow have been demonstrated in the previous section.
22 Ibid 167-168.
23 Alison L. Deutsch, ‘WhatsApp: The Best Facebook Purchase Ever?’ (Investopedia, 25 June 2019)
<https://www.investopedia.com/articles/investing/032515/whatsapp-best-facebook-purchase-ever.asp> accessed 20 October 2019.
It is evident that data can provide a boost for virtually any business in terms of development and improvement of their services. This can, in turn, enhance competition and lead to development of new markets. In such a case, an important question arises. Namely, whether it is desirable that data is made available to companies that might not be in a position to collect it effectively themselves.
In order to discuss this question in more detail, this thesis will focus on a single scenario, and in particular on the benefits of data present on search platforms.
The data-enhanced innovation present in these markets is significant and may lead to creation and entrenchment of a dominant position. The relevant factors for assessment of a dominant position in data-driven markets and specifically the markets for online search and search advertising will be, thus, discussed in the second chapter on the example of Google.
Nevertheless, before discussing how Google’s activities can be used to establish and maintain dominance in a digital market, it is first important to consider the goals of competition law enforcement.
2. Data collection as a threat to competition
2.1 Why do we regulate competition?
In order to identify how data collection can lead to harm to undisturbed competition between undertakings active on data-driven markets, this chapter will first revisit some of the fundamental ideas behind competition law. The stance taken on this matter by the Court of Justice in T-Mobile is that competition law is ‘designed to protect not only the immediate interests of individual competitors or consumers but also to protect the structure of the market and thus competition as such’.24 This statement indicates there are at least three objectives of
competition law: protection of competitors, protection of consumers and protection of the competitive process itself. Protection of competitors is motivated by the idea that a smaller, entrant, company might not be able, absent of government intervention, to achieve the
24 Case 8/08 T-Mobile Netherlands BV and Others v Raad van bestuur van de Nederlandse
necessary scale to maintain itself on the market before it is eradicated by the natural forces of unregulated competition.
A certain degree of dichotomy can be spotted between the first and the last objectives, namely protection of competitors and protection of the competitive process itself. The protection of competitors, by definition, interferes with the competitive process as it would naturally unfold had regulation not been present. The ideal outcome of regulating competition is a balanced state in a market where competition between undertakings arises as a consequence of the natural process of rivalry trough improvement of products and lowering of prices. At the same time, competition law aims to interfere where this process would be distorted by anti-competitive behaviour of an undertaking with disproportionate powers in the market. The CJEU has expressed this idea in Intel v Commission as follows: ‘not every exclusionary effect
is necessarily detrimental to competition. Competition on the merits may, by definition, lead to the departure from the market or the marginalization of competitors that are less efficient and so less attractive to consumers from the point of view of, among other things, price, choice, quality or innovation’. In other words, competition by means of innovation is desirable. The
concept of ‘fair competition’ is also of relevance here. It would be considered ‘unfair’, for instance, if an undertaking indulged in a practice intended to denigrate competitor’s product and thereby distort competition.25 On one hand, competitive threats can force a company into
developing cutting-edge products for ever-lower prices. An outcome that is considered desirable. Such success, however, especially solitary, will often lead to establishment of a dominant position for its reaper. On the other hand, a dominant position can be abused to forcefully ensure its retention which is considered to be an undesirable outcome. In addition to protection of competitors and competition itself, the core goal pursued by competition law has been said, in the recent years, to be the protection of consumer welfare.26 This reasoning seems
logical as, forgetting for a moment the economic prosperity and innovation harvested by the wider society, it is the consumers who, ultimately, benefit from the abundance of choice available on the market or suffers from the scarcity thereof.
In the following sections, taking into account the above listed goals of competition law, this thesis will consider reasons why competition authorities should be interested in the effects of
25 Richard Whish and David Bailey, Competition Law (9th edition, Oxford University Press, 2018) 23.
data collection on competition and explores how data is already changing the ways in which companies are able to gain dominance and, in particular, considers Google’s case as the main example.
2.2 Google’s data fortress
Google, as the biggest owner of multimedia in the world, benefits greatly from the variety of platforms it offers its services on.27 It is due to its unprecedented position in terms of its ability
to collect and utilize user data that this thesis examines its example as a case where the essential facilities doctrine may provide remedy to distortion of competition. This section explores in more detail how network effects experienced by an online search platform lead to increase in scale and scope in data-driven industries. Its aim is to demonstrate that data collection is likely to lead to establishment of dominant position and, moreover, that high-volume data collection is accompanied by increase in entry barriers from the perspective of a market entrant.
The basic idea of network effects is that the more users use a platform, the better off will its other users be. Similarly, the more users enter search queries to Google’s search engine, the more data its algorithm has to improve itself and the better the search query results for all users will be. In the context of a data-driven market, one can observe two types of network effects; scale and scope induced. While the former relates to the volume of users, and their data, that a platform has access to, the latter relates to the diversity of subject matters the data contains as a result of the variety of portals through which the data is collected. For example, in addition to its search engine, Google collects data through a variety of platforms including Gmail, Google Maps and YouTube.
2.2.1 Scale-induced network effects
A significant advantage of a search engine is that it can reap the benefits of data-driven network effects beyond any single local market.28 This is demonstrated by an example of a resident of
London who searches for images of Big Ben. Once the search results are displayed to the user in form of web links, images etc., they will be able to precisely identify which items from the offered list of option are relevant to their query and choose those. The search engine will then
27 Stucke and Grunes (n 8) 211. 28 Ibid 172.
make a record of the search query as well as the subsequent behaviour of the user in terms of the selected results. The next time someone searches for Big Ben, the relevance of the results displayed will have been directly induced by the Londoner’s search and subsequent choice of result.29 This phenomenon can be described as network-effects induced by the scale of data.
The underlying mechanism is the positive feedback loop created by the repeating trial-and-error process which leads to improvement of the algorithm operating the search engine.30
Google provides search services to users free of charge and competes with other companies on the merit of quality of the results it can display. It relies on the input of its users for further development of its search engine. The underlying algorithm gets an opportunity to re-organize the hierarchy of results it shows upon the entry of a particular query by a user. If a link originally displayed at the bottom of the page attracts the attention of users, the algorithm will re-evaluate its relevance and move it up in the display ranks.31 The feedback Google receives
directly enhances, thus, traffic volume on search queries.
When Microsoft and Yahoo! decided to merge in 2010, the Commission was presented with arguments which confirm the importance of scale for success in the search engine industry. It was argued that the merger would have positive effects on competition in the market for general online search as it would provide the merged entity with a better position to compete with the strong market leader, Google.32 The increase of scale resulting from the merger of the two
platforms, was argued, would make the entity a more credible alternative to Google.33
Microsoft argued that prior to the merger each of the parties was held back from improving its search service by the low volume of search queries it received. Conversely, a ‘large’ search engine had better chances at spotting relevance issues of their search results as it possessed a larger sample of queries to learn from.34 The lack of queries has been argued to be an important
factor specifically in the case of less-frequently searched or as referred to by the European Commission in Microsoft/Yahoo - ‘tail’ queries.35
29 Ibid 203. 30 Ibid 170. 31 Ibid 173.
32MICROSOFT/ YAHOO! SEARCH BUSINESS Commission Decision COMP/M.5727 [2010] OJ L-2985, para
211.
33 Ibid 160. 34 Ibid 162.
To explain why a smaller search engine might not be able to improve fast enough to capture the attention of a wide scale of customers, let us assume that two search engines, a larger and a smaller one, attract a random cross-selection of users for their internet search platform. The larger search engine receives, on average, 5 billion search queries a day while the smaller one only receives 5 million. For more popular topics, both search engines might receive a sufficient number of queries, say 50 million and 50,000 respectively. For less-popular, interest specific, ‘tail-queries’, however, the smaller search engine might only receive 500 queries per day, as opposed to the larger search engine’s 500,000. Consequently, the smaller search engine might perform just as well as the larger one on searches for popular topics but struggle greatly to provide relevant results for less-popular topics.36 Under this scenario, the larger search engine
will directly benefit, not only from the size of its current userbase but also from every unsatisfied user of the smaller search engine who seeks information about interest-specific topics and chooses to look for alternatives on the market.37
Despite Microsoft’s merger with Yahoo!, Google’s dominance in the market for search engines remains unthreatened. In 2013, the Commission has observed that Google has over 90% share in the market for its general web search service in all of European Economic Area.38 In
September 2019, when Bing and Yahoo! held 2.25% and 0.8%, respectively, Google’s share on the European market was reported to be a staggering 93.68%.39
2.2.2 Scope-induced network effects
Another aspect of Google’s success can undeniably be attributed to the variety of platforms it provides its services on. It is said, for example, that Google scans the content of emails sent or received by its Gmail service to create user profiles and provide targeted advertising.40
Activities undertaken by a user active on a Google platform are taken into consideration and will be reflected in that user’s search results.41 Google’s platforms are diverse. Users can search
36 Ibid. 37 Ibid.
38 Commission, ‘Antitrust: Commission seeks feedback on commitments offered by Google to address
competition concerns’ (EC Europa, 25 April 2013)
<https://ec.europa.eu/commission/presscorner/detail/en/IP_13_371> accessed 15 December 2019.
39 Statcounter, ‘Search Engine Market Share Europe October 2018 - October 2019’
<https://gs.statcounter.com/search-engine-market-share/all/europe> accessed 11 October 2019.
40 Stucke and Grunes (n 8) 188.
41 Frank Pasquale, The Black Box Society: The Secret Algorithms That Control Money and Information
YouTube for videos, send emails via Gmail and find their intended destination on Google Maps, amongst others.42 Each of these activities are different facets of a user’s interest profile
which Google creates and uses to improve its services.
The majority of Google’s revenue comes from monetizing the advertisement space it offers on its platforms. In 2014, advertisement generated income accounted for 89 per cent of Google’s total revenue.43 The predominant business model is the ‘cost-per-click’ monetization where
Google charges the advertiser each time a user actually clicks on the ad displayed to them.44 In
using such model, understanding user preferences is key to delivering competitive advertising services. The variety of platforms owned by Google increases the diversity of user data it collects and analyses. This then translates into improved quality of advertisements it can offer to its advertising customers. As Google puts it, the goal of AdWords, its main auction-based advertising program is ‘to deliver ads that are so useful and relevant to search queries or web
content that they are a form of information in their own right’.45 All of this demonstrates that
data has the power to fuel the positive network effects in a digital market. The model is simple: all else equal, the more data a company has the faster it is able to improve and expand its services and the more money it will be able to receive from advertisers.
2.3 Threat to competitors, the competition process and consumer welfare
The competitive strength Google gains as a result of data-collection is undisputable. Before it is discussed whether competition law should interfere, this section will return to the three subjects competition law aims to protect, namely competition, competitors and consumers, and reflect on how Google’s activities could be affecting them.
Starting with the process of competition itself, it appears that companies like Google have been able to gain a dominant position in various digital markets. The market share of Google in the market for search engines has been oscillating around 90% for the past ten years.46 Google has
42 Ibid 68.
43 Stucke and Grunes (n 8) 196.
44 Google’s Annual Report Pursuant to Section 13 Or 15(D) of the Securities Exchange Act of 1934 (for the
fiscal year ended December 31, 2013)
<http://www.annualreports.com/HostedData/AnnualReports/PDF/NASDAQ_WSTG_2018.pdf> accessed 15 November 2019, 4.
45 Ibid.
46 Statcorner, ‘Search Engine Market Share Worldwide January 2009 - September 2019’
expanded the catalogue of its services and the platforms these services are offered on. Despite the existence of other factors explaining Google’s success, data-collection has been argued to have been a key element propelling the company forward. At the present-day Google receives billions of search queries every day.47 Recalling the competitive power such voluminous
dataflow creates, the question arises whether one can speak of meaningful competition between a market entrant and Google at all. It appears that competition in the market for search engines has been stalled if not eradicated. In that sense, it could be argued that the widespread data-collection practice exercised by tech-giants today is highly inducive to market dominance and distortion of competition.
However, dominance in and of itself does not constitute a breach of competition law and can even have positive effects for product development and consumer welfare. After all, an undertaking that does not have to focus on its economic survival can invest more time and resources into perfecting the product it offers on the market. In any case, it seems reasonable to conclude that the state of data-dominance in data-driven markets is an undesirable phenomenon, at least from the perspective of protection of competitors and the competitive process itself. The existence of positive effects on innovation of data-induced network effects and the consequent benefits to consumers are acknowledged but will not be elaborated on in this thesis.
The following chapter will examine the applicability of the existing European competition law regime on dominance, and the abuse thereof, in the context of data-driven markets. Special emphasis will be put on Google’s dominance in the market for search engines.
3. Google’s case from a competition law perspective
The previous chapter demonstrated that Google’s capacity to collect user data and utilize it to further improve and expand the variety of its services is unprecedented. It is due to the existence of network effects and a positive feedback loop that this extensive data collection poses a threat to competition. Data of each user that Google is able to attract to its platform will be missing
47 Danny Sullivan, ‘Google now handles at least 2 trillion searches per year’ (Search Engine Land, 24 May
2016) <https://searchengineland.com/google-now-handles-2-999-trillion-searches-per-year-250247> accessed 12 October 2019.
from a dataset of its competitor who is not able to provide an adequate alternative. Although multi-homing is argued to be a common practice on online platforms due to low switching costs for users, practice shows that Google currently does not have a real competitor in the market for general online search. Due to the lack of data about user search preferences available to them, other market players are simply not able to improve their service fast enough, relative to Google, to be able to provide a viable alternative on the market. Data is, indeed, so important for companies providing online search and search advertising services that one cannot help but consider it as essential. It is for this reason that this thesis examines the essential facilities doctrine as a possible remedy to the competitive threats unfolding on data-driven markets.
3.1 Essential Facilities Doctrine
The concept of the ‘essential facilities doctrine’ stems from the idea that dominance in one market can be used as a leverage to ensure dominance in another, downstream market.48 The
holder of the essential facility tries to expand the dominance it has in the upstream market into a derivative, downstream, market precisely by refusing to supply its competitor on the latter market with a facility necessary for conducting business there.49
Also known as ‘refusal-to-deal’, the doctrine stands as an exception to the freedom to choose one’s contracting partner that the dominant undertaking normally enjoys50 It encompasses a
wide range of facilities the withholding of which by a dominant undertaking can be found to disrupt competition.51 In fact, there is no exhaustive list of scenarios in which the doctrine
ought to be applied. This is because the essential facilities doctrine was not introduced in European legislation on competition law but has, instead, been developed, in various forms, through the jurisprudence of the European Court of Justice. As of the present day, however, it has never been applied in the context of a request to access to data of a competitor.
48 Jonathan Faull and Ali Nikpay, Faull and Nikpay: The EU Law of Competition (OUP 2014) 464.
49 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 209. 50 Faull and Nikpay (48) 463.
The essential facilities doctrine was originally applied to traditional infrastructure such as ports in Sealink52 or raw production materials in Commercial Solvents53. In the landmark case Oscar
Bronner, the ECJ clearly set out the conditions for application of the doctrine.54 Firstly, the
refusal of providing access must ‘be likely to eliminate all competition […] on the part of the person requesting the service’. The elimination, in this case, relates to the downstream market on which both the requestor and the dominant undertaking operate. Secondly, the refusal must ‘be incapable of being objectively justified’. Thirdly, the facility must be ‘indispensable’ to such a degree that there must be ‘no actual or potential substitute in existence’ for it.55
With the progression of technology and subsequent development digital markets the nature of facilities that have been claimed to be essential to conduct business became increasingly more abstract. In Microsoft, for instance, the General Court was asked to rule on Microsoft’s obligation to disclose the interoperability information with its Windows operating system with competitors.56 Taking into consideration the power undertakings controlling large datasets already have in existing digital markets and foreseeing a widespread emergence of IoT products, it could be only a matter of time until a dominant competitor’s dataset is claimed to be essential by a market entrant.
The following section will outline the essential facilities doctrine as it exists today and attempt to apply it to a hypothetical scenario in a data-driven market. The hypothetical scenario used will be that of an undertaking active on the market for online search advertising (the requestor), requesting access to the data collected and controlled by a dominant undertaking active on the market for general online search (the facility holder). The facility holder would in this case, in addition to being active on the market for general online search, also be active on the market for online search advertising. This hypothetical scenario is intended to imitate the currently real situation where Google, as the dominant search engine provider in on the European market, is also active on the market for online search advertising. This scenario allows for a detailed consideration of Google’s case, thereby showcasing the practical significance of the general question of this thesis.
52 Sealink/B&I Holyhead: Interim Measures (Case IV/34.689) Commission Decision 94/19/EC [1992] OJ L
378.
53 Joined Cases 6 and 7-73 Istituto Chemioterapico Italiano S.p.A. and Commercial Solvents Corporation v
Commission of the European Communities [1974] ECLI:EU:C:1974:18.
54 Case 7/97 Oscar Bronner, Opinion of AG Jacoba [1998] ECLI:EU:C:1998:264, para 35.
55 Ibid 41.
3.2. Dominant position of Google
In the interest of keeping the hypothetical scenario as close to the real situation on the two relevant markets as possible, Google will be used as a benchmark of a dominant undertaking active on the upstream market for general online search. A precondition to assessing applicability of the essential facilities doctrine to a specific case is the existence of a dominant position of the facility holder on the upstream market. Google’s dominance in the market for general online search is undisputed at the present day. The European Commission has identified the tech-giant’s dominance in various online markets in several of its press releases.57
In the context of online search specifically, the Commission has identified a market for general online search and found that Google holds over 90% market share in this market in various member states.58
Thus, as there has been no dispute about the ability of Google to maintain a dominant position in the market for general online search, the existence of dominance will be assumed for the purpose of the following analysis of the hypothetical scenario at hand.
3.2.1 Upstream market for online search and downstream market for search advertising
In addition to existence of a dominant position by an undertaking on the upstream market, applicability of the essential facilities doctrine requires existence of a downstream market. Two relevant, separate, markets thus have to be identifiable before the doctrine can be applied to the hypothetical scenario outlined in the previous section.
There are three issues this section aims to address. Firstly, it deals with the question whether the market for general online search and the market for online search advertising constitute two separate markets for purposes of applicability of the essential facilities doctrine. Secondly, whether the relation between the market for general online search and the market for online search advertising is such that the former can constitute an upstream market and the latter a
57 Commission, ‘Antitrust: Commission fines Google €4.34 billion for illegal practices regarding Android
mobile devices to strengthen dominance of Google's search engine’ (EC Europa, 18 July
2018) <https://ec.europa.eu/commission/presscorner/detail/en/IP_18_4581> accessed 15 December 2019.
58 Commission, ‘Antitrust: Commission fines Google €1.49 billion for abusive practices in online advertising’
(EC Europa, 20 March 2019) <https://ec.europa.eu/commission/presscorner/detail/en/IP_19_1770> accessed 12 November 2019.
downstream market. Finally, this section shortly addresses the possibility of defining a hypothetical market for data as an alternative upstream market.
Up to the present day, a distinction between different types of markets for online search has not been made in the context of applicability of the EFD. Inspiration will, therefore, have to be drawn from existing decisions of the European Commission to identify the relevant markets for the hypothetical scenario at hand.
In addition to the already mentioned decision of the European Commission where it found Google to have above 90% market share in the market for general online search59, another
relevant market categorization can be found in an investigation the Commission started on Google in 2010. There, the Commission made a distinction between two interrelated markets with regards to Google’s search engine: the market for internet search and the market for online search advertising.60 In a related press release, the Commission expressed a concern that
Google was abusing its dominant position in three interrelated online markets, namely those for web search, online search advertising and online search advertising intermediation.61
Furthermore, in Google Shopping, the Commission has considered that Google’s favourable treatment of its comparison-shopping service in the results displayed by its search engine constitutes an abuse of Google’s dominant position in the general internet search market. In other words, the preliminary finding was that Google’s dominance in one market, namely the one of general internet search, was used to create advantage in a related market for vertical search for online shopping.62 Commission’s findings imply existence of two related but
separate markets: the one for general internet search and the one for vertical search, in this case, for online shopping.
59 Commission, ‘Antitrust: Commission fines Google €1.49 billion’ (n 57)
60 Inge Graef, ‘Market Definition and Market Power in Data: The Case of Online Platforms’ (2015) 38(4) World
Competition: Law and Economics Review 473, 491
61 Commission, ‘Antitrust: Commission seeks feedback on commitments offered by Google to address
competition concerns’ (EC Europa, 25 April 2013)
<https://ec.europa.eu/commission/presscorner/detail/en/IP_13_371> accessed 10 October 2019.
62 Commission, ‘Statement by Commissioner Vestager on antitrust decisions concerning Google’ (EC Europa,
15 April 2015) <https://ec.europa.eu/commission/presscorner/detail/en/STATEMENT_15_4785> accessed 10 December 2019.
The findings of the European Commission lead to the conclusion that there exist, at the present day, various markets related to online search on which Google is active and, in certain cases, holds a dominant position. Out of the variety of markets identified, two markets are relevant for the purpose of analysing applicability of the EFD to the hypothetical scenario at hand, namely market for general online search and market for online search advertising.
The second question that remains to be answered is whether the relation between the market for general online search and online search advertising is such that the former can constitute an upstream market and the latter a downstream market.
As explained before, the idea behind the essential facilities doctrine is prevention of leveraging of dominant position in one market by strengthening one’s position in another, derivative market. The derivative or downstream, market should comprise of the product that the requestor will offer once it has been granted access to the required.63 The Commission’s
reasoning in Sealink demonstrates that the delineation of the upstream market will depend on how the downstream market is defined. In that case, a ferry operator, B&I, requested access to the port of Holyhead in north Wales in order to operate its ferry service to and from Ireland. The port was owned by Sealink, which also operated on the market for ferry services. At the time, there were three routes ferries between Ireland and the Great Britain operated on. One of them, the central corridor route, predominantly used port Holyhead for departures from and arrivals into the Great Britain. For the purpose of that case, the Commission has identified a downstream market for ferry services to and from Ireland via the central corridor route. The Commission then, on basis of the downstream market identified, defined the upstream market as one for provision of port facilities for passengers and ferry services on the central corridor route. It was the terms in which the Commission defined the downstream market that allowed for the narrow interpretation of the upstream market and the finding of indispensability of Holyhead port for provision of services by B&I. Had the Commission considered the downstream market to be one for all ferry services between Ireland and Great Britain, including ones operated on the other corridor routes, the central Holyhead port would not have been seen as essential facility for B&I to provide its services.64
63 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 254. 64 Whish and Bailey (n 25) 717.
In the context of online platforms and for the purpose of this thesis, a downstream market for online search advertising will be used to analyse applicability of the essential facilities doctrine to an access to data request. As explained earlier in this thesis, the business of online advertising relies on utilizing user data to match the content of paying advertisers with users searching for information on an online platform such as Google’s search engine. It is precisely and solely the data collected on its search engine platform that allows Google to provide targeted-advertising services. There exists, therefore, a clear hierarchy between the business of online search which is a source of user data and the business of online advertising which so inherently relies upon this data. In fact, the business of online search advertising is derivative of the business of online search in the literal sense of the word. It is on basis of this observation that this thesis argues that the markets for general online search and online search advertising are in such relation to one another that the former constitutes an upstream market and the latter a downstream market for purposes of application of the essential facilities doctrine.
In contrast to already identified markets such as that for general online search, literature has discussed the idea of a hypothetical market for data itself.65 The concept of a potential or
hypothetical upstream market was developed by the European Court of Justice in the IMS
Health case.66 The jurisprudence on the essential facilities doctrine prior to IMS Health
required existence of a factual upstream market on which the requested facility is or has been traded. With potential and hypothetical upstream market as an option, it is no longer necessary for the requested input to have been already traded as an independent product by the dominant undertaking.67 In theory, therefore, the fact that online providers like Facebook or Google do
not actively license their user data to third parties like, for example, Twitter does, does not preclude competition authorities from imposing an obligation on these providers to grant competitors access to their datasets.68 Consistently with the IMS Health judgment, the
European Commission, in its Guidance on enforcement priorities in applying Article 102, states that it is not necessary for the refused product to already have been traded; it is sufficient that there is demand from potential purchasers.69 A request to purchase a dominant search engine
provider’s data could constitute such demand.
65 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 115-116
66 Case C-418/01 IMS Health GmbH & Co. OHG and NDC Health GmbH & Co. KG [2004]
ECLI:EU:C:2004:257.
67 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 254. 68 Ibid.
69 Communication from the Commission — Guidance on the Commission's enforcement priorities in applying
Moreover, Graef has argued for consistency of defining a market for data with the concept of ‘specialised assets’ listed in the EU Horizontal Guidelines70 as one of the factors for
determination of competitive credibility of an R&D pole.71 The application of the concept of
specialised assets is said to make the definition of the relevant market dependent on the capabilities and resources that are necessary for a company to have in order to be able to innovate.72 Acceptance of an interpretation of the concept of specialised assets as
encompassing data would allow competition authorities to take a form of potential competition into consideration whereby online platform do not merely compete in the product markets for specific services, but compete also in a broader market for data itself; data which can be deployed to improve quality and relevance of services they offer.73 Such interpretation would
constitute the forward-looking approach to market definition competition on data-driven markets requires.
Definition of a market for data, however, could be subject to normative challenges. The expected difficulty in defining a market for data is the perceived lack substitutability of different sorts of data. Traditionally, a relevant market consisted of all the products which are found to be substitutable for one another on basis of their characteristics, prices and intended use.74 The European Commission, however, has so far not made a distinction between different
types of user data. In fact, it seems to consider online data as a homogenous commodity. For example, in Google/DoubleClick, the Commission referred to different online portals by companies among which were Microsoft and Yahoo as players who ‘have the ability to collect
large amounts of more or less similar information that is potentially useful for advertisement targeting’.75 Similarly, in the Facebook/WhatsApp merger decision, the Commission
considered Google, Apple, Amazon, eBay, Microsoft, etc. as market participants collecting user data alongside Facebook.76 On basis of the above observation, Graef has argued that the
70 Communication from the Commission - Guidelines on the applicability of Article 101 of the Treaty on the
Functioning of the European Union to horizontal co-operation agreements (EU Horizontal Guidelines) [2011] OJ C11/1
71 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 115. 72 Ibid 114
73 Ibid 115
74 Commission Notice on the definition of relevant market for the purposes of Community competition law
[1997] OJ C 372/5, para 7.
75 Inge Graef, ‘Market Definition and Market Power in Data: The Case of Online Platforms’ (n 60) 500. 76 Facebook/WhatsApp (n 17) 188-189
Commission considers data collected by companies active on the internet as substitutable.77
There are, however, important differences to be identified between different types of data and the way it is utilized in different online markets. A distinction relevant to the analysis at hand can be observed, for example, between data gathered by a social network provider about the demographics and interests of its users which it uses for targeted advertising on social media platforms, and the information about user preferences gathered by a search engine through observation of entered search queries and subsequent result selection which the provider uses to display relevant ads. The substitutability between the former and the latter type of data seems relatively limited.78
To summarize, it is not yet clear whether and how could an upstream and downstream market be defined in case of a claim for a dataset as an essential facility. In case of a downstream market for online search advertising, it seems logical to look for dominance in an upstream market for general online search as this market is the source of data so crucial to search advertising, namely the search behaviour of users. An alternative scenario proposes identification of a hypothetical market for user data itself. In such a scenario, the downstream market could be directly the one for general online search. Such categorization could be meaningful as it would allow potential competitors of Google to request its data directly for the purpose of competing on the market for general online search, the market where Google has consistently held a market share of around 90%.
For the purpose of the case study at hand, nonetheless, the more conventional approach to market delineation will be used. The upstream market will be that for general online search while the downstream market will be that for search advertising. The main service on the latter market is targeted advertising which is provided on basis of the user data collected on the upstream market for general online search.
3.2.2 Application of the indispensability test
In assessing whether a requested facility is essential for conducting business on the downstream market, the ECJ will apply the so-called ‘indispensability test’. A facility was held to be indispensable by the Court when there are no alternatives to conduct the same business without
77 Inge Graef, ‘Market Definition and Market Power in Data: The Case of Online Platforms’ (n 60) 500. 78 Ibid 498
it,79 and if there are ‘technical, legal or even economic obstacles capable of making it
impossible, or even unreasonably difficult to duplicate the requested facility.80
In line with the indispensability test established by the ECJ, the European Commission considers a refusal to supply a facility to be an enforcement priority in cases where the facility concerned is ‘objectively necessary to be able to compete effectively on a downstream market’.81
Indispensability of data has not yet been examined by the European Commission; its past considerations of the nature of data, however, give some indication on the potential standpoint of the Commission in case of an essential facility claim. As described previously, the Commission seems to generally categorize online data as a homogenous commodity the collection of which is a widespread practice exercised by numerous players active on digital markets. Such conclusion can be drawn from the line of argumentation used in merger cases like Google/DoubleClick and Telefonica/Vodafone/EverythingEverywhere, where the Commission adopted the idea that user data will continue to be available to competitors even after the combining of datasets of large incumbents as a result of a merger.82 In
Google/DoubleClick the Commission argued that ‘similar data’ was at the time available to
Google’s competitors including Microsoft and Yahoo! and, if needed, could be also be purchased from third party service providers.83 If data is considered in terms of being a
homogenous facility collected widely by various internet players, a finding of indispensability seems unlikely.
It is submitted, however, that considering data as a homogenous commodity does not account for the wide variety of information different datasets contain and, therefore, excludes the possibility of a particular dataset being unduplicable due to the specific context and circumstances it was created in. Considering the uniqueness of a specific dataset, in fact, might be decisively relevant for assessing its indispensability under the essential facilities doctrine. The data used by companies to operate a search engine or an online search advertising service may only be obtainable through servicing customers directly and thus not be readily available
79 Case 7/97 Oscar Bronner [1998] ECLI:EU:C:1998:569, para 43.
80 Ibid 44
81 Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty (n 69) 81. 82 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 264-265.
for purchase from third parties.84 Coming back to the concept of data-induced network effects
and positive feedback loop, market entrants may lack the ability to generate the momentum necessary to be able to attract users and create their own datasets. An incumbent online service provider, as opposed to market entrants, enjoys the ability to improve its services proportionately to the number of trial-and-error tests it is able to run its search algorithm through. As explained in the previous chapter, a company like Google is able to leverage the large flow of data accessible to it to monitor demand for information and adjust its product accordingly. Under such market conditions, it is very difficult to imagine an entrant with access to only a fragment of the data available to the incumbent would be able to effectively compete on the quality of services offered.
If one considers online data to be a homogenous commodity, it is unlikely to pass the Oscar
Bronner indispensability test which requires there to be absolutely no other methods to conduct
the same business in the absence of the requested facility. The recognition that data collected by a search engine provider is unique and might not be easily substitutable by online data collected through different channels is paramount for assessing its indispensability correctly and with regard for all relevant factors. If the factor of uniqueness would be ignored, an argument could be made that the requesting entrant could develop its own search platform to obtain the necessary data, in analogy with the Court’s reasoning in Oscar Bronner. 85
On the contrary to Oscar Bronner, the Microsoft approach to the assessment of indispensability was developed in the context of a digital facility and thus might be better suited to consider data.86 In Microsoft, the Court of First Instance turned away from the strict requirement of no
alternatives to conduct the same business and held, instead, that it is necessary for competitors to be able to interoperate with Windows on an ‘equal footing’.87 In other words, the General
Court took the view that interoperability of information was indispensable because it was the only means for the requesting competitor to reach an equal position with the dominant undertaking.88 Application of the Microsoft indispensability standard to data of a dominant
general online search provider could establish indispensability, especially if combined with an argument for existence of entry barriers created by the limited access to data market entrants
84 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 265.
85 Aristeidis Demiroglou, ‘Essential Facilities Doctrine and Intellectual Property Rights: Approaches under the
Competition Law’ (LLM thesis, International Hellenic University 2016) 17.
86 Case T-201/04 Microsoft v Commission [2007] ECLI:EU:T:2007:289.
87 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 172. 88 Demiroglou (n 85) 35
are subjected to. Indeed, and as mentioned previously in this thesis, it has been argued that due to the existence of network effects on multi-sided platforms like online search platforms, users and advertisers gain greater value from operating on the platform with the largest existing userbase.89 Consequently, an entrant platform might experience great difficulty attracting
customers for both sides of its platform. This, in turn, will eradicate its ability to collect user data which is essential for development and upkeeping of a competitive level service. Form such perspective, it seems unreasonable to conclude that requested data of an incumbent search engine operator is readily available to competitors simply because some online data is might be available for purchase or collection.90 Even if other data would available for purchase from
another competitor or a data broker, the relevant question would be whether this data is of sufficient quality for the entrant to attract a minimum number of users required to reach the critical mass of its own data needed to successfully launch a new service that can survive on the market.
If the data of a dominant undertaking is necessary to provide a service of an equal or at least comparable quality to the one offered by it, and that data is not available to a competitor due to existence of network-effects and positive feedback loop on the side of the incumbent, data-induced entry barriers may exist.91 The presence of such market circumstances on the relevant
markets in the hypothetical scenario at hand could lead to a finding of indispensability of the requested data under the Microsoft indispensability standard.
It should be pointed out, however, that Microsoft did not appeal the judgment of the General Court and, consequently, the Court of Justice did not have a chance to express its opinion on the case and whether it endorses application of the lower standard for the indispensability test.92
It is, therefore, not clear whether the lowered threshold will be applied in future judgments the subject matter of which could be, amongst others, indispensability of data.
In addition to the jurisprudence of the ECJ, the European Commission’s guidelines on enforcement priorities in applying Article 102 provide a useful tool for assessing indispensability of data in the hypothetical scenario at hand. The Commission will typically
89 Graef, ‘Data as Essential Facility: Competition and Innovation on Online Platforms’ (n 10) 249. 90 Ibid 250.
91 Ibid 251 92 Ibid 173
