I. Introduction
The adoption of the Digital Markets Act (DMA or Regulation) follows a clear premise.Footnote 1 EU competition law is deemed ineffective to police anti-competitive conduct in digital markets.Footnote 2 The alleged failure derived from slow-paced antitrust sanctioning proceedings as well as the complex construction of theories of harm revolving around novel types of digital conduct justify its appearance in the EU’s regulatory framework.Footnote 3 According to such a narrative, competition authorities took too long to set forth their cases and, even when they managed to do so, intense litigation slowed the enforcement of those antitrust provisions. Even when litigation resulted in a victory for enforcers, remedies came short of working as the antiseptic means to restore the competitive process to its original undistorted state.Footnote 4 In sum, EU competition law failed to deliver effective enforcement with respect to the fast-paced, complex, and intricate nature of digital players and ecosystems.Footnote 5
The DMA’s effective enforcement stands as the means to address such an enforcement failure. The regulation mandates its regulatory targets (gatekeepers) to submit their compliance reports setting forth the technical implementations and changes to their business models they believe to satisfy the threshold of the DMA’s effective enforcement.Footnote 6 The European Commission (EC) faces a daring task in assessing those compliance reports, precisely because the DMA stands as a hybrid regulatory instrument.Footnote 7 Similarly to the application of antitrust rules, the EC holds the power to enforce the DMA’s provisions via its punitive powers. In fact, the EC opened six non-compliance proceedings to determine the regulation’s breach, suggesting that legal disputes with the gatekeepers in litigation may become commonplace.Footnote 8 These initial results question the call to ensure a quick and effective implementation of the new rules.Footnote 9
From the substantive perspective, the DMA’s obligations do not correspond with a single idea of enforcement. The EC has constantly repeated that the DMA is not output-oriented.Footnote 10 The regulation establishes market opportunities for business users that they must seize so to activate the DMA’s real potential. As a result, it is quite complex to determine where effective enforcement lies since impacts on consumers are not directly secured nor ensured via the DMA’s application.
Against this background, the paper responds to the question of where we should be looking at if we are to determine if the DMA’s enforcement is effective. One cannot simply assert that effective enforcement is achieved when a certain number of competitors populate the gatekeeper’s core platform services. Structural challenges may persist when several competitors are present in the same market. Therefore, the paper takes recourse to the notion of “metrics” to flesh out where does the threshold of effective enforcement lie.
The task is challenging from a theoretical perspective, since there are very few indications by the letter of the law pointing towards a particular direction. Both the EC and the gatekeepers have shied away from providing any indication of how success for the DMA should be defined. In fact, the EC directly transferred the whole responsibility of this task onto the gatekeepers by compelling them to detail within their compliance reports how each one of their technical implementations complied with the broader objectives of contestability and fairness and with each provision’s goal.Footnote 11 To this call, gatekeepers evaded any type of definite response.
In response to the loud silence, the paper uncovers the DMA’s secret metrics for success by engaging with the regulation’s main objectives, ie, contestability and fairness, and the implicit policy goals underlying each one of the substantive provisions contained under Articles 5, 6 and 7. Notably, the paper uncovers the following five main policy goals enshrined by the DMA: (i) market modelling; (ii) consumer choice; (iii) eliminating restrictions that favour the platform’s openness; (iv) neutralising competitive advantages; and (v) enhancing transparency. By combining both aspects of the DMA’s successful enforcement, the paper sets forth the European Commission’s enforcement capabilities. That is, the practical challenges it will face in applying its provisions, correlated to its policy goals. By doing so, the paper highlights the relevance of setting forth those policy goals, since only by uncovering one can analyse, in all honesty, whether the DMA may render effective and successful results in the market.
The paper is structured as follows. Section II explores the meaning of the legal interests protected by the DMA and provides a classification of the provisions accordingly. Section III outlines the DMA’s long-term policy goals by examining its provisions in relation to the EC’s enforcement capabilities. Section IV analyses Alphabet’s compliance with Article 5(4) DMA as a case study to demonstrate the correlation and challenges in the DMA’s effective enforcement. Section V concludes.
II. Metrics for defining the European Commission’s effective enforcement
Article 1(1) DMA sets out the Regulation’s main objectives in terms of ensuring contestable and fair digital markets. These goals come within the wider purpose by which the DMA aims at levelling the playing field of digital markets.Footnote 12 In light of the economic characteristics commonly associated with most digital platforms (eg, extreme scale economies, very strong network effects or a significant degree of dependence on both business users and end users),Footnote 13 the DMA reverses the antitrust error-cost framework by establishing per se obligations upon the designated targets.Footnote 14 Once a gatekeeper is designated under the Regulation, it must comply with the mandates set forth in Articles 5, 6, and 7 in a period of six months.Footnote 15
The shift from the application of antitrust rules to the regulatory approach is substantial. Instead of the burden of initial intervention and proof lying with the enforcer, the regulatory target must demonstrate how it integrates compliance solutions into its business model in line with the DMA’s objectives of contestability and fairness.Footnote 16 In this context, one would expect the DMA’s obligations and aims to be eminently easy to apply and pursue in practice. EC officials have constantly remarked on the fact that the DMA sets out clear rules. With this idea in mind, gatekeepers would know exactly what compliance should look like.Footnote 17
However, such a legal certainty is not a given. As opposed to the obligations under Article 5, termed self-executing obligations,Footnote 18 Articles 6 and 7 establish a set of obligations open to further specification, through the venue of regulatory dialogue illustrated in Article 8(2). Therefore, there is an implicit recognition that Articles 6 and 7 DMA at least are not so clear-cut regarding the technical transformations they require.Footnote 19
From the substantive perspective, the provisions apply irrespective of the underlying functioning of the gatekeeper’s business model. The DMA does not directly address the formula by which gatekeepers claim to create and capture value in digital markets. Hence, the same bar of compliance applies to all core platform services (CPSs) catered by gatekeepers, regardless of the extent to which they feed on strong network effects, data advantages, or the mechanics of market tipping.Footnote 20
The concepts of fairness and contestability act as touchstones to the DMA’s enforcement in more than one way.Footnote 21 First, they are the main objectives of the Regulation and influence all the EC’s actions in interpreting the DMA’s provisions. Accordingly, each one of the solutions presented by the gatekeepers must correspond to these objectives. Second, contestability and fairness permeate the regulatory mandates contained under Articles 5, 6, and 7 DMA. Therefore, they play a role in fleshing out the policy outcomes that each of the twenty-three mandates outlined in the aforementioned articles seek to achieve.
1. Contestability as potential competition
As set out under Article 1(1) and Recital 32 DMA, contestability is a means to an end. It is a means to eliminate barriers to entry and expansion undermining the ability of undertakings to contest, based on competition on the merits, the gatekeeper’s position in the market as well as to impact the innovation potential of the wider platform economy. This objective is clearly linked to the economic features of CPSs that imbalance competition and innovation.Footnote 22
Recital 32 states that the lack of contestability is not a matter of the number of competitors which populate the market. Instead, in order to promote inter-platform competition, the locus of attention is focused on the presence of barriers to entry or expansion hindering the exercise of competition at the platform level.Footnote 23 The Regulation does not define what economic characteristics of the digital platforms are to be understood as barriers to entry and expansion undermining inter-platform competition. The DMA sets out in the abstract the economic features that have led digital markets to tip the scales in favour of gatekeepers. But it does not indicate whether all of them are barriers to entry and expansion or even whether they may be reversed via the imposition of antitrust-like remedies. Economists have been discussing for decades now whether a particular economic characteristic is a barrier to entry and expansion.Footnote 24 Assuming that the DMA takes the widest conception possible of those barriers, it builds on the traditional definitions of economic and legal barriers set out by Bain in the 1950s.Footnote 25 Entry to a market is assumed unprofitable until information asymmetries and market imperfections can be compensated.
The DMA presumes that all these economic features apply equally to all CPSs and gatekeepers across the board. Economic reality trumps this one-size-fits-all approach. Some CPSs are not as contestable as others. For instance, switching costs are not as intense on web browsers as they are on online social networking services. In a similar vein, within the same category of CPSs it may well be the case that contesting a gatekeeper’s position may be easier than disputing another incumbent’s situation in the market. One such example is that of a number-independent interpersonal communications services (NIICS). The EC has only designated two NIICS belonging to the same gatekeeper (Meta’s WhatsApp and Messenger). However, the same economic characteristics (and, thus, the barriers to entry and expansion hindering the competitors’ capacity to contest their market position) do not apply to them with the same intensity and degree.Footnote 26 Even though both build on strong network effects, Meta’s WhatsApp service is much more prominent in its position as the by-default messaging service end users access in the EU. The same network effects apply to some extent to Messenger, but they only influence market outcomes in proportion to their link with the CPS’s integration with Facebook.
Contestability under the DMA is primarily adversarial. In other words, it can only be enhanced when one or several business users compete with the gatekeeper’s position, regardless of the value proposition each of them caters to end users. Therefore, it may be the case that the DMA may only facilitate the replacement of gatekeepers without addressing diversification. Following the terms of Recital 32, increased contestability could simply mean that business users have a credible opportunity to replicate the gatekeeper’s operations.Footnote 27
Against this backdrop, the paper translates the wider contestability objective into an identifiable metric. To this end, the paper identifies contestability with those provisions and scenarios promoting potential competition, understood as the springboard for existing and new sources of competition.Footnote 28 Indeed, by eliminating barriers to entry and expansion, the DMA strives to trigger the emergence of competitors in two fundamental manners. First, by enhancing the competitive situation of existing rivals, which have populated the market alongside gatekeepers in the past. Second, by creating favourable market conditions to foster market entries at the inter-platform level with the capacity to successfully compete against gatekeepers. Under Section II.C, the paper sets out those DMA’s provisions that (individually or in conjunction with the goal of fairness) pursue one of these manifestations of potential competition.Footnote 29
2. Fairness as conflicts of interest: value appropriation and conditions of access and competition
According to the DMA, gatekeepers impose unfair terms and conditions upon their business users to hinder them from fully capturing the benefits of their own contribution to the markets. The DMA’s depiction of fairness addresses these imbalances between the rights and obligations of business users where the gatekeepers obtain a disproportionate advantage (intra-platform competition).Footnote 30 Thus, the fairness objective is purely redistributive in nature.Footnote 31 Since gatekeepers have unfairly appropriated monopoly rents from the value that business users create on their platforms, the regulatory tool aims to correct this by redistributing them to business users.Footnote 32
The Regulation does not point out what type of redistribution should apply, namely whether surplus must be allocated depending on each business user’s value brought to the market or whether the appropriation of value should take place irrespective of their economic profits. The value of fairness under the DMA cannot be categorised into a single manifestation narrowing the gap between an unfair and a fair outcome.Footnote 33 More than a tractable metric for identifying effective enforcement, fairness is a value bearing an amorphous nature, depending on the individual consideration of the configuration of the market and each CPS’s economic characteristics.
Following upon the DMA’s silence, the paper translates fairness into a single definition applying across the board to identify unfair conduct in the eyes of the regulation, ie, the presence of conflicts of interest. The DMA’s scope of application predetermines this finding. Indeed, for an undertaking to qualify as a gatekeeper, it must provide a CPS which is an important gateway for business users to reach end users, aside from bearing a significant impact on the internal market and enjoying an entrenched and durable position in doing so.Footnote 34 Therefore, the presence of a potential conflict of interest is implied in the designation of the gatekeeper.
A clear comparison can be made with reference to those cases where the Court of Justice (CJEU) has recognised that the dominant undertaking’s conflict of interest predetermines market outcomes.Footnote 35 Following this same line of reasoning, the DMA embeds fairness in its foundational rationale by implementing this principle through its per se mandates. The DMA seeks to eliminate conflicts of interest concerning both the access conditions and competitive landscapes of CPSs. On the one side, the DMA removes all discriminatory conditions of access so that potential competition may crystallise. This is the reason behind the fact the Regulation highlights that contestability and fairness are intertwined, thus stating that a single provision may address both goals.Footnote 36 On the other side, the DMA revamps previously unfair distribution channels within ecosystems by compelling gatekeepers to adjust to digital business models that lack market dominance.Footnote 37 In other words, the DMA weaponises the goal of fairness to redistribute both value and control throughout gatekeeper ecosystems.
The redistribution of control across ecosystems entails co-responsibility on the part of business users. As opposed to the confrontational nature of contestability, fairness as outlined in the DMA is rooted in the Socratic paradigm, wherein each economic operator should be rewarded according to their efforts. Consequently, fairness is implied to apply to both the activities of gatekeepers (who may be rewarded for their role in shaping digital ecosystems) and business users.
3. The classification of the provisions according to their metrics
Stemming from the intertwinement of both metrics, each of the DMA’s provisions can be categorised into one of three categories illustrated in Table 1.
DMA’s provisions depending on their metrics
Several provisions aim at narrowing down the conflicts of interest embroidered into the conditions of access and/or competition to the gatekeeper’s business models in isolation. This is only logical. The Regulation is mainly based on the presence of a gatekeeping power and on the related risks created by the leveraging of their advantages from one area of their activity to another. Therefore, such provisions pursue the disintermediation of the designated gatekeepers from their prominent positions within the markets they operate in. For instance, the DMA discontinues the gatekeeper’s capacity to leverage non-public data generated by its business users when it competes with them, under Article 6(2).
A few obligations seek to restore potential competition as a standalone. These provisions report a quasi-automatic increase in the number of competitors entering the market as a result of their implementation. As means of example, Article 6(4) DMA cements potential entries at the downstream and upstream level regarding app distribution by compelling gatekeepers to allow and technically enable third-party app stores and apps to interoperate with their operating systems.
Finally, most provisions have a hybrid nature since they impact on both existing and potential competition whilst eliminating conflicts of interest.
After all, the Regulation has not been designed as a closed system of protection reporting clear legal interests per each provision. Although one could identify certain provisions to confer the power to business users to narrow the gatekeeper’s conflicts of interest to a minimum, the absence of these conflicts of interest will also entail, in the medium to long-term, a higher propensity and incentive for entrants to compete on the merits with the gatekeeper, thus fostering inter-platform (rather than just intra-platform) competition.
III. The matrix in practice
The DMA compels the EC to effectively enforce its provisions. From the first round of compliance reports issued by the seven gatekeepers designated in September 2023 and April–May 2024, the task looks moving-target-like and complex.
As a first step, it is worth noting that there are three types of DMA mandates depending on the scope of their application. The twenty-three provisions do not necessarily apply to all CPS categories. As illustrated in Table 2, most provisions only apply to specific CPSs, while several provisions are even wider in scope than the CPS categories.
Scope of DMA’s provisions
Irrespective of their scope, the provisions do not apply in a vacuum. The solutions put forward by the gatekeepers in their compliance reports do not exclusively impact the services strictly falling under the DMA’s scope of application. In most cases, the provisions target secondary or complementary services belonging to business users. Therefore, the gatekeepers’ adjustments to their business models have a clear external vocation, rather than representing a mere internal restructuring.
When descended into reality, the legal interests of contestability and fairness can be translated into a myriad of policy outcomes with two different origins.
On the one side, the DMA seeks to restore the competitive conditions that the antitrust framework allegedly failed to achieve in digital markets. By doing that, the DMA aims at achieving solutions competition law is deemed unable to ensure. This is particularly salient if one looks at the self-preferencing prohibition under Article 6(5) DMA. Inspired by the EC’s investigation in Google Shopping, the DMA imposes an outright ban deriving from a novel theory of harm.Footnote 38 A similar example can be drawn out from Article 5(2) DMA, which has been inspired by the German Facebook case.Footnote 39
On the other side, the DMA pursues independent and complementary policy goals of its own. To the extent that the regulatory instrument remains distinct from competition law, considerations of public policy can be integrated without the need to address how they intersect and overlap with the need to increase consumer welfare. Consequently, the DMA eliminates the concept of consumer harm, while also incorporating consumer protection-focused policy decisions into the regulatory framework. For instance, the gatekeepers reporting extend to the gatekeepers’ requirement to submit audited reports detailing the consumer profiling techniques they employ within their CPSs under Article 15. The primary aim of this obligation is to promote transparency and provide a clear understanding on the gatekeepers’ profiling activities.
In the next sub-sections, the paper disentangles the different policy goals pursued by each of the mandates contained in Articles 5, 6 and 7 DMA.
A two-step process must, therefore, necessarily apply to the DMA, since all its provisions do not necessarily target the same policy goal, nor do they protect the same legal interests. When confronted with the question of whether the gatekeepers’ compliance solutions demonstrate effective enforcement, the first step requires the enforcer to establish what legal interest is pursued. Section II.C already conducted this exercise, as shown in Table 1.
The second phase of the analysis translates these legal interests into long-term expectations. Declared contestability and fairness take various forms when applied into practice. As opposed to the legal interests, policy goals remain obscured by a myriad of policy choices embedded in the Regulation. Section III.A uncovers the policy goals behind each provision and Section III.B outlines the enforcement capabilities the EC will need to address in relation to these goals.
1. The DMA’s policy goals
The DMA’s legal interests outlined under Section II co-exist with additional policy goals pursued by each of its provisions. From the perspective of regulatory theory, the DMA brings together aspects of both a goals-based and a rules-based regulation.Footnote 40 It shifts the responsibility of intervention onto the subjects of the Regulation, compelling them to evaluate the most effective way to adhere to its goals of contestability and fairness. At the same time, the obligations under Articles 5, 6 and 7 DMA provide for specific prescriptions and proscriptions of conduct. Rather than investigating whether the gatekeeper’s compliance is consistent with the goals set out by the Regulation, the EC analyses whether they comply with the substantive provisions.
Building upon this idea, the policy goals can be plotted across a spectrum of the desired regulatory transformations mandated by the DMA, including market modelling, openness, neutralisation of competitive advantages and transparency. Under Table 3, the paper sets out the provisions included under each one of those points on the axis.
DMA’s provisions according to their metrics and policy goals
As depicted in Table 3, on one side of the spectrum the DMA functions as a market modelling tool. The EU legislature targets at setting out a particularised view of the economic interactions taking place by the market forces. The regulatory design of some of the provisions carve out a particular view of the architecture and design of digital ecosystems. As the outcome is apparently predetermined, the regulatory intervention prevents gatekeepers from selecting one option from a variety of potential solutions that could meet the compliance standard. For example, Article 5(2) prohibits data combinations across core platform services. Thus, business models based on behavioural advertising are seen as presumptively undesirable.Footnote 41 EC representatives have even highlighted that the DMA requires gatekeepers to offer users a less personalised alternative to its services, which may consist of contextual advertising.Footnote 42 Table 3 includes all those provisions under the DMA seeking to impose a particular regulatory design upon the gatekeepers.
Moving along the spectrum to more alleviated forms of intervention, a wide range of provisions build upon the DMA’s preference for the openness of digital ecosystems.Footnote 43 The policy goal introduces accessibility to the means of competition where the gatekeeper is the stronghold to the catering of the service. Accordingly, on the one hand, some of the mandates of the provisions aspire to open markets to trigger more consumer choice. For instance, Article 6(4) allows alternative operators to distribute their app stores and apps on the gatekeeper’s ecosystems without relying on proprietary technologies. On the other hand, the paradigm of openness entails that business users should not be hindered from competing in gatekeeper environments. Due to this reason, several provisions unfasten the restrictions gatekeepers may impose upon the entry of certain functionality on their CPSs. As means of an example, Article 5(4) prohibits the gatekeeper’s anti-steering restrictions.
Moreover, some of the provisions pursue the detachment of those competitive advantages which, in principle, the gatekeeper enjoys due to its condition as a prominent player in digital markets. Such advantages are not identified as choke points to be eliminated by the least restrictive measures imposed by the Regulation. On the contrary, these provisions aim at neutralising the gatekeeper’s position within their CPSs. The clearest example is enshrined in Article 6(2) DMA, which neutralises the gatekeeper’s capacity to leverage business user data generated on its CPSs to compete with them. All the gatekeeper’s CPSs are comprised under the obligation as a matter of scope and every type of competition between the gatekeeper and the business user remains captured. In parallel, Article 6(5) introduces the self-preferencing prohibition hindering the gatekeeper’s capacity to treat more favourably, in ranking, related indexing and crawling, its own services and products vis-à-vis those of its business users. Under the assumption the provision is not sufficiently wide, Article 6(5) adds on that the gatekeeper shall apply fair, transparent, and non-discriminatory conditions to such ranking.
Finally, the least intense form of regulatory intervention is that of enhancing transparency, especially in online advertising services. As acknowledged in Recital 45, the choice is based on the failure of EU data protection regulation.Footnote 44 The conditions under which gatekeepers provide online advertising services to business users are considered opaque and non-transparent because of the gatekeeper’s practices and the complexity of modern-day programmatic advertising. Thus, the legislator imprints the need to restore transparency in favour of business users in online advertising services via Articles 5(9) and 5(10) to resolve a failure which is not necessarily indicative of the presence of gatekeeping power. Alternatively, the transparency policy goal permeates consumer-protection objectives into the competition policy-based regulation by different means by, for instance, approximating the DMA’s content to the spirit of the P2B Regulation.Footnote 45 Article 5(6) is a good proxy for illustrating this point. Indeed, Article 5(6) does not shape gatekeeper decision-making into any given direction. It recognises a right to business and end users so they can exercise it before public authorities without any type of impediment deriving from the target’s gatekeeping power.
From a quantitative perspective, such an analysis demonstrates that the DMA’s policy goals are not embodied in the least intrusive means of the neutralisation of competitive advantages nor of the enhancement of transparency. On the contrary, most of the DMA’s provisions target a particular view of how digital markets should look like from a policy perspective. Therefore, the Regulation cannot be said to be agnostic regarding the business model transformation it imposes upon the gatekeepers. It points towards a clear preference for choosing market outcomes via the market modelling provisions under Articles 5(2), 5(3), 6(11), and 7 DMA. In parallel, it mandates ecosystem openness, irrespective of the gatekeeper’s configuration of its business model and decision-making processes, prior to the full application of the obligations.
Alternatively, from a qualitative viewpoint, Table 3 confirms a finding already pointed out in Table 1. Most obligations are not directly fine-tuned to strengthen potential competition in the market. In fact, there is a clear preponderance of hybrid mandates alongside those provisions aimed at narrowing down the gatekeeper’s conflicts of interest in different forms.
2. Enforcement capabilities
Stemming from the theoretical background of the DMA’s legal interests and policy goals, the EC’s effective enforcement can only be measured if confronted with practical reality. The seven designated gatekeepers already applied the Regulation’s obligations across their business models and submitted their compliance reports in March and October–November 2024. EC officials have already termed some of the solutions presented by the gatekeepers as blatant infringements of the spirit and letter of the Regulation.Footnote 46
Two main cornerstones to the DMA’s enforcement loom over the EC’s capability in transforming the Regulation’s mandates into reality, namely, information asymmetries and the provisions’ interplay with other pieces of the EU law and ongoing antitrust proceedings.
The DMA addresses the problems of the lack of information at the EC’s disposal to engage with the dynamics of digital platforms by shifting the responsibility onto gatekeepers. They are required to submit compliance reports and the auditing of their consumer profiling techniques.Footnote 47 However, reviewing the submitted compliance reports, such a solution does not necessarily appear to be decisive in terms of narrowing information asymmetries vis-à-vis the European Commission.
For instance, the prohibition under Article 5(2) is supposed to be self-executing and, therefore, applied by default. From the technical perspective, this shift entails the CPS data infrastructures’ transformation into siloed datasets which cannot interact with each other, absent the end user’s consent. In practice, Alphabet applied this approach by restricting data flows of personal data across its eight core platform services vis-à-vis the siloing of its non-CPS designated services.Footnote 48 The same enforcement strategy has not been followed by other gatekeepers, such as ByteDance, which refused to put forward substantial technical solutions. Given that TikTok’s advertising services are an integral part of the TikTok entertainment platform, no combinations of personal data apply from different services when it creates user profiles for personalised advertising. Therefore, it asserted that the current configuration of its data infrastructure already complies with the obligation under Article 5(2).Footnote 49
If the EC may gain some knowledge of the gatekeeper’s data processing activities via the obligation under Article 15 DMA, without the target’s active engagement in unveiling its own data infrastructure, a profound interpretation of the compliance report cannot be performed.Footnote 50 At a first glance, when monitoring the provision’s enforcement, the EC can only trust the gatekeeper’s assessment. It is not completely unsurprising that those provisions directed at modelling the market in a particular direction are the most impacted by this challenge. Bearing in mind they devise an idea of the expected market outcome, those results may conflict and oppose the policy goals set out by other pieces of EU regulation, providing sufficient grounds for tension in the EC’s enforcement capabilities, especially in light of its obligations deriving from the principle of sincere cooperation under Article 4(3) TEU.
Furthermore, as the DMA is a piece of legislation within the wider corpus of EU law, the EC’s enforcement capabilities cannot be measured in a vacuum. Despite the Regulation’s assertion that its application takes place without prejudice to any other piece of EU regulations and to the application of competition law, the distinction of where the EC’s enforcement starts and ends is not straightforward.Footnote 51 The EC may have to look outwards to enforce the DMA’s mandates. For instance, when sharing search data generated on online search engines under Article 6(11), the gatekeeper’s anonymisation task cannot be interpreted without reference to the understanding of anonymisation under EU data protection regulation. Due to this reason, the DMA provides for several fora of discussion with data protection supervisory authorities to substantively engage on these points of law.Footnote 52 In parallel, some of the gatekeepers’ compliance solutions mimic the remedies already offered within antitrust proceedings in advance of the DMA’s adoption. This is the case of both Meta and Amazon with regard to the technical implementations of the obligations under Articles 6(2) and 6(5). Their solutions did not go any further in scope or substance than what they had already convened with the European Commission in their Facebook Marketplace and Amazon Buy Box and Marketplace proceedings.Footnote 53
Aside from both characteristics, the provisions following distinct policy goals bear different challenges regarding the EC’s enforcement capabilities.
Circling back to those provisions pursuing the neutralisation of competitive advantages, for instance, the EC is at a clear crossroads. Information asymmetries persist with the gatekeepers, creating issues not only with their overall incentive for disclosing data but also with the specific challenges arising from the implementation of the DMA’s provisions. If gatekeepers were to disclose their compliance strategies to reduce information imbalances when submitting their reports, assessing compliance with such provisions would still be entirely unclear. Because the obligations are mainly negative in nature and require an abstract demonstration of neutrality, gatekeepers can effectively obscure the decision-making process, making it more difficult for the EC to monitor enforcement.
As a matter of example, Amazon’s compliance with the data siloing obligation under Article 6(2) DMA demonstrates this point. In the illustration of its compliance solution, the gatekeeper did not substantially engage with the process of decision-making underlying its operations. Instead, it referenced, in the abstract, its efforts to silo data through the different technical systems it has in place. By this token, Amazon remarked the presence of various automated systems, algorithms, models, and tools that feed into the decisions where it may be perceived to act in competition with third-party sellers via selection, inventory, and pricing decisions. In summary, Amazon declared that, upon a review of the data inputs drawn from each of its automated systems, it could confirm that none of them ingest or use non-public third-party seller data. No further reference nor explanation was provided by the gatekeeper. Despite the reversal of the burden of intervention, the gatekeeper did not present any evidence to the effect of demonstrating compliance with the negative obligation. The EC is, thus, expected to trust the gatekeeper’s word and explanations at face value.
Moving to those provisions targeting the openness of platforms and digital ecosystems, the compliance reports show the gatekeepers’ clear predilection to avoid relaying power out of their hands. Gatekeepers can no longer completely block access to their platforms as they could before the DMA came into effect. However, they have found ways to delay business users’ requests for access by implementing entitlement procedures.
Apple’s enforcement strategy is quite salient in this respect, as it was arguably the digital ecosystem with the smallest degree of openness prior to the DMA’s application. Pursuant to Article 6(4) DMA, the gatekeeper is required to allow alternative operators of app stores and apps to distribute these services via different means to its proprietary App Store. However, Apple has not relinquished all its decision-making capacity when it comes to app distribution on iOS.Footnote 54 Alternative app store operators must go through an entitlement process, managed and designed by the gatekeeper, to receive authorisation to operate on iOS. Apple has not included the entitlement requirements into its first compliance report, but it updated and uploaded information on its developer’s webpage the conditions these alternative operators will have to comply to exert the opportunities for openness provided by the DMA.Footnote 55 Such rules include the need to follow the notarization process, enabling Apple to filter through requests for access focusing on reasons of security, privacy and the maintenance of device integrity. Without a positive answer from Apple on notarization, the business user will not be able to benefit from Article 6(4).Footnote 56 On top of this process, Apple establishes additional requirements to authorise the business user’s operations as an alternative app store, such as submitting a new binary for the app store’s sole distribution on iOS in the EU or providing Apple with a standby letter of credit from an A-rated financial institution in the amount of EUR 1.000.000.
Similar challenges involve the evaluation of the effective application of provisions pursuing market modelling and openness. While the DMA aims at introducing opportunities business users must grasp to thrive in the CPSs’ markets, some compliance reports show that the venues for the enhancement of inter and intra-platform competition are a matter of business users’ discretion. Notably, some gatekeepers present their business users with a binary option. They may either stick with the choices and conditions they had before the DMA came into effect or choose to adopt the compliance solutions proposed under the entitlement conditions set by the gatekeeper. This is precisely the model Apple presented to its business users. However, in this context, compliance with the regulation does not apply by default.Footnote 57
Considering the array of challenges the EC will encounter while monitoring DMA’s enforcement, it is clear that a complex network of obstacles lies ahead. These hurdles are not straightforward in theory and vary in practice, depending on the gatekeepers and their enforcement strategies outlined in the compliance reports.
IV. Putting theory into practice: a case study of Alphabet’s proposed compliance with the anti-steering prohibition
Evaluating effective enforcement is a significant challenge. As illustrated in the previous Sections, given the DMA’s broad spectrum relating to its legal interests and policy goals, enforcement is anticipated to be multi-faceted, nuanced, and essentially unpredictable in terms of the outcomes expected from the Regulation’s mandates.
Therefore, the paper proposes a practical method to assess whether a specific compliance solution aligns with the concept of effective enforcement or necessitates further scrutiny under the measures outlined in the DMA. It fundamentally reveals the obscured aspects of the DMA’s enforcement strategy, aiming to balance its recognized legal interests with the often-overlooked long-term policy goals its provisions express and aim to achieve. In doing so, the paper illustrates a dual-layered enforcement approach that must not only uphold the DMA’s broader objectives but also align with the policy direction embedded by the EU legislator. This dichotomy within the DMA’s framework is crucial for anticipating the EC’s enforcement capabilities and determining how to address them effectively.
To examine the necessity of considering the secretive nature of long-term policy goals, the paper employs a case study of the non-compliance procedure initiated by the EC against Alphabet. The procedure addresses potential infringement of Article 5(4) DMA related to the Google Play intermediation service.
According to Alphabet’s compliance report, it now provides additional means for developers to communicate and promote offers to end users as well as by providing them with the capacity to directly conclude contracts via these means. Alphabet has opened this possibility for app developers via its new External Offers program. Developers must agree with the terms and conditions of that program to enjoy the possibility of steering their users to promotional offers. Aside from that, participation in the program by the app developer is subject to Google’s approval. For instance, some of the eligibility criteria fleshed out by Alphabet include the fact that developers must only direct end users to their own digital features or that they are directly responsible for providing support to users in their external transactions. In this respect, the most salient aspect of the compliance solution relates to the fee structure Alphabet imposes upon developers when they redirect their end users to promotional offers. Under the assumption the app developer acquired those end users due to its presence on Alphabet’s Google Play, the gatekeeper charges a 5 per cent initial acquisition fee on the offers catered through the in-app digital features and services during the two years following the initial external transaction on top of an ongoing services fee of up to 7 per cent for those same offers. Only after these two years stemming from the end user’s initial acquisition, the developer may opt out from receiving Play services and paying the ongoing services fee.
The first step of the analysis necessarily stems from its categorisation in line with the legal interest it purports. Article 5(4) DMA seeks to promote intra-platform competition by narrowing down gatekeeping power in setting rules driving downstream competitors away from the digital ecosystem. The rationale underlying the anti-steering provision relates to the fact that a gatekeeper will not earn revenues from those services its downstream competitors realise outside its digital ecosystem.
Article 5(4) pursues the legal interest of reducing conflicts of interest as a standalone metric. Upstream competition among platforms is not directly concerned and, as such, the provision does not fall within the definition of a hybrid obligation as depicted under Section II.C. Establishing the provision’s legal interest sets the path for the anticipated results that one can intuitively expect the provision to achieve. In the case at stake, the legislator may intend to boost traffic from downstream services offered through digital ecosystems to their own websites by separating the completion of a transaction from the operator’s presence in a specific digital ecosystem. Therefore, an increase in the number of app developers offering such functionalities and the prevalence of more promotional offers providing those links would serve as initial indicators of effective compliance with the provision.Footnote 58
Once the legal interest has been identified, one must set out the provision’s scope in relation to the CPSs included to compare it with the scope of the compliance solutions proposed by the gatekeeper. According to the terms of Article 5(4), the provision is wider in scope to the CPS categories. It compels the gatekeeper to allow business users to communicate and promote offers to end users acquired via its core platform service or through other channels, and to conclude contracts with those end users. Looking at Alphabet’s proposed solution, the initial acquisition and ongoing services fees imposed on developers as a take-it-or-leave-it choice does not seem to align with the provision’s scope. Indeed, on the one side, the anti-steering provision spans through all types of communications and channels. Thus, the absence of a compliance solution for the rest of Alphabet’s CPSs and channels is not coherent with the obligation. On the other side, Article 5(4) does not solely focus on end users benefiting from a CPS, but on all categories of end users who depend on the gatekeeper’s ecosystem. In this context, it becomes possible to analyse the terms introduced by Alphabet regarding its fee structure when end users are directed to promotional offers from in-app services provided by app developers.
Continuing with the analysis, it is critical to determine the provision’s nature in the context of the broader policy goals pursued by the EU legislator. Article 5(4) aims at enhancing the openness of the ecosystem at the downstream level by removing previous restrictions on directing end users towards promotional offers. To this end, Article 5(4) includes a clear and active mandate to eliminate anti-steering restrictions from the terms and conditions imposed by gatekeepers on their counterparts. This enables transactions to take place outside of the gatekeepers’ ecosystems. Both the legal interest and policy goal remain, therefore, closely correlated in terms of their identification, even though one does not necessarily automatically follow from the other.
The EC’s enforcement capabilities closely align with both the legal interest and policy goal of the provision. As the provision aims to promote openness in downstream markets, the primary enforcement challenges arise from Alphabet’s entitlement processes. Indeed, Alphabet’s proposed compliance solution does not prescribe openness by default. Instead, the steering of users will only occur with Alphabet’s prior authorization of developers’ entitlements to provide link-outs within its services.
Finally, the potential side effects stemming from the provision’s compliance must be taken in mind. In the case of Article 5(4), there is an undeniable interplay with the separate obligation imposed by Article 6(12), according to which the gatekeeper must apply fair, reasonable, and non-discriminatory conditions of access for business users to its software application stores, online search engines, and online social networking services listed as CPSs. Therefore, within the context of Article 5(4), the EC must assess how fair conditions are set through developers’ entitlements, which should be included in the broader analysis of compliance with Article 6(12) DMA. However, what is deemed fair under Article 5(4), aimed at reducing conflicts of interest, might not be regarded similarly under a distinct provision.
In summary, the case study of Alphabet’s compliance with Article 5(4) DMA yields mixed results but reveals several clear findings. At first glance, it is relatively easy to discern whether the provision serves one legal interest over another. If the provision addresses issues within the downstream market concerning the CPS it operates in, then conflicts of interest are at play. Therefore, Alphabet’s solution should aim to rectify any imbalances in bargaining power imposed on business users in comparison to Alphabet’s own position. However, the new fee structure and the requirement for app developers to pay an ongoing service fee for Google Play for two years contradict this spirit.
Understanding the provision’s policy goal is not crucial for assessing compliance with it at face value, but it is pivotal for two fundamental aspects of the analysis. First, it defines the long-term objective the provision aims to achieve. If compliance solutions meet the legal interest (ie, fairness) but fall short of achieving the desired degree of openness, there will still be room for the enforcer to encourage the gatekeeper to implement a modified version of its compliance solution. The enforcer will be able to display its punitive and non-punitive powers.
Secondly, elucidating a provision’s policy goal is crucial for highlighting potential challenges the EC may face in its enforcement efforts. Regarding Article 5(4) DMA, which aims to instil openness, the most significant challenge associated with fulfilling this hidden promise is intermediation. The DMA is unlikely to completely eliminate gatekeeping power when a dominant player still benefits from operating within an ecosystem. This correlation becomes apparent when considering Alphabet’s entitlement process to offer its in-app link outs to business users. Without first establishing the long-term policy goal, it would be difficult to determine whether the entitlement process should be evaluated based on fairness or if enforcement challenges arise from elsewhere.
Indeed, this is precisely what the paper has untangled for each provision, paving the way for the EC’s effective enforcement. It is not entirely clear what legal foundations will influence the EC’s enforcement strategy, apart from the broader concepts of contestability and fairness. The paper has shown that an additional layer, substantiating the DMA’s direction into specific policy goals, is crucial for understanding where the Regulation’s success lies. This also guides gatekeepers and enforcers in narrowing down their choices in terms of ecosystem design, architecture, and enforcement approach. To hold both gatekeepers and the EC to account, the paper illustrates the common ground they can reach by acknowledging the DMA’s policy goals. As a consequence, the EC’s enforcement actions should be constrained to these specific goals.
V. Concluding remarks
The adoption of the DMA has been accompanied by significant doubts and criticism, particularly due to its controversial relationship with competition law. Inspired by antitrust investigations, the new Regulation finds its roots and essential rationale in an alleged antitrust enforcement failure. Therefore, once adopted and applied, the main question is how to determine the DMA’s success. Namely, how to identify the conditions under which the implementation of the new rules has been effective for each obligation by achieving results that competition law would not be able to ensure.
This paper addresses this question by providing the legal interests and policy goals to measure the DMA’s success. Notably, it shows that such a process requires unveiling the DMA’s hidden policy goals. Indeed, although contestability and fairness are the proclaimed protected legal interests, they do not represent the outcomes the EU legislator aimed to embed in the Regulation. Despite the EC’s desire to keep these aims somewhat hidden, four main policy goals influence the DMA’s provisions: market modelling, openness, neutralizing competitive advantages, and enhancing transparency. The EC’s enforcement capabilities are directly correlated with the long-term policy goals pursued by each provision.
Translating fairness and contestability into clear policy goals and thus unveiling the secret of the DMA provides two significant benefits. It guides and constrains gatekeepers in developing solutions to comply with the new rules. At the same time, it holds the EC accountable for its enforcement strategy, ensuring its actions are directed toward achieving specific market outcomes.
Open access
