A Game of Hide-and-Seek: The European Commission Does Not Designate Microsoft Ads, Bing and Edge and Apple’s iMessage as DMA Gatekeepers
The DMA seeks to capture gatekeeper conduct. For that, the regulation applies to those targets of the regulation satisfying the legal category of a gatekeeper. If an undertaking is not a gatekeeper as per a designation decision issued by the European Commission (EC), then it will not remain captured by the regulatory instrument.
In September 2023, the EC designated six gatekeepers with respect to twenty-two core platform services (CPSs) and left the fate of four CPSs belonging to Apple and Microsoft in the air. The designated gatekeepers and their corresponding CPSs all directly followed the quantitative thresholds set out under Article 3(2) DMA and did not reach the high standard of the rebuttal of the presumption of Article 3(5) DMA. Instead, these four CPSs presented sufficiently substantiated arguments demonstrating they did not satisfy the requirements under Article 3(1) DMA.
Following the rebuttal of the presumption, the European Commission triggered four market investigations under the terms presented under Article 17(3) DMA to determine whether they should fall outside of the scope of designation. In February 2024, the EC closed the market investigations and found that Apple and Microsoft should not be designated as gatekeepers for those CPSs.
The contradiction of the terms under Article 3(5) and Recital 23
Most of the European Commission’s designation procedures under the DMA followed the quantitative path. This was the least resource-intensive and fastest way to capture the targets of the regulation. Thus, it was only reasonable to expect that most of its designation decisions touched upon the quantitative thresholds set out under Article 3(2) DMA. However, it was not entirely clear how it would interpret the means the DMA puts forward to avoid capturing false positives – via the rebuttal of the presumption of Article 3(5) DMA – and false negatives – through the qualitative designation enshrined in Article 3(8) DMA -.
As far as the rebuttal of the quantitative presumption is concerned, the terms of Article 3(5) are not completely aligned with the interpretative provision under Recital 23. Article 3(5) establishes the gatekeeper may demonstrate sufficiently substantiated arguments to prove that it, exceptionally, does not satisfy the requirements listed in Article 3(1) due to the circumstances the relevant core platform service operates in. Reading the provision in isolation, any type of argument, of a qualitative or a quantitative nature, could demonstrate to be sufficiently substantiated to surpass the first hurdle of rebutting the quantitative presumption.
However, the European Commission’s designation decisions did not prove to be so welcoming of any type of argument. On the contrary, in each instance when a gatekeeper tried to put forward an argument slightly related to a qualitative aspect of its business model or functioning, the European Commission outright rejected it. For instance, the EC dismissed most of ByteDance’s arguments advocating against TikTok’s designation under the DMA, given that they fell outside the scope of the designation exercise altogether (para 142 of ByteDance’s designation decision). For example, ByteDance argued it is a challenger to the rest of digital platforms (para 111). As such, it should not have been designated by the European Commission.
By adopting this stringent approach, the EC followed the terms presented by the faintly contradictory terms of Recital 23. Under its terms, the EC “should take into account only those elements which directly relate to the quantitative criteria”. In principle, thus, Recital 23 opposes the rationale underlying Article 3(5) on its own terms. In my own mind, the line of reasoning presented under Recital 23 makes sense if the designation is performed under the quantitative means of Article 3(2) DMA. If the EC designates a gatekeeper based on quantitative criteria, then it is only reasonable to expect the rebuttal of that same presumption to contextualise those same criteria into the bigger context. Recital 23 provides some examples of the circumstances to justify the rebuttal, such as the impact of the undertaking providing CPSs on the internal market beyond its revenue or market capitalisation, its size in absolute terms or the undertaking’s CPS considering the overall scale of its activities.
The European Commission’s decisional practice surrounding its application on the terms of the rebuttal under Article 3(5) vis-à-vis those presented under Recital 23 is, at most, irregular. Despite the EC outright rejecting most qualitative arguments presented by gatekeepers seeking to rebut the quantitative presumption, it accepted them in two cases. First, it accepted the rebuttals Alphabet and Microsoft presented for their messaging services Gmail and Outlook.com based on the fact the gatekeepers did not appear to exert any control over the operations of their end users or business users to enable them to make those users dependent on their services (para 146 of Alphabet’s designation decision and para 129 of Microsoft’s designation decision). The argument stems from the interpretation of Recital 20, which is an interpretative provision unrelated to the rebuttal process and describes the reasons underlying the quantitative presumption. In a similar vein, the EC has, lately, accepted X Ads and TikTok Ads’ rebuttal of the presumption directly at the stage of designation without taking recourse to the mechanism of Article 17(3) DMA, although we do not yet know what arguments sustained their non-designation yet. Second, it accepted Samsung’s rebuttal of the presumption accepting the gatekeeper’s argument that its Samsung Internet Browser is not characterised by strong lock-in effects, and it does not have the ability to leverage its position in web browsers to the advantage of its other digital services (para 45 of Samsung’s designation decision). Let’s remember that the European Commission rejected those same arguments in ByteDance’s rebuttal (paras 110 and 111 of ByteDance’s designation decision).
In fact, several judges of the General Court are currently confronting this same contradiction in trying to disentangle the terms of both provisions for resolving ByteDance’s appeal of its designation decision (for comment on the appeals and the interim measures of the case, see here). Last 29 April, the General Court held its first hearing on any issue related to the DMA and the judges enquired both ByteDance and the EC on whether and how both provisions may be reconciled.
Enter the scene: ecosystems!
In waiting for the GC’s response, the European Commission has given a decisive step towards the wider notion of Article 3(5) as opposed to Recital 23 by deciding to not designate Apple’s iMessage as well as Microsoft’s Bing, Edge, and Advertising. The non-designation decisions illustrate the fact the rebuttal of the presumption is, in reality, a two-step process with distinct thresholds of intervention required from the gatekeeper at each of those two stages.
Whilst the first prong of the process requires the gatekeeper to present qualified arguments to justify the prima facie exemption from the presumption, the second stage expands the terms of the narrative in line with Article 3(5) by enabling gatekeepers to pose wider questions of law to the EC for scaping the designation. The non-designation decisions demonstrate this point with reference to the notion of digital ecosystems.
Throughout the different designation decisions issued in September 2023, the European Commission was adamant in defending the notion that digital ecosystems do not play a role in the designation nor in any substantial way in the context of the DMA. In fact, the EC was not wrong in bringing forward such a point. As much as one reads and re-reads through the DMA’s provisions, the notion of ecosystems is referenced thrice in relation to: i) their structurally extremely difficult conditions to challenge or contest by existing or new market operators (Recital 3); ii) their importance in the wider digital environment of the gatekeeper’s operations (Recital 32); and iii) as the sustaining grounds for the gatekeeper’s provisions of their number-independent interpersonal communications services (Recital 64).
Throughout Article 3, however, the concept of an ecosystem remains unnoticed. In fact, the European Commission recognised “nothing (in the DMA) suggests that having an ecosystem is an absolute pre-requisite to be an important gateway for business users to reach end users”. The existence of an ecosystem “contribute(s) to explaining the reasons why the legislators laid down the conditions under which certain CPSs constitute an important gateway for business users to reach end users” and in, fact, the reference under Recital 3 of the DMA states that only some undertakings exercise control over ecosystems, “which makes clear that not all gatekeepers have such control over an entire ecosystem of services in the Union” (para 130 of ByteDance’s designation decision). Thus, the undertaking must not necessarily have an ecosystem to be designated as a gatekeeper. In all fairness, throughout its designation decisions, the EC kept silent on whether the absence of an ecosystem (or of its structural conditions) could, in turn, sustain a CPS’s exclusion from the DMA’s scope.
The EC’s non-designation decisions responded in the positive. Microsoft advocated for Bing’s, Edge’s, and Ads’ non-designation by arguing that the gatekeeper’s platform ecosystem does not currently sufficiently contribute for any one of them to make them important gateways for business users to reach end users. For instance, in the decision accepting the three rebuttals to the presumption when discussing Bing’s designation, Microsoft argued that despite the gatekeeper offers a comprehensive set of digital products and services, some of which are offered in an integrated way, forming an ecosystem of products and services, Bing does not benefit from any significant advantage due to its vertical integration. Furthermore, Edge and Microsoft Ads do not reinforce Bing’s importance because they are not important gateways for business users to reach end users, either (paras 36 to 38 of the decision). In a similar vein, the EC accepted Microsoft’s line of reasoning sustaining Edge should not be designated because, despite the gatekeeper’s efforts in pre-installing the web browser on Windows and incentivising OEMs to set Edge as a default, it did not sufficiently contribute to Edge being an important gateway in the sense of Article 3(1). In turn, the EC terms Microsoft’s ability to use Edge as a lever to drive usage of other Microsoft products as limited and insufficient (paras 63 to 66 of the decision). Finally, one of the main arguments sustaining Microsoft Advertising’s non-designation also revolved around the fact that Microsoft’s platform ecosystem does not currently result in sufficient scale effects benefiting the CPS. In spite of the comprehensive set of digital products and services within the ecosystem, the data-driven advantages and network effects did not entail Microsoft Ads acquired a significant scale of use across the metrics of business or end users (paras 88 and 96 of the decision).
From the above, one can easily derive the EC’s clear predilection to expanding the terms a gatekeeper may present once it has surpassed the high threshold of the ‘sufficiently substantiated arguments’ required under Article 3(5) DMA. Qualitative considerations not even mentioned in the letter of the law are appraised to avoid false negatives making it to the designation. The difference with the threshold of intervention established under Article 3(5) DMA is uncanny, especially if one bears in mind that those same arguments were not submitted by the gatekeeper in the notification phase leading to the September deadline.
But what about the broader picture?
Even though establishing the requisite legal standard to the designation and rebuttal processes is key to securing the DMA’s effective enforcement with regards to those undertakings not falling within the false negative category, the EC’s non-designation decisions demonstrate nuance and precision in excluding the four CPSs from the DMA’s scope of application.
The four CPSs were not designated based on the same reason: they did not meet the requirement of being an important gateway for business users to reach end users, as enshrined in Article 3(1)(b) DMA. This may seem an unimportant statement, but it brings some nuance into the rebuttal process. Undertakings cannot rebut the quantitative presumption based on the fact they do not meet the requirements under Article 3(1)(a) nor of Article 3(1)(c) DMA. The core of an undertaking’s gatekeeping power lies in its instrumental capacity to make business users dependent on it to reach their consumers, i.e., their end users.
According to the non-designation decisions, this characteristic as an important gateway is at odds with several questions of fact. Aside from the argument relating to the lack of effects radiated by the digital ecosystem, three main reasons were presented as sufficiently justified to exclude the CPSs from designation.
First, the limited overall usage of the CPS within the CPS category from the end user perspective. For instance, Apple’s iMessage did not merit designation due to the low intensity of its usage compared to the main messaging services and even when end users did engage with the CPS, they only did so in smaller intervals of time and less frequently as opposed to the leading messaging services in the market (paras 25 to 29 of Apple’s non-designation decision). The metrics were much more straightforward for Microsoft’s online search engine Bing. The number of search queries and clicks of the end users was low in absolute and relative terms as opposed to the rest of its competitors. They accounted for less than 5% of all monthly search queries on all major online search engines operating in the European Union and 3,3% of clicks across all devices in the Union for the 2020-2022 period (paras 23 to 25 of Microsoft’s non-designation decision). In a similar vein, Microsoft’s web browser Edge only accounted for 5,8% of all webpage views across all types of devices in Europe (para 53 of the decision).
Second, the absolute and relative business usage of the service. The point was made evident for Apple’s iMessage by establishing the small volume of business users and messages it bears, as opposed to other B2C communication channels provided by Meta such as WhatsApp API, WhatsApp Business App, or the Messenger Chat Plugin. The EC even drew into the information obtained for its Meta/Kustomer merger to derive that Apple Business Chat (the previous name of the B2C functionality of iMessage) had a very low penetration rate in the EEA, with less than 50 000 daily conversations in 2020, representing less than 0,2% of the total volume of conversations of EEA over-the-top B2C messaging (paras 29 to 34 of Apple’s non-designation decision). To gauge Microsoft’s Bing B2C aspect, the European Commission only briefly remarked on the fact the limited scale of usage by end users directly demonstrates it is not an important gateway for business users to reach them, regardless of their advertising spend on the online search engine (paras 28 to 31 of Microsoft’s non-designation decision). Alike for Bing, the European Commission did recognise advertising spend is indicative of business usage for Microsoft Advertising services. To that end, it provided the results of its market investigation where most surveyed advertisers and publishers recognised it was, at most, their secondary provider of ad spaces and they did not use its tracking tools (paras 98 to 102).
Third, the lack of control of the gatekeeper over the operations of the CPS. The argument resembles those already accepted by the EC when directly accepting the rebuttals submitted by Alphabet and Microsoft’s Gmail and Outlook.com. At this stage, however, the EC went into more detail to demonstrate Microsoft’s web browser Edge is dependent on Chromium’s Blink browser engine. Thus, the way content is rendered on Edge by business users to end users is predicated on conforming with its competitor’s engine, rather than being an autonomous choice by Microsoft (para 59). In any case, with regard to Microsoft’s Edge, the EC did assert future developments could tip the scales towards the designation. The results of the market investigation showed Edge could become a more important web browser for business users to reach end users in view of developments such as Microsoft’s launch and integration of AI-based services and features (para 69). It is, however, not completely clear whether, in such a circumstance, the EC would be compelled to start the quantitative designation anew or whether it could simply amend the current decision by altering its terms.
Key takeaways
Before the EC’s uneven interpretation of the criteria surrounding the DMA’s designation process, the recent non-designation decisions added some nuance to the analysis, despite they sometimes contradict the terms of the designation decisions the enforcer issued last September. In principle, we already know what usage thresholds will be required of those gatekeepers wishing to dodge designation. All metrics indicated their usage to be below 10%. Thus, any gatekeeper with higher usage metrics may automatically think twice about whether to submit a rebuttal to the quantitative presumption.
From the qualitative perspective, the non-designation decision also provides some breathing space to the EC. Although the notion of ecosystems was completely barred from applying in the realm of Article 3(5), it did play a role in sustaining the gatekeeper’s escape from the DMA. One could even dream about the EC’s interpretation of the DMA’s substantive provisions in line with the complex notion of digital ecosystems and the distributed impacts each one of them may bear once applied.
