Take a Big Basket, Everyone Looks Small: TikTok’s Gatekeeper Designation Hearing Before the Court of Justice (Case C-627/24 P)

A couple of days ago, the Court of Justice of the European Union (CJEU) held its first hearing relating to the DMA’s interpretation. Following the General Court’s (GC) ruling (Case T‑1077/23, see here) finding in favour of ByteDance’s designation of its social network, TikTok, by the European Commission, the designated gatekeeper presents the errors of law encountered in the judgment to make the court reach the conclusion that all its rebuttal arguments did not merit being outright rejected. In turn, the European Commission’s (EC) position still stays strong with regard to the validity of its designation decision and the procedural safeguards it applied throughout.

In this post, I will consider the two main issues discussed at the hearing: i) the legal standard contained in Article 3(5) DMA; and ii) whether the EC infringed the gatekeeper’s right to be heard when attending to its gatekeeper notification. Additionally, I will explain why I believe ByteDance’s defence missed a ripe opportunity by not engaging with the EC’s decision as a holistic assessment.

What is the legal standard for quantitative designation and how to escape it

The DMA sets out three quantitative presumptions in Article 3(2) DMA that correspond with an undertaking’s capacity to be designated as a gatekeeper. Surpassing those three thresholds entails that the undertaking incurs a presumption that it; 1) has a significant impact on the internal market; 2) is an important gateway for business users to reach end users; and 3) has an entrenched and durable position in the market (correlated with the legal requirements set out under Article 3(1) DMA).

In practice, when the undertaking providing any of the core platform services (CPSs) listed under Article 2 DMA meets those thresholds, it notifies the EC within 2 months after the thresholds have been met. In that notification, the gatekeeper includes the relevant information relating to the thresholds for each of the CPSs. Based on the notification, the EC makes its decision in relation to the designation. In ByteDance’s particular case, it notified its potential gatekeeper position on the 3^rd of July 2023 (see para 1 of the designation decision) and was designated as a gatekeeper on September 5, 2023.

Article 3(5) DMA in the moments in between gatekeeper notification and gatekeeper designation. The provision reads that the undertakings providing CPSs “may present, with its notification, sufficiently substantiated arguments to demonstrate that, exceptionally, although it meets all the thresholds (in Article 3(2) DMA), due to the circumstances in which the relevant CPS operates, it does not satisfy the requirements listed in (Article 3(1) DMA relating to an undertaking’s description as a gatekeeper)” (emphasis added).

And here’s where the EC’s and ByteDance’s discussion started to unfold in the hearing. The text above corresponds to the first sub-paragraph of Article 3(5) DMA. The following two subparagraphs illustrate that “where the EC considers that the arguments submitted pursuant to the first subparagraph by the undertaking providing the CPSs are not sufficiently substantiated because they do not manifestly call into question the presumption set out in Article 3(2) DMA, it may reject those arguments (…), without applying the procedure laid down in Article 17(3) DMA (the market investigation procedure)” and “where (it does), the EC may (…) open the procedure laid down in Article 17(3) DMA” (emphasis added).

In practical terms, Article 3(5) DMA provides gatekeepers with the possibility to rebut any of the presumptions under Article 3(2) and then different consequences apply if the EC accepts or rejects those arguments. First, it can accept those arguments outright and decide not to designate the gatekeeper, despite the notification, as it did with Samsung’s notification in 2023 of its Samsung Internet Browser. Second, it can reject those arguments at face value on the designation decision (and, thus, automatically designate the undertaking as a gatekeeper), which matches the reality of ByteDance’s rebuttal of the presumption (para 120 of ByteDance’s designation decision). And third, it can decide to check the validity of those arguments through the market investigation procedure under Article 17(3) DMA. All undertakings that went through the latter process managed to secure their CPSs non-designation under the DMA, notably Microsoft’s Bing, Ads and Edge services (now under appeal before the General Court, prompted by third party Opera) or Apple’s iMessage. I invite you to read my paper on gatekeeper designation for a more in-depth review of the topic.

According to the General Court’s ruling, “it is indisputably apparent from the terms ‘exceptionally’ and ‘manifestly’, in Article 3(5) of the DMA, that the standard of proof required of the undertaking concerned is high, in the sense that the arguments presented by that undertaking must be capable of showing, with a high degree of plausibility, that the presumptions laid down in Article 3(2) DMA are called into question” (para 71) (emphasis added). This was one of ByteDance’s main points of contention in the hearing and the first legal error that it sought the CJEU to address.

ByteDance sought to convince the CJEU by arguing that the EC should have triggered a market investigation as a consequence of the arguments they set forth because they met the legal standard under Article 3(5) DMA. To do that, it reiterated that Article 3(5) DMA provides for two different legal tests in subparagraphs one and three. On one side, the first subparagraph, under the gatekeeper’s impression, applies to exceptional circumstances where sufficiently substantiated arguments are put forward before the regulator, i.e., the highest threshold of rebuttal possible that combines an exceptional circumstance that is sufficiently substantiated. On the other side, subparagraph three requires sufficiently substantiated arguments that manifestly call into question the presumptions set out in Article 3(2) DMA.

In ByteDance’s own viewpoint, the General Court conflated both legal tests by bringing them all under the same umbrella (i.e., a high degree of plausibility), given that it would be excessively difficult for any undertaking to succeed in its rebuttal if that legal standard were to apply. When asked by the Judge Rapporteur and the President of the Court about whether those two distinct legal tests did not, in the end, render equivalent results, the gatekeeper’s representatives stressed that the three tenets would place too high a burden and sufficiently substantiated arguments manifestly calling into question the presumption only require arguments to be put forward that are backed by evidence (not necessarily and conclusive right) and that have a bearing with regard to the presumption. Throughout the course of the hearing, the EC concurred with the GC that Article 3(5) only has one legal standard, depending on the tenets of exceptionality and sufficiently substantiated arguments manifestly calling into question the presumption.

The practical consequences of the finding that ByteDance seeks are twofold. First, that the CJEU recognises that its arguments were sufficiently substantiated in manifestly calling into question the presumption (but fell under the exceptionality threshold contained under Article 3(2) DMA). Second, that gatekeepers can submit prima facie evidence to rebut the presumption, without the need to submit more solid evidence in their notification, as opposed to the GC’s rejection of evidence of such nature (paras 70 and 71 of the ruling). On this last point, ByteDance’s arguments were quite sound in focusing on the object of the rebuttal, rather than on its nature. In the gatekeeper’s view, it is not the actual designation that the gatekeeper must seek to rebut, but rather the application of the presumption, as stated under Article 3(5) DMA. One has to concede that it is reasonable to believe that an undertaking cannot anticipate the breadth and depth of evidence that it will need to rebut the presumption when interactions happen in a one-off interaction. Due to this reason, it makes sense that the gatekeeper must first rebut the presumption and, after that, in the market investigation process, seek to rebut the gatekeeper designation itself.

The EC did not oppose this view, although it did emphasise the following questions by Advocate General Biondi that the burden of proof stayed with the undertaking throughout the notification process, right through to the market investigation. In other words, the EC acts as a sounding board to the gatekeeper’s arguments (and checks their reasonableness and validity) but does not engage with additional evidence that may question the undertaking’s position in the market, even if it may conflict with its characterisation as a gatekeeper. On top of that, the EC emphasised that it is entitled, and not forced, to trigger a market investigation under its margin of discretion, given that Article 3(5) DMA requires it when the legal standard is met. That is the reason why it decided not to pursue this path in cases where it was particularly clear that the rebuttal was to be accepted. The applicant contested this point by arguing that Article 3(5) conveys the options available to the EC, and not the circumstances under which the market investigation is triggered. In any case, according to ByteDance, the word ‘may’ must be interpreted as an empowered that is prompted by the triggering circumstances laid down in the law, which is further reinforced with the fact that Article 17(3) DMA declares that the EC shall conduct the investigation.

If the CJEU were to find in favour of ByteDance, it would open the gate for prima facie evidence to be submitted in the rebuttal of the presumption, but I concur with the sitting judges that opting for one interpretation or the other will not automatically lead to a substantial tempering of the legal standard that the gatekeeper must meet to avoid designation. Although ByteDance believes that exceptionality (i.e., something unusual or extraordinary) is a higher standard, my view is that the adverb manifestly places the gatekeeper in a condition where it must present unmistakably clear evidence that requires no further proof, if one were to stay with a literal interpretation of the term.

Multi-homing, market capitalisation and fast-followers as valid arguments

Within ByteDance’s main defences, the gatekeeper further presented reasons of the factual evidence operating in its favour, justifying its de-designation under the DMA, in seeking that the CJEU annulled the EC’s designation decision operating against it. For this particular purpose, ByteDance showcased the three main grounds that sustain its rebuttal of the presumption; 1) the fact that a great part of its market capitalisation lies in its Asian business; 2) the existence of multi-homing; and 3) the presence of fast followers that acted as copycats to its services. By doing so, the gatekeeper argues that the EC failed to show that ByteDance has a significant impact on the internal market (Article 3(1)(a) DMA) and is an important gateway for business users to reach end users (Article 3(1)(b) DMA).

Regarding the gatekeeper’s significance in the internal market, Article 3(1)(a) DMA establishes alternative thresholds that can be met. Either the undertaking achieves an annual EU turnover equal to or above EUR 7,5 billion in each of the last three financial years, or its global average market capitalisation or equivalent fair market value amounts to at least EUR 75 billion in the last financial year, and it provides the CPS in at least three Member States. The EC can choose to apply one or the other. In this particular case, ByteDance did not meet the former but surpassed the latter. Thus, the EC presumed its significance in the market as a consequence.

When it was given the chance to respond to the EC’s initial findings, the gatekeeper voiced out its concern that this “business ha(d) no connection to the Union, face(d) very different competitive dynamics, and operate(d) in a distinct regulatory, linguistic and cultural environment. Therefore, (…) the global fair market value (threshold) is not indicative of TikTok’s current or potential future ability to monetise users in the Union” (para 100 of ByteDance’s designation decision). The EC termed the argument as irrelevant because “the fair market value of the undertaking (operates) as a whole and not in relation to a specific geography” and actually that it “stands in direct contradiction with Article 3(2)(a) DMA, which (…) is aimed at capturing the financial capacity of the undertaking concerned, including its access to financial markets and the resulting ability, for example, to acquire other, innovative undertakings providing similar services” (para 124).

The General Court established that the EC had incurred in an error in its reasoning because “the undertaking concerned may demonstrate, on the basis of a number of sufficiently substantiated arguments, including its low Union turnover, that, despite its global market value (…) it has only a limited presence on the internal market, so that that market value does not reflect, on account of the circumstances in which the relevant CPS operates, a potential to monetise its users in the Union in the near future” (para 91 of the GC’s ruling). Thus, it had been wrong in terming the reasoning irrelevant.

However, the GC went into considering whether the China-focused market cap would merit its characterisation as a sufficiently substantiated argument manifestly calling into question the application of the presumption. In this sense, the GC noted that “the Commission rightly considered that the fact that ByteDance’s fair market value was mainly due to its operations in China was irrelevant” because “the global fair market value (does not draw) any distinction as to the geographical origin of that value” (para 104), so that it does not automatically have a bearing on “ByteDance’s ability to use its value generated in China or elsewhere in the world to strength its position in the European Union” (para 105).

In any case, the GC accepted that the burden of proof stayed with the gatekeeper throughout designation and, for instance, indicated that it was up to the undertaking to prove whether facing cultural and regulatory barriers in the EU made a difference in terms of its designation as a gatekeeper because they meant that it “could not benefit, in the internal market, from its global fair market value” (para 106). Given that ByteDance did not produce such evidence, it could not say that the presumption was called into question. Here’s where the dichotomy that ByteDance addressed to the CJEU manifests relating to the rebuttal to the presumption or, rather, the rebuttal to the gatekeeper designation as a whole. The gatekeeper made the same point before the CJEU, trying to rely on factual evidence (that the Court cannot, in principle, review) proving that it did not merit designation.

Notwithstanding, the gatekeeper’s line of reasoning prioritised the discrediting of the important gateway requirement via low multi-homing evidence and a low presence in the competitive landscape of social networks. In its interactions with the European Commission, the undertaking defended that multi-homing by TikTok business users and end users demonstrates the absence of lock-in effects (para 103 of ByteDance’s designation decision). Following the economic literature, the EC established that the former does not necessarily correlate with the latter, nor does it preclude the possibility of weak contestability and unfair practices (para 136).

Within the discussion, the GC settled that the presence of multi-homing in social networks “is a common practice” (para 183 of the General Court’s ruling) and that its asymmetric nature “can be explained by the economic and historical context in which online social networks operate in the EU”, i.e., Facebook’s and Instagram’s dominance in the space prior to TikTok’s launch in the internal market. Thus, “it was to be expected that a large proportion of TikTok users were already using Facebook and Instagram prior to the launch of TikTok and would continue to do so, in parallel, thereafter” (para 198). The argument held, in the GC’s view, insubstantial value for rebutting the application of the presumptions.

During the hearing, ByteDance submitted three responses to the GC’s definition of an important gateway in the sense of Article 3(1)(b) DMA. First, it accounts for the DMA’s broader context and emphasises that high user numbers (as required under the corresponding threshold under Article 3(2)(b) DMA) entail that there is a sense of economic dependency between the undertaking and its business users, following the spirit of Recitals 20 and 75. It follows, then, that in the presence of other effective channels where business users can access the same nature of social networks, economic dependency does not arise. Second, it alleges that the GC distorted the evidence presented to it by saying that social network users multi-home and, thus, it is not a relevant factor for examination. Given that its main rival in the market, Meta, holds a powerful ecosystem that keeps users locked-in and multi-homing is low, one cannot simply do away with the effects of multi-homing with respect to its CPS. And third, it points to the competitive landscape of the market (based on evidence that the GC found inadmissible) whereby it only occupies 8-12% of the market.

To respond to such allegations, the EC briefly noted that the mere existence of multi-homing is not enough to show low levels of lock-in and, still less, to try to rebut the quantitative presumption. As a matter of fact, the EC highlights several points that it had already manifested in its designation decision, notably the fact that TikTok has a significantly higher engagement rate than other online social networking services, with end users spending significantly more time on TikTok as compared to other social networks (para 141 of ByteDance’s designation decision).

To top it all off, the designated gatekeeper went on to reiterate that several social networks acted as fast followers to its short-video format (para 110). Therefore, no presumptive power can be said to be applied with respect to contestability, in ByteDance’s own words in the hearing. The EC did not respond to those allegations, based on the fact that the allegations laid outside of the scope of the General Court’s ruling.

Although the judge rapporteur sought to clarify ByteDance’s position with regard to what exact passages of the GC’s ruling the applicant considered to be in error, he even asked about those instances where the GC mentioned the potential impact of the lack of an ecosystem on the undertaking’s position. ByteDance dodged the question by going back to its main arguments surrounding multi-homing and network effects. In any case, through his questions, the case’s judge rapporteur managed to settle that the errors of law pointed out by ByteDance were more procedural than substantive in nature.

Assuming the factual nature of most of ByteDance’s arguments with regard to its characterisation as a gatekeeper, one can be fairly optimistic about the EC’s designation decision holding sufficient ground with regard to this particular aspect.

The crux of the designation procedure

During the summer of 2023, the EC was faced with a new set of procedural rules that applied to some undertakings for the first time in its and their history. According to the DMA’s Implementing Regulation, gatekeepers had to clearly identify which of the three cumulative requirements set out in Article 3(1) its rebuttal arguments related to and, for each of them, explain why the relevant CPS exceptionally did not satisfy the requirement, despite the fact that it might meet the corresponding threshold under Article 3(2) DMA (Article 2(3) of the Implementing Regulation). Prior to the notification, ByteDance and the EC held four meetings on the undertaking’s potential designation decision (para 345 of the GC’s ruling). Following the undertaking’s notification of their potential gatekeeper position, the EC was bound to produce its findings in the short period of time of 45 days, as provided by Article 3(4) DMA.

Nothing in Article 34 DMA (regulating the undertaking’s right to be heard and access to the file) nor in the Implementing Regulation compelled the EC to address an SO-like to the undertaking prior to its designation. However, as recognised by ByteDance and the EC during the hearing, out of precaution and given that Article 41 of the Charter applies in any case, the regulator sent a letter to the gatekeeper illustrating the responses to its rebuttal of the quantitative presumption only 10 days after the notification had been received. 15 days later to having received ByteDance’s response to its preliminary views, the applicant and the EC held an additional meeting (para 345). And here’s where the dance starts.

ByteDance describes the letter as the preliminary findings that the EC must issue in the midst of a non-compliance or specification proceeding when the regulator triggers its punitive and non-punitive powers (quod non). Therefore, it raises its nature way above its object and holds the EC accountable for the substantial changes introduced thereof involving, for instance, the regulator’s findings that ByteDance’s lack of an ecosystem or high multi-homing were irrelevant to its designation. To this effect, ByteDance argues that the EC added five entirely new facts to its designation decision to make it robust enough for its judicial scrutiny, based on information that it extracted from online articles and media blogs (as can be seen in footnotes 143, 145, 152, 153 and 155 of ByteDance’s designation decision). In response to the allegations, the EC recognises that this letter could not have contained its complete assessment since it was issued only 10 days later to the undertaking’s notification.

Based on the disparity of the expectation created by the letter and the final result, the designated gatekeeper intended to show the CJEU that the GC was wrong in excusing the EC’s behaviour under the premise that the gatekeeper already had the chance to respond to those allegations and decided to refuse to do so in the administrative proceedings. This is the case, for example, for ByteDance’s argument that it lacked an ecosystem, that the GC recognises that “the Commission had not informed the applicant of its view that, as a matter of fact, ByteDance did have an ecosystem” (para 353 of the GC’s ruling). Given that the GC considered that the rest of the grounds set out in the designation decision were sufficient in themselves to justify the EC’s finding that those arguments were not sufficiently substantiated so as to manifestly call into question the presumption (para 356). In parallel, however, the GC highlighted that the “undertaking concerned cannot submit, for the first time before the Court, arguments or evidence, under Article 3(5) DMA, which it had not submitted during the administrative procedure in rebuttal of the presumptions” (para 234).

In my own view, introducing the doctrine of exhaustion of administrative remedies is problematic for a number of reasons (which you can read here), and the gatekeeper’s position before the EC’s potential gatekeeper designation is always framed as a lose-lose situation against it. When the EC grants a period of allegations for the gatekeeper to contest its preliminary views (that it is not forced to do) on the designation, it is setting a trap that the gatekeeper cannot easily escape. On one side, when the gatekeeper takes the EC’s word at face value and does not insist on those problematic aspects of designation, then it will suffer the consequences under judicial scrutiny. On the other side, if the gatekeeper maintains its position during the designation process, then the EC only attends to those arguments in the final version of its designation decision, placing it in a prominent position to have the last word regarding any particular aspect of the DMA’s implementation. Building on the idea that the EC only considers throughout designation (and, potentially, in its market investigation procedure under Article 17(3) DMA) that the arguments that the gatekeepers raise have to be checked against the economic and legal reality, the regulator’s use of the informal letter operates more in its favour than it does against it, to the extent that it indirectly narrows the terms of the discussion with the undertaking.

Whether this is something that the CJEU can resolve in connection with ByteDance’s rights of defence and to be heard is a completely different matter, bearing in mind that ByteDance failed to portray how the decision would have been different in the absence of the two alleged procedural irregularities. This is precisely where I believe that the gatekeeper’s defence would have profited from setting out in the hearing one of its main points of contention: the consequences of the arguments for the rebuttal on the overall designation decision (aka what ByteDance references as a lack of a holistic assessment on the EC’s part).

Silver bullet or one robust corpus of evidence

Stemming from the discussion on the legal standard that follows from Article 3(5) DMA, I find that the General Court’s statement that none of ByteDance’s arguments “taken individually or as a whole in the context of each of the presumptions laid down in Article 3(2) DMA, were (…) sufficiently substantiated so as manifestly to call those presumptions into question” (para 333 of the GC’s ruling), quite puzzling. Considering that ByteDance’s appeal of the designation decision before the GC and now before the CJEU followed the expedited procedure, this is the first time that any EU court has pronounced itself with respect to the designation process, despite the many appeals pending before the General Court.

It is in this context that I find the GC’s passage above mindboggling. In the counterfactual exercise that all of ByteDance’s arguments would have succeeded before the General Court, i.e., that it does not have an ecosystem that does not benefit from significant network effects, that there are no significant lock-in effects, that the scale of TikTok is smaller than that of some of other online social networking services and that the level of engagement of advertisers and business users registered with the platform is minimal, the win would have been tainted by the outcome. The General Court would not have backpedalled the EC’s designation decision and declared that those points met the rebuttal standard under Article 3(5) DMA, neither individually nor taken as a whole.

This demonstrates that none of those arguments could have acted as a silver bullet that would have resulted in the General Court’s annulment of the decision. Maybe such types of allegations simply do not exist under the DMA, and that’s fair. In turn, it also illustrates that the EC’s designation decision must be regarded as a corpus of evidence that is so sound from the legal and economic perspective that it cannot be simply knocked down by a round of punches. Ironically, that cannot be right if the EC’s designation decision relies on the application of a presumption that requires of very little economic evidence being set forth before the undertaking’s eyes. One of the two options must be right, or both at the same time, within the strange regulatory world of the DMA.

In any event, I believe that ByteDance could have pointed out these inconsistencies to at least make the CJEU doubt about how designation and rebuttal work. From the place where I am standing, it seems that designation will always follow through when the EC has made its mind that it is necessary (because when it doesn’t, a market investigation follows).