The long-awaited judgment from the European Court of Justice (“CJEU”) in the Google Shopping saga has finally arrived. On 10 September 2024, following a case that started with an investigation in 2010 and spanned nearly one and a half decades, the CJEU rejected Google’s appeal and delivered a huge victory for the European Commission (“Commission”).
This blog will first briefly address the background of the case, followed by a summary of the key findings by the CJEU and, finally, the wider implications of the judgment.
Background
In 2017, the Commission fined Google €2.4 billion for abusing its dominant position on the market for online general search services in 13 EEA countries by favouring its own comparison shopping service (“CSS”) over competing CSSs. This abuse consisted of two practices: first, Google actively promoted its own CSS by displaying it higher up in the general search results page (“SERP”), and second, Google simultaneously demoted competing CSSs by using ranking algorithms which led to competing CSSs appearing further down in the search results. These two practices led to a significant decrease in traffic to rival CSSs, which weakened their position on the market, even resulting in some rivals exiting the market altogether. The novelty of the case lay in the Commission considering this anti-competitive leveraging by Google as an independent form of abuse, now widely known as “self-preferencing” (although technically neither the Commission nor the Courts referred to it as such, rather calling it “leveraging” or “favouring”). Google appealed the Commission’s decision to the General Court (“GC”), which in 2021 dismissed the appeal in nearly all parts and upheld the Commission’s original fine. For a more extensive analysis of the GC judgment, please see our earlier blog post here.
CJEU Judgment
Google raised four grounds of appeal before the CJEU, all of which were dismissed. Below, we discuss the most significant parts of the Court’s reasoning.
- The applicability (and the end) of the Bronner criteria
Since the beginning, Google has maintained that the case concerned the alleged refusal of access to the Shopping Unit on the SERP. Following the CJEU’s judgment, this discussion can finally be put to rest. In short, the Court sided with the Commission and the GC in stating that the Bronner criteria did not apply to the case at hand.
First, the CJEU reiterated the applicable case law in refusal to supply cases and agreed with the GC that not every issue of access requires the conditions set out in Bronner to be applied (para 81). The CJEU emphasised that the description of the conduct at issue “makes it clear that the conduct concerned the discriminatory positioning and display on the general results pages of Google’s general search service and not access to the boxes” (para 99). In essence, although the case concerned access in general, the CJEU endorsed the GC’s finding that it concerned the conditions for access to Google’s SERP and not access to the Shopping Unit at the top of the SERP as Google claimed. As rival CSSs had already been given access to the SERP, the Bronner criteria were not applicable (para 113).
A similar question on the applicability of Bronner has been raised in another ongoing case involving Google, the Android Auto request for a preliminary ruling. In her recent opinion on the case, AG Medina also stressed the limited applicability of Bronner, arguing that it should only apply in very specific cases where the infrastructure at issue has been developed for the exclusive use of the dominant company. Although the reasoning of the CJEU in Google Shopping and AG Medina in Android Auto differs to an extent, the general consensus seems to be that Bronner is the exception, not the rule, even in cases relating to access.
- Anti-competitive leveraging as an independent form of abuse
Second, the Court considered the circumstances to be considered in determining whether Google’s conduct was anti-competitive. The Commission had particularly focussed on three factors, i.e. (i) the importance of traffic generated by Google’s SERP for CSSs, (ii) user behaviour, and (iii) the fact that diverted traffic from Google’s SERP accounted for a large proportion of traffic to competing CSSs and could not be effectively replaced by other sources (paras 141 and 169-173 of the GC judgment).
Google argued that these circumstances did not concern the nature of its conduct and, therefore, did not provide a valid basis for deciding whether the conduct departed from competition on the merits (para 126). The GC had previously concluded that the Commission used these specific circumstances to place Google’s promotion of its own CSS and the demotion of rivals “within a particular context” (para 145).
The CJEU confirmed that demonstrating actual or potential effects must, in general, be done “in light of all the relevant factual circumstances” (para 167), which includes the circumstances concerning not only the conduct itself but also the market(s) in question and the functioning of competition on those markets which actually makes a lot of sense. The Court, therefore, concluded that the circumstances identified by the Commission were essential in demonstrating the potential anti-competitive effects of Google’s conduct and should have been considered since they “enabled those practices to be placed in the context of the two relevant markets and the functioning of competition on those markets and were thus capable of demonstrating that potential exclusionary effects on the downstream market (…) were due not to the merits of that service but to those practices combined with the specific circumstances identified” (para 171).
This set the scene for another profound aspect of the judgment, where the CJEU stated that more favourable treatment of its own services by a dominant undertaking in isolation is not per se anti-competitive. To quote the CJEU in paragraph 186: “It is important to add that it cannot be considered that, as a general rule, a dominant undertaking which treats its own products or services more favourably than it treats those of its competitors is engaging in conduct which departs from competition on the merits irrespective of the circumstances of the case.” (emphasis added)
Does this mean that by contrary analysis self-preferencing is generally allowed, as some commentators are suggesting? We would argue no. In essence, the CJEU confirmed that this type of self-preferencing behaviour cannot be assessed in isolation and that consideration must be given to the circumstances and the context within which it occurs. Therefore, in specific cases, favourable treatment may be an abuse, but it cannot be stated as a general rule. Regardless, we do not think this part of the judgment should be read as an invitation for self-preferencing conduct.
- AEC test is neither mandatory in all abuse of dominance cases nor relevant in the current case
Last, Google claimed that the GC was wrong to hold that the Commission did not need to examine whether Google’s self-preferencing conduct was capable of foreclosing equally efficient competitors. Google argued that a formal AEC price-cost test deployed in cases concerning pricing-related practices should be distinguished from a general principle according to which Article 102 TFEU is not aimed at protecting less efficient competitors (para 253). According to Google, in line with that principle, it should always be assessed whether allegedly abusive conduct can foreclose as efficient competitors.
The CJEU rejected Google’s argument. It held that even though Article 102 TFEU is not aimed at protecting less-efficient competitors, not every finding of an abuse of dominance requires proof that the conduct is capable of excluding as-efficient competitors (paras 263-264). The CJEU then referred to its previous findings in Intel on how the AEC principle becomes particularly relevant where the dominant undertaking submits during the administrative procedure that its conduct is not capable of restricting competition and producing the alleged foreclosure effects, supported by evidence (para 265). Then, the Commission must not only assess the extent of the dominant position but also the possible existence of a strategy aiming to exclude as efficient competitors (para 265). The CJEU then reiterated that the Commission must demonstrate the infringement of Article 102 TFEU, which may include relying on the AEC test where relevant (para 266). However, the Court accepted the GC’s reasoning that in the current case, “it would not have been possible for the Commission to obtain objective and reliable results on the efficiency of Google’s competitors in the light of the specific conditions of the market in question” (since the conditions of competition were distorted by Google’s anticompetitive behaviour) (paras 267-268). The Court, therefore, confirmed that an AEC test is not mandatory when applying Article 102 TFEU and that it was not relevant in the case at hand.
The Court’s stance on the (non)necessity of the AEC test was to be expected as it confirms the approach taken in its previous judgments in Intel and Unilever Italia, extending the line of reasoning beyond cases on rebate schemes or exclusivity clauses. Whereas the Court seemed to focus on the fact that it was neither relevant nor possible to obtain information on the efficiency of Google’s competitors in the current case, AG Kokott reasoned differently in arguing why she deemed the AEC irrelevant in the case. She highlighted the role that less efficient competitors can play in exerting competitive pressure on the dominant undertaking, particularly in markets where it is unlikely for other undertakings to be as efficient as the dominant undertaking.
What are the broader implications of the judgment?
The judgment is a historic win for the Commission. Commissioner Vestager even called it a “catalyst for change, inspiring a more vigilant and proactive approach to regulating big tech and ensuring a fairer digital marketplace”. Already before the final judgment, “self-preferencing” had made waves and even inspired legislation such as the Digital Markets Act (“DMA”) . Although Article 6 (5) of the DMA imposes a per se prohibition on self-preferencing, that article is narrower with regard to the conduct it covers. It only applies to a limited number of companies that have been designated as “gatekeepers” by the Commission and applies specifically to ranking.
The CJEU judgment confirmed the status of self-preferencing as an independent form of abuse to which the refusal to supply criteria laid down in Bronner should not apply. This validates the approach taken by the Commission so far and enables both the Commission and the National Competition Authorities to hold dominant undertakings liable for leveraging their dominant position on one market to exclude competition on another market, potentially leading to increasing scrutiny by authorities towards vertically integrated dominant companies. Nevertheless, the Court neither deemed self-preferencing a per se abuse to be prohibited in each case nor did it give companies the go-ahead to favour their own products. What will be determinative is the circumstances of each case as a whole, including the characteristics of the market.
As for the AEC test, the Court very clearly stated that it only needs to be applied where relevant. In many (digital) markets, such as in the current case, where high-entry barriers and strong network effects make the emergence of an as-efficient competitor virtually impossible, the application of the AEC test will be both irrelevant and impossible. In any case, its scope of applicability has been narrowed down to price-related conduct. This aligns with the Commission’s draft Article 102 Guidelines, which refer to the price-cost test as ‘generally inappropriate for assessing whether non-pricing practices depart from competition on the merits’ (para 56). However, this does not mean that the Commission does not have to demonstrate the existence of (at least) potential anticompetitive effects, but there is no systematic obligation on the Commission to rely on the AEC test in doing so.
Finally, this loss means that Google must now face CSSs (and potentially other affected Verticals) in civil courts throughout Europe for damages suffered from 2008 until the day of (future) compliance, likely ushering in a new era of litigation, including further spotlight on Google’s compliance with Commission decision and the DMA.