ECJ’s Google Shopping Judgment: The End of a Long Saga

The European Court of Justice released its long-awaited judgment¹ in the Google Shopping saga last week, finally putting to bed close to fifteen years' of scrutiny into Google's practices of favouring its own comparison shopping service (Google Shopping) over rival shopping services.

In its ruling, the ECJ upheld the General Court's earlier judgment² which had rejected Google's appeal over the European Commission's decision³ to fine it €2.42 billion for abusing its market dominance as a search engine by systematically favouring Google Shopping in its general search results.

The overall outcome of the ECJ's reasoning in Google Shopping is perhaps unsurprising to competition law practitioners - given the unwavering direction of travel of the case. The ECJ judgment nevertheless raises a number of interesting points and leaves a number of questions unanswered.

Key takeaways

• Refusal to supply. The judgment confirmed that not every issue of access necessarily requires the application of the Bronner test of refusal to supply. The ECJ found the Bronner doctrine applies in circumstances where a dominant firm refuses to grant a competitor access to infrastructure which it has developed for its own business needs. However, the ECJ ruled that the Bronner test is not applicable in cases where there is no outright refusal of access to infrastructure - but rather access granted on discriminatory terms (such discrimination being assessed under separate forms of potential abuse).
• Competition not on the merits. The ECJ accepted Google's arguments that, to establish an abuse of dominance under Article 102, a two-pronged test applies: (i) that actual or potential anticompetitive effects arise from the abusive conduct; and (ii) that the conduct falls outside of "competition on the merits". However, in assessing the latter requirement, the ECJ rejected Google's arguments that only circumstances relating specifically to Google's conduct are relevant to the assessment. Instead, the ECJ held that, in assessing "competition on the merits", relevant circumstances regarding the characteristics of the market or the nature of competition are capable of characterising the conduct as falling outside of the scope of competition on the merits.
• Causality and counterfactual. The ECJ maintained that the causal link is one of the essential elements of a competition law infringement and that, as a result, the burden of proof for such causal link (and hence the counterfactual analysis) lies with the Commission. However, the ECJ found that the counterfactual analysis is just one way to establish causality. Where establishing a credible counterfactual may be "arbitrary or even impossible" (para 231), the Commission cannot be required to systematically establish a counterfactual and can rely on other evidence to establish causality.
• "As-efficient competitors". The ECJ reiterated earlier case law that it is not the objective of Article 102 to ensure that less efficient competitors remain on the market but also remarked that this statement did not imply that an abuse of dominance finding does not always require a showing that the conduct was capable of excluding an as-efficient competitor. With respect to the AEC test, the Court held that this is just one way to establish an abuse of dominance.

Background

Crucial to the Commission's finding of abusive conduct was that Google had given Google Shopping an unlawful advantage in two ways through:

• The promotion of Google Shopping - by displaying its results in a more eye catching and prominent manner in Google's general search results (being displayed at the top of search results or in reserved "boxes"); and
• The demotion of Google Shopping's competitors - by showing those pages only as general search results (i.e., in blue links, without rich format), thereby pushing the pages down the list of search results due to various adjustment algorithms. Google did not apply the same adjustment algorithms to Google Shopping.

The Commission found that such practices had potential anticompetitive effects on the comparison-shopping market, as necessary and irreplaceable search traffic was diverted from competing providers whilst traffic to Google Shopping increased. Moreover, this had potential anticompetitive effects on the general-search market by generating increased click-through revenue to Google via Google Shopping traffic.

Google unsuccessfully appealed to the GC, arguing that the decision contained errors in law and fact in concluding that Google's conduct was abusive and likely to have anticompetitive effects. In a clear win to the Commission, the ECJ quashed Google's last attempt to overthrow the decision.

Main findings of the ECJ

The application of the Bronner "refusal to supply" test

Central to both the GC and ECJ appeals was Google's argument that its conduct in question was a refusal to supply access to its "boxes" and that the Commission had failed to establish the conditions of a "refusal to supply" under the Bronner "essential facilities" doctrine. However, both the GC and ECJ distinguished the case on the basis that the Commission had accused Google of discriminatory access - and not an outright refusal - to its general search page.

First, the ECJ found that the boxes in which Google Shopping's results had been displayed did not constitute a separate "facility" from Google's general results pages (noting that both the boxes and general search results were generated on the same results page). As Google's competitors had access to the general search page, no refusal of access to a facility had occurred. Second, even if the boxes were a separate facility for the purposes of Bronner, Google's competitors had not requested or been denied access to the boxes. Indeed, the Commission had not challenged Google's failure to provide competitor access to those boxes (or the general search page), but the overall "discriminatory positioning and display on the general results pages of Google's general search results" (para 99). In this regard, the ECJ also rejected Google's argument that the Commission's suggested remedy to provide competitor access to the boxes was evidence of the existence of a refusal to supply case. The ECJ agreed that this was one such remedy acceptable to the Commission, but noted that the Commission equally accepted the removal of the boxes altogether such that competitors would be displayed with equal prominence on its general search pages. Thus, the ECJ held, Bronner did not apply.⁴

The ECJ therefore found that these practices constituted, and had been assessed, as separate forms of leveraging abuse under Article 102. For these practice, "indispensability" was not a necessary condition to establish abuse but likely a relevant factor for any effects-based assessment (para 111).

The assessment of "competition on the merits"

The ECJ rejected Google's arguments that the Commission had failed to establish as a separate condition from a finding of actual or potential anticompetitive effects that its conduct departed from "competition on the merits".

The ECJ accepted that to categorise a conduct as an abuse of a dominant position, it is necessary to demonstrate both (i) the actual or potential anticompetitive effects arising from such conduct, and (ii) the existence of competition not on the merits. However, the ECJ rejected that the Commission's assessment had been purely speculative, finding that the Commission had justified its analysis on "suspect elements in the light of competition law" (namely an unjustified differential treatment) and second, "of specific relevant circumstances relating to the nature of the infrastructure giving rise to that difference in treatment" (para 158).

Indeed, the ECJ acknowledged that there is no general rule prohibiting a dominant undertaking from treating its own products or services more favourably than those of its rivals, nor that engaging in such conduct would constitute conduct departing from "competition on the merits" irrespective of the circumstances of the case. However, it pointed out that neither the GC nor the Commission had merely noted the existence of such favourable treatment by Google of Google Shopping, but "having regard to the characteristics of the upstream market and the specific circumstances identified" the conduct (namely the promotion of Google Shopping and demotion of competitors) "was discriminatory and did not fall within the scope of competition on the merits" (para 187).

With regard to the "specific circumstances identified", the ECJ also rejected Google's arguments that such circumstances could only be relevant if they related solely to Google's conduct at issue. Citing European Superleague⁵the ECJ found that relevant factual circumstances include those that concern the market in question and the context in which the conduct arises. For Google Shopping, three specific circumstances of the case were relevant to the assessment, being: (i) the positive network effects resulting from increased traffic to comparison shopping providers (i.e., the more a comparison shopping service is visited, the greater the relevance and usefulness of its services and the more merchants are inclined to use them); (ii) the behaviour of users when searching online (i.e., that users typically concentrate on the first three to five search results, and pay little to no attention to the remaining results); and (iii) the impact of diverted traffic (i.e., that the proportion of traffic diverted from competing shopping services could not be replaced by other sources). Whilst not arising from Google's conduct, the ECJ thus held that such circumstances were "relevant circumstances capable of characterising the existence of practices falling outside of the scope of competition on the merits" (para 162).

Counterfactual and causality

The ECJ rejected Google's arguments that the Commission had failed to establish a causal link between the diversion of search traffic from competitor shopping providers and Google's conduct, as it had not carried out the required counterfactual analysis. In doing so, the ECJ made a number of important clarifications regarding the counterfactual analysis.

First, the ECJ recalled that a "causal link is one of the essential constituent elements of an infringement of competition law which is for the Commission to prove" and rejected that an undertaking invoking a counterfactual analysis to challenge the Commission's assessment would reverse this burden of proof (para 224). In the ECJ's view, the GC correctly invited Google to provide evidence to counter the Commission's findings.

The ECJ clarified that the counterfactual analysis is just one of various tools at the Commission's disposal to evidence a causal link between a conduct and its effects. In some scenarios (including in Google Shopping), identifying a credible counterfactual scenario can be "arbitrary or even impossible" and that "the Commission cannot be required to systematically establish such a counterfactual scenario" (para 231). In line with the GC, the ECJ held that the Commission permissibly drew a correlation between the abusive conduct and the anticompetitive effects by observing the evolution of the markets and of its market participants as a valid alternative to the counterfactual analysis.

Moreover, the ECJ found that the relevant counterfactual in Google Shopping would necessarily involve both elements of Google's abusive practices (i.e., the promotion of Google Shopping and the demotion of competing comparison-shopping services) because "it was only by being combined that these two practices at issue influenced user behaviour in such a way that traffic from Google's general results pages was redirected" (para 244). The ECJ therefore concluded that "an appropriate counterfactual scenario should have also made it possible to examine the likely development of the market in the absence of both of those practices" (para 245). It rejected Google's attempt to demonstrate the counterfactual as based on the singular alternatives (i.e., the promotion of Google Shopping, or the demotion of competing comparison-shopping services).

The significance of 'as-efficient competitor' concept and the role of the AEC test

Finally, the ECJ rejected Google's arguments that, in assessing whether its practices were abusive, the Commission should be required to consider the effectiveness of competing comparison-shopping providers and should have applied the AEC test.

With respect to the concept of the 'as-efficient competitor', the ECJ noted that "the objective of [Article 102] is not to ensure that competitors less efficient than the dominant undertaking remain on the market" (para 263). The ECJ specified that this statement does not imply, however, that any finding "is subject to proof that the conduct concerned is capable of excluding an as-efficient competitor" (para 264).

As to the narrower question of the role of the AEC test, the ECJ noted (citing Intel v Commission⁶)that the AEC test is likely to be relevant where the dominant undertaking has produced such evidence to show that its conduct is not capable of restricting competition.

The ECJ found that the Commission is required to demonstrate the abuse of a dominant position by considering "various criteria", with the AEC test being just on of them (and only to the extent the test is relevant) (para 266)). In Google Shopping, the conduct in question consisted of discrimination that allowed Google to divert traffic from competing comparison-shopping service providers. Given that the ability of competing shopping services depended on the availability of traffic - as "increased investment in alternative sources was not an 'economically viable' solution" - the ECJ held that the GC was right to conclude that "it would not have been possible for the Commission to obtain objective and reliable results concerning the efficiency of Google's competitors" (para 268).

Concluding remarks

In a month of upheaval (Illumina / Grail), the Google Shopping judgment is a win for the Commission (and Commissioner Vestager). Nevertheless, whilst a definitive ruling, the judgment raises a number of interesting points and several open questions remain:

• First, although the ECJ was clear that the Bronner test applies only to refusal to supply cases (and not to cases of discriminatory access), this largely seems to have turned on a rejection of the "boxes" as separate infrastructure from the general search page (meaning that there had been no refusal of access to an essential facility) and the identification of the relevant conduct as a combination of practices which go beyond refusal. The judgment did not provide, however, any clear framework as to how to determine the correct "infrastructure" (or facility) for the application of the Bronner test. Nor does the ECJ link its conclusions to the underlying key rationale for the higher indispensability standard in Bronner, i.e.,the incentive to innovate and the impact on dynamic competition.
• Second, the dividing line between lawful and abusive self-preferencing remains unclear.⁷ The ECJ noted that there is no general rule prohibiting a dominant firm from treating its own products or services more favourably than those of its competitors, and such conduct must therefore be assessed under an effects-based analysis. The outcome in Google Shopping seems to have been largely driven by the "specific circumstances" of the case (including the existence of positive network effects and the behaviour of users online) which influenced the magnitude of the effect of the conduct on competitors and moved lawful self-preferencing to abusive conduct. The ECJ's Google Shopping judgment has similarities with the MEO case⁸ on discrimination which suggests the Court may have more to say in future cases regarding the potential scope of a discrimination abuse.
• Third, the ECJ (at least implicitly) endorsed the two-pronged test of (i) competition not on the merits and (ii) actual or potential anticompetitive effects as required to establish an abuse of a dominant position. At the same time, it made clear that factors relating to market characteristics, rather than conduct characteristics, could establish that certain practices deviate from competition on the merits, which begs the question as to how far one must go to analyse the nature of the conduct on the market as a standalone factor.
• Furthermore, the ECJ indicated that determining counterfactual scenarios is just one way to establish causality and that the burden for proof for causality, as a constituent element for an abuse, lies with the Commission. However, one may query the categorical statement of the Court that in cases where the conduct consists of a combination of practices (in this case (i) the display of the Google Shopping result in the boxes and (ii) the demotion of the results of Google Shopping's rivals), the correct counterfactual is a world where neither of those practices takes place, particularly where each individual practice on its own is not held to be unlawful. Arguably, in those circumstances the correct counterfactual should be the likely (lawful) conduct which the dominant firm would pursue in the absence of the abuse.
• Finally, on the question of the role of the "as-efficient competitor" and, more specifically, the AEC test, the ECJ took a balanced view. The judgment highlighted that the objective of Article 102 is not the protection of less efficient competitors and pointed to the important role of the AEC test in rebutting presumptions of certain by object abuses. At the same time, the ECJ acknowledged that a finding of an abuse does not require demonstration that the conduct in question is capable of excluding as an-efficient competitor. Furthermore, the ECJ stated that the AEC test is neither mandatory nor always relevant.

[1]Google and Alphabet v Commission (Case C-48/22 P), 10 September 2024.

[2]Google and Alphabet v Commission (Case T-612/17), 10 November 2021.

[3] Commission Decision, Google Search (Shopping), Case AT.39740, 27 June 2017.

[4] The ECJ did, however, reiterate that Bronner can apply even in refusal to grant access cases involving infrastructure developed by a dominant undertaking for the purposes of its own business and owned by it. Here the ECJ cited judgments of 26 November 1998, Bronner, C-7/97, EU:C:1998:569, paragraph 41, and of 12 January 2023, Lietuvos geležinkeliai v Commission, C-42/21 P, EU:C:2023:12, paragraph 79 and the case-law cited).

[5]European Superleague Company, C-133/21, EU:C:2023:1011.

[6]Intel v Commission, C-413/14 P, EU:C:2017:632.

[7]2024 Draft Guidelines on the application of Article 102 of the TFEU to abusive exclusionary conduct by dominant undertakings.

[8] MEO - Serviços de Comunicações e Multimédia SA v Autoridade da Concorrência, C-525/16, EU:C:2018:270.

Public link

Please use the above public link if you want to share this noodl on another website.