Competition law & privacy. This may not be a decision on the merits, but it is already rich in lessons. In particular, this case is the first time that the Autorité de la concurrence, i.e. the French Competition Authority (hereinafter FCA), has dealt with a dispute on the borderline between competition law and data protection law.

Context. This litigation follows a 22 June 2020 conference for app developers at which Apple Inc. (hereinafter Apple) announced that it would update its iOS operating system to enhance the privacy of its customers. Specifically, Apple planned to implement, as of September 2020, a feature called App Tracking Transparency (hereinafter ATT prompt) involving the display of a window for users to consent to being tracked on various websites or mobile apps for advertising purposes, prior to any use of the Identifier for Advertisers (hereinafter IDFA), i.e. an Apple unique identifier[1].

The FCA’s referral. It is in this context that the associations Interactive Advertising Bureau France (hereinafter IAB), Mobile Marketing Association France (hereinafter MMA), Union des entreprises de Conseil et d’Achat Media (hereinafter UDECAM), and Syndicat des Régies Internet (SRI) filed a complaint to the FCA. These associations consider that the implementation of this system and the obligation imposed on app developers to use it to access the IDFA constitute an abuse of dominant position. According to the applicants, these practices infringe in particular the prohibition on imposing unfair trading conditions laid downin Article 102(a) of the Treaty on the Functioning of the European Union (hereinafter TFEU) and the prohibitionon imposing a supplementary obligation laid down in Article 102(d) TFEU[2].

This is why, in addition to the referral on the merits, they have requested interim measures to be taken against Apple, namely firstly, the suspension of the ATT prompt until a decision on the merits has been adopted or an acceptable solution has been found following negotiations between the industry and Apple; secondly, the obligation for Apple to enter into a dialogue with industry stakeholders to find an acceptable solution for obtaining permission for user tracking that is not only complies with European privacy legislation but also less intrusive to users and harmful to competition in the relevant markets[3].

The applicability of EU law. The FCA’s review involves a number of steps, starting with the applicability of EU law. As a reminder, National Competition Authorities (hereinafter NCAs) are obliged to apply EU law when the practices at issue are capable of appreciably affecting trade between Member States. At this stage of the investigation, there is little doubt that it applies: firstly, the practices complained of are global in scope[4]; secondly, they are likely to make it more difficult to enter the market for the distribution of apps on iOS[5]; and thirdly, Apple’s turnover far exceeds the de minimisthreshold set out in the European Commission’s (hereinafter Commission) notice on the effect on trade concept[6]. The FCA then examined Apple’s position on the relevant markets (I) and the alleged abuses (II).

 

I. Apple’s position in the relevant markets

Analysis of the situation. Having defined the relevant markets (A), the FCA concluded that Apple had a monopoly position in this case (B).

 

A. The relevant markets

The relevant markets defined by the FCA.As it stands, the FCA has delineated three relevant markets: the market for the distribution of apps on iOS – in which Apple allegedly carries out the alleged practices – and the markets for install ads on iOS and mobile app installation attribution services – in which their effects occur[7].

The market for the distribution of mobile apps on iOS. As regards, first, the market for the distribution of mobile apps on iOS, the FCA does not exclude at this stage the delimitation of such a market due, in particular, to the Commission’s previous decision on Google Android. In that case, the Brussels authority considered that Google and Apple were not competitors, as they did not operate on the same market. Google was in the market for marketable operating systems for smartphones, while Apple did not license its iOS operating system because of its proprietary business model[8]. In this respect, it is tempting to see this decision as part of the trend for NCAs to refer to previous decisional practice to delimit the market, without necessarily seeking to differentiate the players and practices involved, in order to overcome the difficulties associated with their definition[9]. However, such a methodological shortcut is more easily justified here insofar as the present decision has not been rendered on the merits. Moreover, the FCA does not simply cite the Google Android case as a precedent; it also points out that Apple’s App Store is neither substitutable from the point of view of users nor from that of advertisers. Firstly, as regards users, it seems that they are not in a position to switch ecosystems in the event of an increase in the price of paid apps on iOS, or a decrease in their quality, because of their attachment to Apple branded products[10]. Secondly, for app developers, the Apple Store is the only way to reach iOS users, who « are a particularly strategic and attractive ‘‘target’’, as they generate more revenue than users on Android »[11].

Secondary markets. Moreover, these passages dedicated to the market for the distribution of mobile apps on iOShave also given rise to developments relating to « secondary markets »[12]– or « after-sale markets », or even « aftermarkets » – which the authors of the Crémer report had suggested to apply to digital ecosystems[13]. In this case, Apple considers that the competitive pressure exerted on the primary market for the sale of mobile handsets is an obstacle to the characterisation of its dominant position on the secondary market for the distribution of apps[14]. However, Apple’s argument does not convince the FCA. The FCA recalls, for all intents and purposes, that the exclusion of dominance in the secondary market implies the fulfilment of a number of criteria identified by the Commission in Pelikan/Kyocera:« (i) consumers canmake an informed choice including the prices they will face in the future; (ii) consumers are likelyto make such a choice accordingly; (iii) a sufficient numberof consumers would adapt their purchasing behaviour at the level of the primary market in the event of a price increase or a quality deterioration on the secondary market; (iv) this adaptation will take place within a reasonable timeframe »[15]. At this stage, the FCA considers that these conditions are not met for several reasons: firstly, it is the price of the mobile device that is decisive in the choice of the latter, but not that of the apps. Secondly, as Apple has developed a real ecosystem based on both hardware and software, it is illusory to think that users are likely to change their mobile device because of the increase in price or the decrease in quality of apps in the App Store, as this would ultimately mean sacrificing their investment in the Apple ecosystem[16]. On this last point, the FCA thus seems to consider that the considerable switching costs borne by users within the Apple ecosystem cannot exclude Apple’s dominant position in this secondary market[17]. As a number of players have developed real ecosystems, like the Big Tech, it will be interesting to see how often the aftermarket analysis will be used in future litigation.

The geographical scope of this market.As regards the geographical scope of this market, the FCA tells us that it is potentially global[18].

The market for mobile app install ads on iOS. Next, with regard to the market for mobile app install ads on iOS, it is interesting to note that this market has never been defined before, either in the decisional practice of the authorities or in the case law[19]. However, the delimitation of such a market is not really surprising. Indeed, the segmentation of the online advertising market is due, on the one hand, to the tendency to differentiate products and services in digital markets and, on the other hand, to the fact that the functionalities of the products or services in question play a predominant role in the analysis of substitutability. In this case, this type of advertisement would have its own characteristics, such as the direct action of the user on the advertisement, i.e. the downloading of the app which is the subject of the ad[20].

From a geographical point of view, the limits of this market would seem to be limited to the national territory[21].

The market for attribution services for mobile app installations on iOS. Finally, with regard to the market for attribution services for mobile app installations on iOS[22], it should be noted that, unlike the previous markets, these allocation services are not specific to a particular operating system. Indeed, almost all of them are available on both iOS and Android[23].

Furthermore, the geographic scope of this market would be at least Europe-wide[24].

 

B. Apple’s monopoly situation

Apple’s monopoly on the market for the distribution of mobile apps on iOS.At this stage of the investigation, it appears that Apple is in a monopoly situation on the market where the allegedly abusive practices occurred – i.e. the distribution of mobile apps on iOS[25]–; the Apple Store being the only distribution channel for mobile apps on iOS devices. The very evidence of this solution reminds us, if any were needed, of the importance of the definition of the relevant market(s) in the characterisation of the dominant position.

 

II. The abusive practices of which Apple is accused

The freedom of dominant undertakings and large digital platforms. As a preliminary point, the FCA recalls – and rightly so – that the freedom of trade and industry enjoyed by undertakings, including those in a dominant position, allows them to shape their business model, products and services as they see fit, provided in particular that they do not contravene competition rules. Although useful, the originality of the decision on this point lies elsewhere. Indeed, what makes it unique in this respect is the application, for the first time, of these considerations to so-called « structuring » platforms, which also enjoy the status of « gatekeepers ». The FCA’s use of this terminology is not insignificant; in this way, it is in line with a previous contribution[26], but also, and above all, with the Commission’s proposal for a regulation[27].

With these clarifications, the FCA then examines whether Apple imposed unfair trading conditions (A) or a supplementary obligation (B) as alleged by the complainants.

 

A. The imposition of unfair trading conditions

Comparison with the Facebook case. This is undoubtedly the most interesting aspect of this decision. As the abuse alleged against Apple is similar to the one highlighted by the Federal Cartel Office (hereinafter FCO) in the German Facebookcase, it may be tempting to draw a parallel between the two decisions. However, the two cases actually diverge on a number of points.

Difference in applicable law. Firstly, in the Facebookcase, the FCO relied – wrongly for some people[28]– on the specificities of German competition law, whereas in this case it is EU law that is applied by the FCA. Moreover, it is fortunate that the concepts forged long ago by European decisional practice and case law still find effective application in the digital environment[29]. In this sense, the FCA refers for example to the BRT/SABAM, United Brandsor Tetra Pakdecisionson the standard of proof applicable to the assessment of the imposition of unfair trading conditions[30].

Difference in the way privacy issues are dealt with.  Secondly, unlike the FCO, the FCA has not itself applied and interpreted data protection law. The present decision is the result of the collaboration between the FCA and the French data protection authority, the Commission Nationale de l’Informatique et des Libertés (hereinafter CNIL) – the latter having been asked for its opinion by the FCA on the personal data protection aspects of the practices at issue[31].

Difference in the assumptions addressed by these cases. Finally, and most importantly, the assumptions considered in these different cases are not the same: to put it simply, in the Facebook case, privacy is an element in determining the abuse; whereas this case is more in line with the assumption that privacy is an element in justifying the abuse. Indeed, privacy enhancement seems to be the reason why Apple implements its ATT prompt. In any event, this is what emerges from the FCA’s assessment, both of the objective pursued by Apple and of the necessary and proportionate nature of its measure. 

Apple’s objective of protecting personal data. Firstly, with regard to Apple’s data protection objective, the FCA considers that the ATT prompt is part of Apple’s long-standing commercial policy and, moreover, responds to a growing consumer demand in this area. In so doing, it concludes that this strategy is not anticompetitive in itself and is a legitimate exercise of Apple’s commercial policy[32].

Necessity and proportionality of the ATT prompt. Secondly, with regard to the necessity and proportionality of the ATT prompt, the complainants tried to demonstrate that this was not the case, even going so far as to argue that this device would have « an alarming tone which likely to scare the user »[33]. In any case, their arguments did not convince the FCA, for whom the ATT promptappears neither excessive nor disproportionate with regard to the interests of consumers who wish to have control over their personal data for advertising purposes, nor with regard to those of app developers, even though the vast majority of them finance their services through personalised advertising[34].

Furthermore, the FCA considers that Apple acted in good faith in postponing the implementation of the ATT prompt to a later date in order to give time to app developers to comply with this new technical way of collecting consent[35].

Secondly, although the complainants argue that the collection of consent via this mechanism is not compliant with the GDPR, the FCA points out that it is not, according to Apple, a substitute for the requirements of the regulation. It is a complementary measure or, to put it another way, an additional protection measure which, in any case, « does not appear to be inconsistent with the letter or spirit of the GDPR »[36].

Also, the fact that part of the window text is fixed and cannot be changed by developers is not a problem from the FCA’s perspective. Indeed, it is always possible for them to complete it by displaying an additional window: either before, in order to inform users about the data collected by this means; or after, in order to make them change their mind following their refusal for example[37]. The CNIL concludes that this device would allow developers to obtain informed consent and thus comply with regulatory requirements in this area[38]. Moreover, the FCA adds that such a standardisation of ATT prompt is likely to facilitate consumer choice[39].

Finally, with regard to Apple’s claim that it did not apply the ATT prompt to its own advertising service Apple Search Ads, it appears that Apple does not have to do so since it does not use any individual tracking technique for Internet users[40]. At this stage of the investigation, it does not appear from the evidence in the file that Apple is imposing, via its ATT prompt, more stringent treatment on app developers than it is imposing on itself[41]. However, the referral on the merits will make it possible to ensure that Apple is not implementing a « self-preferencing » practice[42].

Some additional thoughts. In conclusion on this point, this decision appears satisfactory in several respects: first, this decision appears to be well-founded, as Apple’s practice seems to ultimately improve the protection of its users’ privacy. Secondly – and this is partly in line with the above – as this decision is the result of a collaboration between the FCA and the CNIL, this only reinforces the finding of its legitimacy in terms of both competition and data protection. Indeed, after having already been tested in 2014 in the context of interim measures imposed on GDF Suez[43], this decision only confirms that collaboration between the FCA and the CNIL is not only desirable, but also possible in digital markets[44] where overlaps between the two bodies will inevitably increase in the future.

While this decision is instructive, it does not answer a number of difficulties. For example, what happens when privacy protection is not the real reason for the practice? One example is Google, which has announced its decision to phase out third-party cookies from its Chrome browser and replace them with a set of proposals that are part of the “Privacy Sandbox” initiative. This practice is currently being investigated by the Competition and Markets Authority[45](hereinafter CMA) and the Commission[46], but also by some fifteen US states, including Texas[47]. Although Google claims that its objective is to preserve the privacy of its users, these practices could ultimately favour the advertising activity of its « walled garden », i.e. its ecosystem, which is likely, paradoxically, to infringe users’ privacy more than third-party cookies. However, Google is not an isolated case, far from it. Indeed, although Apple has said that its ATT prompt was not intended to boost its advertising business on its devices, the Financial Times has revealed that Apple is in the process of expanding its advertising business[48]. There is no doubt that in this case, the investigation of the case on the merits will, if necessary, allow the FCA to correct its position and sanction Apple for what looks like « self-preferencing ». But, generally speaking, there is a real risk that the invocation of privacy concerns could undermine the effectiveness of competition rules[49]. While there is no remedy for the hypocrisy of Big Tech, the solution may lie in limiting the internal use of data by these platforms in the field of advertising[50]. Moreover, among the obligations incumbent on gatekeepers, Article 5 (a) of the Digital Markets Act (hereinafter DMA), which is actually aimed at the case of Facebook, provides for an obligation in this regard[51].

 

B. The imposition of a supplementary obligation

The non-exhaustive list of abuses under Article 102 TFEU. In this respect, it is important to recall, as the FCA does, that even if Apple’s imposition of a supplementary obligation on app developers does not correspond to the example mentioned in Article 102(d) TFEU[52], the abuses listed by this provision not being exhaustive, this practice may in any case be prohibited under Article 102 TFEU[53].

The absence of a supplementary obligation imposed on developers by Apple. With these clarifications, the FCA considers that the ATT prompt does not impose a supplementary obligationon developers since it is, in the end, only a « new way of implementing a pre-existing mechanism »[54]. Moreover, this ATT prompt would, according to the FCA, be linked to the main service, i.e. the distribution of apps from the App Store. According to the FCA’s reasoning, the distribution of apps implies the provision of a certain number of ancillaryservices to consumers so that they can download them with confidence. The documents in the file show that the protection of privacy is one of these[55]. Consequently, this abuse cannot be characterised here.

On interim measures.As none of the abuses alleged by these associations against Apple had been proven at this stage of the investigation, the FCA did not consider it necessary to examine whether the conditions of ArticleL. 464-1 paragraph 2 of the code de commerce, i.e. the French Commercial Code(hereinafter C. com) necessary for the pronouncement of interim measures had been met in this case[56].

 

Conclusion. Obviously, the result reached by the FCA in this decision does not prejudge the outcome of the decision on the merits. This is fortunate because, as noted above, there will be new developments in this litigation. In the latest news, several German companies have filed a similar complaint to FCO[57]. Let’s hope that these cases will determine the extent to which privacy can justify an abuse of dominance. Furthermore, the Facebook case was referred to the Court of Justice of the European Union[58](hereinafter CJEU), so it may be some time before we know how far privacy can be an element in determining abuse of dominance. While the FCA has announced that digital and the DMA discussions in particular will be high on Adlc’s agenda for 2021[59], this year’s meeting will already be marked by a focus on privacy, and it promises to be an exciting one!

 

Vincent GIOVANNINI

 

 

[1]FCA, decision 21-D-07 of 17 March 2021 regarding a request for interim measures submitted by the associations Interactive Advertising Bureau France, Mobile Marketing Association France, Union Des Entreprises de Conseil et Achat Media, and Syndicat des Régies Internetin the sector of advertising on mobile apps on iOS, paras 18-30. For more details about the IDFA, see ibid, paras 5-17.

[2]Ibid, paras 2-3.

[3]Ibid, para 95.

[4]Ibid, para 98.

[5]Ibid.

[6]Ibid, para 99.

[7]Ibid, para 101.

[8]Google Android(Case AT.40099) Commission Decision C(2018) 4761 [2018], paras 238-267.

[9]Jean-Christophe. Roda, ‘Le marché’, in Jean-Baptiste Racine (ed), Le droit économique au XXIe siècle – Notions et enjeux(LGDJ 2020), 510.

[10]FCA (n 1), para 111.

[11]Ibid, para 112.

[12]Luc Desaunettes-Barbero and Etienne Thomas, Droit matériel européen des abus de position dominante : Textes et commentaires(Bruylant 2019), 63 : « A secondary market is a market comprising a product that is complementary to a product sold upstream on a primary market ». For more details, see Miguel Sousa Ferro, Market Definition in EU Competition Law(Edward Elgar 2019), 170-180.

[13]Jacques Crémer, Yves-Alexandre de Montjoye and Heike Schweitzer, Competition policy for the digital era (2019), 48.

[14]FCA (n 1), para 113.

[15]Ibid, para 114 (emphasis added).

[16]Ibid, para 115.

[17]Ibid, para 116.

[18]Ibid, para 117.

[19]Ibid, para 119.

[20]Ibid, para 121.

[21]Ibid, para 124.

[22]In short, attribution services help advertisers to measure the effectiveness of their advertising campaigns. See ibid, para 126.

[23]Ibid, para 127.

[24]Ibid, para 128.

[25]Ibid, para 130.

[26]The Autorité de la concurrence’s contribution to the debate on competition policy and digital challenges (19 February 2020).

[27]Commission, ‘Proposal for a regulation of the European Parliament and of the Council on contestable and fair markets in the digital sector (Digital Markets Act)’ COM/2020/842 final.

[28]Wouter Wils, ‘The obligation for the competition authorities of the EU Member States to apply EU antitrust law and the Facebook decision of the Bundeskartellamt’, [2019] 3 Concurrences, 58-66.

[29]Fayrouze Masmi-Dazi, ‘Ciblage publicitaire iOS14 : l’Autorité de la concurrence sur le front de la régulation des big techs’ (2021) Dalloz actualité.

[30]FCA (n 1), para 143.

[31]Ibid, paras 54-64.

[32]Ibid, para 147.

[33]Ibid, para 151.

[34]Ibid, para 153.

[35]Ibid, para 154.

[36]Ibid, para 157.

[37]Ibid, para 158.

[38]Ibid.

[39]Ibid, para 159.

[40]Ibid, para 161.

[41]Ibid, para 162.

[42]Ibid, para 163.

[43]FCA, decision 14-MC-02 of 9 September 2014 regarding a request for interim measures submitted by the company Direct Energie in the sectors of gas and electricity.

[44]Masmi-Dazi (n 29).

[45]CMA, ‘CMA to investigate Google’s ‘Privacy Sandbox’ browser changes’ (8 January 2021).

[46]‘E-000274/2021 Answer given by Executive Vice-President Vestager on behalf of the European Commission’ (23 April 2021) <https://www.europarl.europa.eu/doceo/document/E-9-2021-000274-ASW_EN.pdf>accessed 27 April 2021.

[47]The State of Texas and al. v Google LLC, Case 4:20-CV-957-SDJ, 15 March 2021 (amended complaint).

[48]Hannah Murphy et Patrick McGee, ‘Apple to boost ads business as iPhone changes hurt Facebook’ Financial Times(San Francisco, 22 April 2021), <https://www.ft.com/content/5527ddd1-77a8-4cd0-82fd-4568be5da80f> accessed 27 April 2021. While Apple already sells search ads for its App Store, it plans to add a second ad slot, in the “suggested” apps section of its App Store search page.

[49]Damien Geradin, ‘GDPR Myopia: How a Well-Intended Regulation ended up Favoring Google in Ad Tech’, [2020] <https://poseidon01.ssrn.com/delivery.php?ID=082112002094023099124016095096017117099072067012079043078015093123092109102064006011021023022028055040113122030122109064005127107025064010065101016066082098115023054001094106124120082027112075087068121125121122017118119094026082093095096027102024009&EXT=pdf&INDEX=TRUE> accessed 27 April 2021, 36.

[50]Ibid, 23.

[51]COM/2020/842 final.

[52]This is traditionally associated with tying and bundling practices. See Catherine Prieto et David Bosco, Droit européen de la concurrence – Ententes et abus de position dominante(Bruylant 2013), 1002-1003.

[53]FCA (n 1), para 167.

[54]Ibid, para 169.

[55]Ibid, para 170.

[56]Ibid, paras 173-175.

[57]Javier Espinoza, ‘German groups file Apple antitrust complaint as it makes privacy changes’ Financial Times(Brussels, 26 April 2021), <https://www.ft.com/content/0a48d9aa-244b-4945-b2a0-01c68683544a> accessed 27 April 2021.

[58]Oberlandesgericht Düsseldorf, 24 March 2021, Kart 2/19 (V), available at https://www.justiz.nrw.de/nrwe/olgs/duesseldorf/j2021/Kart_2_19_V_Beschluss_20210324.html?s=09.

[59]FCA, ‘Après une activité très soutenue en 2020, l’Autorité de la concurrence annonce ses priorités pour 2021, qui seront centrées sur l’économie numérique’ (23 December 2020) https://www.autoritedelaconcurrence.fr/fr/communiques-de-presse/apres-une-activite-tres-soutenue-en-2020-lautorite-de-la-concurrence-annonce accessed 27 April 2021.