Meta Platforms Inc & Ors v Dr Liza Lovdahl Gormsen

This is the single judgment of the Court. By a judgment dated 20^th February 2023 ([2023] CAT 10), the Competition Appeal Tribunal (“the CAT”) declined to certify a proposed collective claim pursuant to which the class representative, Dr Liza Lovdahl Gormsen ("the CR"), alleged that the applicants (“Meta”) had abused a dominant position by, in substance, extracting commercially valuable data from Facebook users without offering any payment for that material. 

In that judgment the CAT declined to certify the claim but instead ordered a stay and permitted the CR to submit a draft amended claim for consideration. A draft Collective Proceedings Amended Claim Form (“the CPACF”) was duly submitted. In a second judgment ([2024] CAT 11) the CAT approved of the amendments and certified the new (revised) claim, upon an opt-out, aggregate damages, basis. The CAT held that the amended claim was “clearly” arguable and was capable of being case managed to trial. The CAT refused an application by Meta for permission to appeal this ruling.

Meta now seeks permission to appeal from the Court of Appeal. Given the importance of the issue the Court deferred the application for permission to an oral hearing for full argument. On 7^th October 2024 this Court heard the application. At the end of the hearing the Court ruled that permission to appeal would be refused with written reasons to follow. These are those reasons.

The facts, as alleged, are taken from the CPACF and, as appropriate, from the expert report served by the CR and the summary thereof (respectively, “the Main Scott Morton Report” and “the Summary Scott Morton Report”). 

The CR pursues collective proceedings against Meta on behalf of all users of its Facebook social network service with a Facebook account who accessed the service at least once whilst in the UK between 14^th February 2016 and 6^th October 2023.

It is alleged that when signing up for an account Facebook required users to agree, upon a take it or leave it (“TIOLI”) basis, terms permitting Facebook to collect share and process users’ data both on and off the Facebook platform. The off-Facebook data consists of a user’s data relating to that person’s activities on websites and apps other than the Facebook social network platform. For example, it is alleged that if any website had a button to "like" or "share" it to a user’s Facebook profile, or a Facebook login option presenting users a choice to login into that third-party website via their Facebook account, then the terms permitted Facebook to track the users’ interactions with those third-party websites and apps even if the user chose not to click those Facebook buttons. This permitted Meta to combine data concerning users’ behaviour on third-party sites with information it already possessed and held about a user from that person's activity on the Facebook social network platform. It is said that this data collection exercise was vast and highly valuable to Meta being monetised in charges to advertisers.

It is also said that some of the off-Facebook data collected was highly sensitive and personal and included, for instance, details of users: "(i) researching their HIV medication on an NHS trust website; (ii) confidentially reporting sexual assault on 'the MET online'; and (iii) tracking their menstruation on mobile apps (including contraception use and mood". In addition, Meta collects data on users’ racial and ethnic origins, political and ideological beliefs, sexual orientation and preferences, intelligence, happiness, use of addictive substances, etc. The central thrust of the allegation is that this extensive, personal, off-Facebook data is extremely valuable to Meta in its commercial relations with advertisers. In the CPACF at paragraph [93] the CR identifies a book which refers to an internal, leaked, Facebook document which the author describes as demonstrating the following:

"Facebook's ability to use its unrivalled and highly intimate data stores 'to predict future behaviour', targeting individuals on the basis of how they will behave, purchase and think... for example, a Facebook service called 'loyalty prediction' is touted for its ability to… predict in individuals who are 'at risk' of shifting their brand allegiance… [to] trigger advertisers to intervene promptly, targeting aggressive messages to stabilise loyalty."

The CR claims that this off-Facebook data is not necessary for the operation of the Facebook service. Meta is only able to extract it, without any form of payment or compensation being made to users, as a result of its dominant position on the personal social network market. The CR contends that this amounts to an abuse in two related ways which "in practice amount to the same thing": first, imposition of an unfair trading condition (the TIOLI); and/or secondly, imposition of an unfair price (i.e. the off-Facebook data).

The focus of the application is upon two issues. First, whether the methodology chosen by the CR to establish an abuse of unfair pricing is arguable as a matter of law; and secondly, as to the logic of the way in which causation was pleaded. In respect of both of these, Meta argues that the CAT erred or at least arguably erred so that permission to appeal should be granted.

The CAT concluded that each pleaded abuse “pass[ed] muster” on a “self-standing” basis such that there was no necessity to consider whether they mutually reinforced one another (Judgment [25]). The CAT considered that in the round the pleaded case was “straightforward to manage” (Judgment [31]) and “clearly arguable” both in terms of the pleaded abuse and the claim for consequential loss and damage (Judgment [32]).

In relation to abuse the CAT considered the arguability of the TIOLI offer and the unfair pricing abuse separately.

In relation to TIOLI the CAT concluded (Judgment [25(1)]) that the TIOLI offer was arguably abusive because in a counterfactual, competitive, market Meta might have been compelled by the forces of effective competition to offer choice to users which could involve monetary compensation. The CAT did identify some complications with the argument which it set out in Judgment [25(1)(iii)] but as to which it commented that they “have barely been hinted at by Meta” which was “keeping [its] powder dry”. None of these prevented the CAT from concluding that the essential claim was arguable.

In relation to unfair pricing, the CAT (Judgment [25(2)]) held that this was arguably an abuse because the “shift” from a price based upon on-Facebook data to an offer based upon on and off-Facebook data was arguably unjustified and excessive and unfair. It described this as an “incremental” increase: Judgment [25(2) (vi)]. The CAT observed that this approach was not the way in which the legal test had typically been perceived but that the test, as formulated in case law, was not to be read as a statute and was open textured and flexible. In Judgment [25(2)(vii)] the CAT stated: “Depending on how the Facebook offering to users has evolved over time, one can see how an excessive/unfair price case might be derived from an incremental increase in price”.

The CAT considered causation, loss and damage in Judgment [27]-[30]. It asked, “whether a causal nexus between infringement and loss and damage has been established” (Judgment [27]). The CAT rejected the submission of Meta that the CR had failed to articulate “… a true connection between the abuses of dominance pleaded and the loss and damage flowing from those abuses” or “the basis for a collective bargain model for establishing the price that would be paid to the Users in the class had the abuses not occurred” (Judgment [29]). The CAT referred to the CPACF which averred that: “… [A]s a result of the abusively unfair bargain, or barter, made by Facebook with UK Users, the Proposed Class Members were not adequately compensated for the economic value of their Off-Facebook Data collected and monetised by Facebook, which resulted from the abusive conduct, and have therefore suffered pecuniary loss...”. Paragraphs [175] – [176] identified the relevant counterfactual:

“175.

In the counterfactual, Users would not have been subject to the unfair trading condition which made the provision of Facebook’s social network services conditional (or, by reason of Facebook’s use of choice architectures and the absence of any effective means of limiting such collection, effectively conditional) on the collection of Off-Facebook Data and/or would not have [been] subject to the unfairly high ‘price’ imposed by virtue of the collection of Off-Facebook Data and/or would not have received an unfairly low zero-price in return for their Off-Facebook Data.

176.

In the counterfactual, Users would thus have benefitted from a fair bargain in relation to the collection of their Off-Facebook Data. The value that would have accrued to Users in that fair bargain represents the loss they have suffered by reason of Facebook’s abuse (given that the PCR contends that Users currently receive nothing in return for the collection of their Off-Facebook Data). Accordingly, the same (or substantially the same) counterfactual and methodology for quantifying damages applies whether, as a matter of legal classification, the abuse is articulated as an unfair trading condition or an unfair price. Scott Morton 1 sets out a plausible or credible proposed methodology for establishing that Proposed Class Members have suffered loss and for estimating the loss suffered by the Proposed Class in the form of the value that would have accrued to them pursuant to a fair bargain in the aggregate…”

The CAT rejected the submission of Meta that properly understood the damages claim was based upon a disgorgement or gains-based measure of loss. The CAT accepted the categorisation of the claim as a conventional form of “negotiating damages” which was well established under English law (Judgment [29(2)]). The CR’s claim was based upon a counterfactual: had the (alleged) abuse of dominance not occurred, Meta would never have received the use of the off-Facebook data, either because the terms should never have been TIOLI or because the price (viewed in isolation of the TIOLI offer) was an unlawful one. Either way Meta obtained something it should not have done, and each member of the class has lost something. The loss was not physical but that was irrelevant since each member of the class had (arguably) sustained a loss in the form of negotiating damages and this sufficed to found a viable cause of action (Judgment [29(3)]). The damage was the difference between the payment users received for their off-Facebook data and the payment they would have received in the counterfactual. The CR advanced, by way of methodology, the “Nash bargaining model” which was set out in detail in the Main Scott Morton Report (Judgment [29(4)]). The CAT considered that this set out a perfectly arguable model for addressing damages.

Meta advances two proposed grounds of appeal, as follows:

“Ground 1: as a matter of law, United Brands, together with all other authority, requires an unfair price for the purpose of competition law to be established by reference to the reasonableness of the relationship between the entire economic value of the Facebook service and the entire price paid by Facebook users. Contrary to that legal requirement, the Class Representative ("CR") disregards the economic value provided by Facebook to users. Instead, the CR adopts an erroneously narrow approach whereby she seeks to establish an unfair price by reference to the relationship between increments of the economic value of Facebook that accrued in the period after Off-Facebook Data was collected and increments of the price, namely the Off-Facebook Data. The Tribunal erred by certifying the proceedings on that basis, which is a decision that is wholly unsupported by precedent or principle.

Ground 2: there is a complete lacuna in both the pleaded case and the methodology as to the causal link between the Alleged Unfair Trading Condition and alleged compensatory loss, such that the proceedings stand to be struck out. There is a leap from the allegation that the imposition of the ‘take-it or leave-it’ condition is abusive to a hypothetical counterfactual bargain pursuant to which it is said that Facebook would have made a payment to users. Therefore, when the Alleged Unfair Trading Condition is viewed in isolation, the CR has no sustainable cause of action and no blueprint to trial. In the Judgment, the Tribunal sought to address this by certifying the proceedings on the basis that the CR seeks negotiating damages of the kind articulated in Morris-Garner v. One Step (Support) Ltd [2018] UKSC 20 (“Morris-Garner”), which would mean that it would not be necessary for the CR to prove causation and compensatory loss in the way that is usually required in tort claims. However, in the PTA Ruling, the Tribunal changed course and stated that the proceedings had not been certified on the basis of negotiating damages of the kind in Morris-Garner. But the Tribunal did not then go on to address the complete absence of a nexus between the Alleged Unfair Trading Condition and compensatory loss. That issue was simply ignored. In those circumstances, the Tribunal erred in certifying the proceedings: whilst there is no pleaded case and no methodology as to the causal link between (a) the Alleged Unfair Trading Condition, and (b) compensatory loss, the proceedings stand to be struck out.”

Mr Tony Singla KC, for Meta, in his clear and helpful oral submissions, distilled the point to its essence. He accepted that the test of unfairness was that set out in Case C-20/7/76 United Brands v The Commission [1978] ECR 207 (United Brands) as adumbrated by the judgments of this Court in Competition and Markets Authority v Flynn Pharma Ltd [2020] EWCA Civ 339 (“Flynn Pharma”), and in London and South Eastern Railway Ltd v Gutmann [2022] EWCA Civ 1077 (“Gutmann”).

Mr Singla contended that the CR’s analysis of unfairness starts from 2014, which was the point in time when it is alleged that Meta first collected off-Facebook data. Value was measured by reference to the off-Facebook data and ignored the entire value of the service. This was evident from a number of paragraphs in the CPACF and in the CR’s expert evidence. These suggested that the CR’s legal and economic analysis ignored on-Facebook data and considered that only the “incremental” part of the service was being valued. Mr Singla resorted to first principles, which can be summarised as follows: Facebook was a very popular service which was greatly valued by users. Even if, to test the argument, Meta was dominant it was still entitled to extract a premium price for a premium service. The value in the hands of the user was to be assessed by the value of the entire service and not just some artificially constructed increment thereof. The value to a user was huge and was justified by the combined value of all Facebook data. According to case law it was clear that it was illegitimate artificially to truncate a service in order to suppress its real value and then to pretend that “value” was only that attributable to the truncated part.

In written submissions the point was put as follows:

“28.

In the premises, and whilst aspects of the CR’s position as regards the Alleged Unfair Price are muddled, what is clear is that the CR does not contend – as she is required to do as a matter of law – that there is no reasonable relationship between the entire economic value of the product (i.e., the Facebook service, including its value to users) and the entire alleged ‘price’. That critical issue does not feature in the pleading at all.

29.

The CR’s proposed methodology exacerbates this fundamental issue. In particular, the methodology does not propose to ascertain that there is no reasonable relationship between the entire economic value of the product and the entire price. Instead, the methodology proposes to examine the relationship between Off-Facebook Data and the economic value of increments of the Facebook service, with a focus on economic value generated since Off-Facebook Data is alleged to have been collected. In particular, it is said that:

“the right approach [under United Brands] is not to look at the aggregate value of Facebook to users, but to assess whether Facebook has struck a fair bargain as it has increased its data extraction over time”;

and

“[t]o conduct my assessment of whether Facebook creates economic value to users that justifies the price charged to users I consider it useful to decompose the value of Facebook into different components [including] the value users obtained specifically because Facebook was able to collect the Off-Facebook Data, i.e., which users would not have received in the absence of Off-Facebook Tracking”.

30.

In adopting that approach, the methodology narrowly and artificially focuses on whether any additional incremental economic value accrued to users in the period since Off-Facebook Data is alleged to have been collected, which necessarily excludes an assessment of the economic value of the Facebook service as a whole. But even more than this: within that already narrow and artificial slice of economic value, the CR’s expert does not propose to take into account all matters that are relevant to economic value having regard to the authorities cited above. For example, the CR’s expert ascribes no value at all to a range of features that Facebook introduced in the period since Off-Facebook Data has been collected and the CR’s expert also states that she will not take into account the value in “the network effect generated for users” in that period.”

(Emphasis in the original)

Mr Singla alighted upon the use by the CAT of the phrase “generally” in Judgment [25(2)(ii)] when it observed that:

“The United Brands test involves consideration of whether a price set by a dominant undertaking is first excessive and - if excessive - unfair. Generally speaking both excess and unfairness are assessed by considering the overall price charged for a service offered by a dominant undertaking in light of the totality of the service provided. Comparable are extremely important in assessing abuse. Incremental increases of price in respect of an existing service that does not change are not, even presumptively, unlawful. A dominant undertaking is entitled to increase price without improving the quality or quantity of its offering, although of course such matters will not be disregarded when considering whether an abuse does or does not exist.”

This was said to be erroneous because whilst the CAT accepted that the correct way in which to analyse value was by reference to the product or service as a whole, this was only “generally” so. In Mr Singla’s submission the expression “generally” wrongly permitted exceptions whereas the case law made clear that there were none, and this was relevant because the CR’s case fell within the impermissible exception, not the orthodox rule.

Mr O’Donoghue KC for the CR responded that Meta mischaracterised the CR’s case and the CAT’s analysis. Again, boiling the argument down to its essentials, he argued that the CR did attribute value to the entire Facebook service; it was not overlooked or ignored. He argued that in their pleading and in the expert evidence the CR, in effect, treated the Facebook service prior to 2014 (when it is alleged only on-Facebook data comprised the price or consideration tendered by a user) as non-abusive and, implicitly, involving the charging of a fair (non-abusive) price for the entire Facebook service. In effect up to 2014 the bargain struck between Meta and users involved a fair exchange of “value”.

He said that all that changed in 2014 when it is alleged Meta introduced the collection of off-Facebook data on a TIOLI basis. Upon the basis, ex hypothesi, that the conduct of Meta pre-2014 was lawful then it was informative and probative to consider whether the price subsequently demanded (i.e. for the off-Facebook data, as well as the on-Facebook data) was justified, or arose only because Meta was exercising its dominant market power to extract, unfairly, an ever increasing price from users for a service which largely remained static in terms of quality. In the CR’s case the increase in price was not explicable by reference to improvements in the Facebook service.

Mr O’Donoghue KC pointed out that an analysis of changes over time in the service and price offered by a dominant undertaking was an established technique generating probative evidence relevant to a determination of abuse: see for example Flynn Pharma (ibid) paragraph [92] citing the judgment of the CAT in Napp Pharmaceutical Holdings Ltd v DGFT [2002] CAT 1. None of this involved ignoring any part of the proper value to be attributed to the Facebook service.

Mr O’Donoghue KC referred the Court to the Main Scott Morton Report which set out various pieces of evidence which it was said established that there was a value to be attributed to the Facebook service as a whole but which also showed that what happened in terms of the price demanded after 2014 was not justified by any change in value. If this be true, then it amounted to strong evidence that the increase in the extraction of a user’s data was unjustified by any reference to value and was an abuse of dominance.

The CR endorsed a point raised in discussion between the parties and the Court to the effect that assumed, as a starting point, that the conduct of Meta in 2014 was fair and non-abusive was favourable to Meta. If it was fair and non-abusive in 2014 (a point Meta did not quibble with) then, as a matter of elementary logic, it was highly relevant to see what happened afterwards. The alternative was to take as a starting point that in 2014 Meta was, in substance, being substantially undercompensated for the value of the Facebook service it provided so was entitled, without there being any risk of abuse, to extract off-Facebook data in order, as it were, to make up the shortfall between value and price. But as to this Mr O’Donoghue KC pointed out that in the expert evidence of Mr Parker, for Meta, it is not argued that as of 2014 Meta was being undercompensated such that it was entitled to extract further consideration. It was argued that this being so, it is simply wrong for Meta to object to what was, on any view, a conventional way of analysing value i.e. by reference to change over time.

Given that the issue for this Court is whether to grant permission to appeal the question is whether, in treating the CR’s case as arguable and permitting the claim to be amended and certified, the CAT arguably erred in law. In forming a conclusion this Court will take into account that what the CAT considers to be arguable is a judgment it makes upon the basis of its evolving experience of how the concept of fairness operates in competition law, in particular in the tech sector. The hurdle for Meta will necessarily be relatively high.

It is the judgment of this Court that there is no basis upon which permission to appeal should be granted. The CAT did not even arguably err.

First, under Ground I Meta formulates its argument as an issue of law. On analysis the issue is not one of law but one of the characterisation of the evidence. The CAT was entitled to conclude that the CR’s theory of harm did not ignore the entire value of the Facebook service. Instead, it took a before and after analysis comparing and contrasting the position pre and post 2014. This was what the CAT meant when it used the expression “incremental”. It was using the expression in a temporal sense. Thus, in Judgment [25(2)(vii)], the CAT stated that an incremental price increase might give rise to an unfair price under conventional legal principle “[i]f (by way of example) the original price (i.e. the consensual provision of On-Facebook Data) can be said to be on the cusp of the excessive and the unfair [because] extraction of additional data (Off-Facebook Data) might be said, for that reason alone, itself to be excessive and unfair because of its purely incremental nature”. This encapsulates the temporal element behind the analysis and highlights that the hypothesis for analysis was that as of the date of the inclusion of off-Facebook data the bargain between Meta and its users was non-abusive i.e. struck a reasonable relationship in terms of value. This is why the CAT referred to the conduct being on the “cusp” of abuse. This approach does not ignore any component of the value of Facebook.

This is also clearly set out in the Summary of the Scott Morton Report which provides a useful overview of the more detailed analysis in the Main Report and sets out a variety of different pieces of evidence that, it is said by the CR, support the allegation of abuse. These include, but are not limited to, a review of the evolution of price (and its proxy, data) over time:

“26.

The following factors support the preliminary view that, for purposes of United Brands Limb 1, Facebook earned ‘excess’ profits from Off-Facebook Tracking and that the Off-Facebook Data was of significant commercial value:

a.

The introduction of Apple’s App Tracking Transparency initiative (“ATT”) serves as a useful ‘natural experiment’ to provide evidence that Off-Facebook Tracking significantly increased Facebook’s revenues and profits. ATT was introduced in mid-2021 and resulted in most users not opting in to being tracked. The financial impact of ATT on Facebook – which Facebook itself estimated immediately to be circa $10 billion for the year 2022 – therefore gives a proxy for the profits associated with Off-Facebook Tracking (albeit a conservative one since ATT only applies to iOS users and does not prevent tracking by Facebook on other Meta apps such as Instagram).

b.

The Report relies on Meta’s public statements as to the financial impact of ATT on Meta globally in 2022 as being $10 billion and, using that figure, conservatively estimates that Off-Facebook Tracking provided Facebook with profits of £2.99bn between 2016-2022.

c.

Using a “before” and “after” approach which compares average revenues per user pre- and post-2014 (when it is understood that Off-Facebook Tracking commenced or began to be monetised), it is possible to identify a material increase in Facebook’s average revenue per-user and profitability post-2014. Compared to growth in the related (but distinct) industry of search advertising, it is estimated that Facebook’s profits from Off-Facebook Data were £5.05bn between 2016-2022 (this figure is higher than the figure derived from the approach relying on ATT, which (as explained above) is conservative).

27.

To apply the second limb of United Brands, the Report identifies the following factors as indicative of unfairness:

a.

Facebook’s services are offered on the basis of “take-it-or-leave-it” terms and conditions. Moreover, the terms and conditions are (or were) opaque and unclear, and choice architectures made it difficult to change default privacy settings even if this was possible. In particular, the Report notes that: (i) Facebook for many years appeared to offer users no, or no effective, means to opt-out of Off-Facebook Tracking; (ii) Off-Facebook Tracking involves a form of “bundled consent”: if a user does not opt out they are deemed to have consented to Off-Facebook Tracking on all of the affected Meta properties and third party apps and websites; and (iii) the Off-Facebook Activity feature introduced by Meta in 2020 has been very ineffective in practice, since it proceeds on an opt-out basis.

b.

There are strong grounds to believe that an outcome under effective competition would have involved a more equitable bargain between users and Facebook. Faced with effective competition, it is highly likely that Facebook would have had to either refrain from Off-Facebook Tracking (or give users a transparent option to opt out) or provide some form of compensation to users. In the presence of reasonably effective competition, competition between platforms would strongly focus on users (who are more likely to “single-home” than advertisers). Rival platforms could offer greater privacy protection, reinvest the profits drawn on the advertiser-side to provide an improved user-facing service, and/or provide financial inducements to users. In any of these scenarios, Facebook would have strong incentives to maintain its advertising revenues by providing a value transfer to users (provided the value to Facebook from Off-Facebook Tracking exceeded the costs to users). The same argument can be reframed in a monopsony context; that is, under conditions of effective competition for users, the work those users carry out viewing adverts for Facebook would have resulted in greater compensation in the form of privacy protection, user interface, or financial transfers. The view that Off-Facebook Tracking (without some corresponding value transfer) would not be possible in a competitive environment is supported by the observation that, prior to achieving substantial market power, Facebook attempted to introduce Off-Facebook Data but had to reverse such measures due to user backlash.

c.

On a preliminary assessment, the great bulk of the value of Off-Facebook Tracking accrued to Facebook rather than to users. The data gathered by Facebook was of very significant commercial value, for the reasons given above. The fact that the vast majority of iOS users refused to consent to being tracked under ATT indicates that users in general dislike being tracked (including through Off-Facebook Tracking) which implies they incurred costs as a result of being forced to subject themselves to such tracking. Moreover, the major consumer-facing features added to the Facebook platform since the monetisation of Off-Facebook Data are not personal social networking functions and generally replicate the features of other platforms, such as TikTok. This again indicates that Facebook captured the majority of the economic value of the Off-Facebook Data for itself rather than sharing it with users by way of improvements to its platform.

d.

Facebook’s price is unfair when compared to appropriate comparators. The Report considers the following available and appropriate comparators:

i.

First, since Facebook operated for several years profitably without engaging in Off-Facebook Tracking, in a period in which it faced at least some competition, it is possible to use this period as a “before” comparator to show that Off-Facebook Tracking in the “after” period is unfair. Indeed, during this “before” period, Facebook sought to compete on quality by differentiating itself from rivals in particular by stating that it would not engage in user tracking.

ii.

Second, while other non-dominant social media platforms may have similar T&Cs to Facebook, they have not been able to strike such similarly skewed bargains or generate comparable profits. The only comparable platforms that gather similar data and are similarly profitable (e.g., Alphabet/Google and its properties such as YouTube) also possess significant market power and may well themselves be dominant in their distinct markets.

iii.

Third, other two-sided markets such as credit cards, food delivery services and ride sharing, have more competition which typically results in a fairer bargain with users. For example, credit card users frequently receive compensation in the form of bonus points or rewards for using their cards, while ride sharing and food delivery service users are frequently rewarded for their loyalty with discounts.

28.

The Report goes on to consider how the “economic value” of Facebook should be accommodated within the unfair pricing assessment given Facebook’s prior position before the Tribunal that the value users derive from Facebook exceeds the costs of Off-Facebook Tracking.

29.

The Report sets out the view that the right approach to assess the fairness of the price is not to look at the aggregate value of Facebook to users, but to assess whether Facebook has struck a fair bargain as it has increased its data extraction over time. Great care is needed with an argument that users revealed their preference by ‘accepting’ Off-Facebook Tracking. Given Facebook’s monopoly position in the relevant market, and the ‘must have’ nature of its service (due in particular to network effects), and the fact that users would be immediately cut off should they refuse, users have little choice but to agree to such terms: they are a condition of receiving the service. In this sense Facebook’s argument is similar to a dominant firm saying that its prices are not excessive because their customers were ‘happy’ to pay those prices.

30.

Furthermore, as noted above, much of the user-facing functionality pre-dates the introduction of Off-Facebook Tracking (which was not therefore necessary to deliver these benefits). Whilst Facebook may have initially provided distinctive value (aligning with the second type of cases in which a price can rise above costs, as identified by the Tribunal in Hydrocortisone), the consequence of it achieving dominance or a monopolist position means that it has “tipped” into the third type of case in Hydrocortisone, where the personal social network market is unjustifiably incontestable.

31.

The Report further observes that, first, based on the material available pre-disclosure, it does not appear that Off-Facebook Tracking can be justified by new services introduced contemporaneously with increased data extraction. Second, on current information, Off-Facebook Data does not appear to be directly used to improve the service to users. Third, Facebook cannot rely on past fixed cost investments to justify the introduction and monetisation of Off-Facebook Tracking as (i) Facebook was already significantly profitable prior to Off-Facebook Tracking, (ii) it could have still been significantly profitable if it had obtained genuine, informed consent for Off-Facebook Tracking from users and/or provided fair monetary compensation to users.”

All these categories of evidence are types of evidence that case law has indicated are capable of being relevant.

Secondly, and in any event, this is an area of law which is, par excellence, new and evolving. The use of data as a proxy for monetary payment is a rapidly increasing phenomenon of modern digital life and as such it is generating a range of new legal issues. As was pointed in out in Flynn Pharma (ibid) the law is flexible and there is no single test which must, necessarily, be applied to determine abuse. In Gutmann (ibid) paragraph [102] the Court referred to established case law which made clear that an abuse could be assessed by reference to the counterfactual where there was no dominance, but it could also be determined by looking at the disputed term itself to evaluate whether in and of itself it was unreasonable or disproportionate. In the judgment of this Court, although the factual scenario might be relatively new, the framework for analysis is not. The CR’s case has a surface logic to it: if there was fair value passing between Meta and users in 2014 upon the basis of payment by way of on-Facebook data alone, then the subsequent extraction of incremental high value data from users for no or no sufficient recompense might be unfair and attributable to the market power of Meta. It might amount to an abuse. Meta will undoubtedly advance a robust case by way of defence which the CR will have to meet. But that is for trial as the CAT acknowledged (Judgment footnote [24]). Mr Singla KC argued that whilst the United Brands test, as developed in case law, was flexible there were nonetheless limits. The law was flexible but still operated within boundaries and limits. This is no doubt correct. However, the CAT has explained why in its view the articulation of the law by the CR was arguable and within those limits. The CAT was perfectly entitled to form this view. Indeed, there is nothing in the approach being mooted by the CR which is outwith normal methodologies. But even if there is novelty in the issues arising it must be for the CAT to delve into such novelties to form a view, and it is not for this Court to seek to cut off such analysis before it has even been embarked upon.

Thirdly, the CR points out that in relation to Facebook specifically an analogous scenario has been held by the Federal Supreme Court in Germany to amount to an exploitative abuse in a judgment which the CAT, the High Court and the Court of Appeal have all received with some measure of approval: see the analysis in Gutmann (ibid) at paragraphs [97] – [102]. The CR summarised the position as follows in written submissions:

“The CR considers that she has a strong case on dominance and abuse. For example, in Germany, the Bundeskartellamt found that Facebook had committed an exploitative abuse by the same conduct (i.e. making user access to Facebook’s social media service conditional on Facebook being allowed to collect and store personal and device-related data generated from visiting third-party web pages/apps and from use of other Meta services (e.g. WhatsApp, Instagram, etc), and to connect these data with On-Facebook Data without user consent): CPACF §151. Per this Court in London and South Eastern Railway Ltd v Gutmann [2022] E.C.C. 26 (“Gutmann LSER CA”) §97, there was “nothing especially startling” about that analysis: indeed, the CA observed both that the CAT below had “strongly endorsed” the Bundeskartellamt decision; and that the High Court in Preventx Limited v Royal Mail Group Ltd [2020] EWHC 2276 (Ch) had treated it as “illustrative of the broad range of potentially abusive trading conditions that a dominant undertaking might impose”. The CR also identifies a long list of further proceedings and regulatory investigations about Meta’s relevant conduct at CPACF §103.”

Accepting of course that the judgment of the German Federal Supreme Court does not bind the CAT in the future, it is at the least a powerful indicator that the point is arguable.

Fourthly, Meta argues that the CR has not met the serious objections that it has raised. In relation to the observation of the CAT that pleadings should not unduly pre-empt or anticipate defence arguments, the present case highlights why that is sound guidance, and why Meta’s arguments will need to be considered in the fullness of time, but not now. It is evident that a CR will, at the certification stage, need to strike a balance between providing a clear articulation of its case and an equally clear methodology and road map, so that the CAT can ensure that the proposed case meets the relevant criteria, on the one hand, and avoid excessive detail and recitation of evidence, on the other hand. The CR has advanced a detailed (overly detailed – as the CAT would have it) evidential case. Meta has responded vigorously but not yet fully or with supporting evidence or disclosure. The CR has responded to these counterarguments. This Court has been the recipient of sophisticated argument, albeit cast in relatively emphatic and black and white terms, which begs a host of questions about the facts and the evidence. At the end of the day, as it was put to us by Mr O’Donoghue, the “battle lines” have been drawn. But the fact that there are battle lines does not mean that the central case is not arguable.

Finally, Meta, especially in its written submissions, argued that the CAT was critical of the CR’s pleading and evidence and in the articulation of its case. Meta prayed these criticisms in aid of its argument that the CAT, inconsistently, then proceeded to find that the CR’s case was sufficiently clear in terms of its articulation of legal principle and in the ability of the CAT to case manage the dispute to trial. Meta highlighted the CAT’s criticisms: the CR’s pleading as to abuse was “difficult to follow” due to “the absence of a clearly articulated and self-standing theory of harm”; the pleaded case “makes a number of averments the significance of which is not completely clear”; the plea as to the alleged unfair trading condition was “open-ended and mixes the pleading of fact (which is essential) and evidence (which should be omitted)”; the CR’s articulation of the alleged unfair price as an unfairly high and an unfairly low price was “unhelpful in terms of understanding precisely what the PCR is saying”; and causation and loss was “less clear than it might be”. The CAT did, indeed, identify various concerns, including about the clarity of the manner in which certain legal and evidential issues had been framed. However, contrary to the submission of Meta, none of this undermines the CAT’s central conclusion which was that, nonetheless, the nub or pith of the CR’s legal case was identifiable and arguable and that the CAT could envision the methodology that would take the case to trial. Standing back the approach of the CAT to engage in some “nudge” case management by highlighting, informally, points it wanted clarified, was pragmatic and sensible. No doubt the CR will take these various hints. It cannot though be right to say that, on the one hand, the CR should be told by the CAT not to anticipate every argument the defendant will come up with, but then for the CAT to accede to an application by the defendants for a summary dismissal upon the basis that the CR had not met the defendants’ case which, as set out above, is still in the throes of evolution. The CAT did not fall into this trap.

For these reasons the conclusion of the CAT that the CR’s case was arguable was not wrong in law, whether arguably or otherwise.

Under the proposed second ground of appeal, set out in paragraph [16] above, Meta complains, as it did before the CAT, that there is a lacuna as to the causal link between the allegedly abusive TIOLI offer and alleged compensatory loss. The CAT disagreed and in Judgment [29(3)] explained in straightforward terms why causation had been arguably pleaded. Expressed shortly, but for the alleged abuse users would have possessed the right to negotiate some form of compensation for the use of their off-Facebook data. The TIOLI pricing offer prevented this. It is said that the CAT adopted a different analysis in its ruling on permission to appeal ([2024] CAT 30). That is not a fair reading of the ruling. There, the CAT pithily summarised its earlier ruling and emphasised that it would need carefully to case manage the case on causation as it proceeded to trial. The proposed Ground does not raise anything arguable.

In oral argument Mr Singla KC refined the point. He argued that the TIOLI offer allegation and unfair price allegation were framed as complementary and alternatives i.e. they were on an “and/or” basis. In relation to the latter (the “or”) if the CR failed to establish that the provision of the Facebook service in its entirety was unjustified by reference to the price paid (comprising both on and off-Facebook data), then the unfair pricing allegation would fail. But in such a scenario there could be no logic or utility in the CR being able to argue that the TIOLI offer was independently abusive. He emphasised that the challenge to the TIOLI clause was not being pursued on some altruistic consumer welfare basis i.e. to establish a point of principle. On the contrary it was pursued in the context of a collective aggregate damages claim, where the only interest of the class was in financial compensation. It necessarily followed that if the price paid was fair and reasonable it would not be abusive and it was redundant to then also argue that the TIOLI clause was an abuse. It could not be an abuse to stipulate that if the customer refused to pay the (ex hypothesi) fair price it could not receive the service.

The difficulty at this stage is that it is not the real thrust of the CR’s case or the way in which it has been understood by the CAT. That treats the TIOLI obligation as an integral part of the allegation that the price is unfair. It is part of the mechanism used by Meta to impose the unfair price. The CR has described the price and TIOLI issues as two sides of the same coin. True it is that the CR casts the TIOLI point, forensically, in “and/or” terms and to this extent there is the possibility that it might ultimately standalone against the backdrop of a finding that the “price” was fair. But, looked at in the round, this is more theoretical than real and the argument as it proceeded before the CAT was not upon the basis of how one ground of challenge would play out if the other failed. Mr O’Donoghue KC said that if the compensatory claim failed then the CR might seek to amend its claim to one of disgorgement of profits. The CAT was unimpressed by such a prospect. Be that as it may, the way in which TIOLI is addressed in the analysis and evidence to date is as part of the overall claim for monetary compensation. The CAT was entitled to treat the issue in this way. To the extent that the issue becomes live upon a stand-alone basis in a context where the unfair price claim has failed, that would be the point in time for the CAT to grapple with the implications. The CAT was right to refuse permission to appeal and we can see no basis upon which we should take a different stance.

These are the reasons why the application for permission to appeal was dismissed. For the avoidance of any doubt nothing said by the Court in this judgment expresses a decided view on any issue arising, all of which are for trial.