Debating the EU “pay or consent” decision against Meta

In a new webinar convened by Oles Andriychuk (University of Exeter), I debated two competition law scholars (Simonetta Vezzoso, University of Trento, and Marco Botta, European University Institute) on the topic of the EU Commission’s DMA enforcement decision against Meta (“pay or consent”). I already covered that decision on EUTechReg in a text on how this will impact Meta’s business and in my notes from a podcast with Eric Seufert. The debate was lively and I think it showed well our contrasting perspectives on both the Meta decision and the Digital Markets Act as a whole.

You can watch the webinar on YouTube, but I’m also including below some notes on the content of our debate.

Conference alert: On September 29-30, I’ll be participating as a panel moderator in RAID - Regulation of AI, Internet and Data conference, in person, in Brussels (even I visit that place from time to time). I believe that registrations are still open.

My analysis centered on the practical consequences of the Commission’s decision, my skepticism about its goals, and the lack of clarity for compliance.

The decision harms Meta’s business users (advertisers). My most significant concern is the direct, negative impact this decision will have on business users, particularly small and medium-sized ones. The Commission, in its own decision, acknowledges that Meta Ads is a great tool for advertisers because of its unparalleled scale and personalization capabilities. For a niche artisanal manufacturer, for instance, the ability to target specific customers isn’t just a benefit, it’s existential to their business model. They don’t have the budget for broad brand advertising like L’Oreal. (Simonetta’s point about a hairdresser in Rome is not a good example: such advertising is more likely to be vital for businesses that are not local; businesses which seek a small number of customers across the entire EU).

Yet, the entire thrust of the Commission’s action seems designed to make this highly effective tool worse for its business users. The goal is to create contestability by thwarting the gatekeeper, but the immediate victims are the European businesses who rely on the service. We are seeing business users being harmed in the name of contestability with only hypothetical promises that it will somehow work out in the end. I worry this will become another hotel-like scandal, where the DMA’s enforcement action ends up hurting the very constituency it was meant to empower. (I’m referring to the complaints from hotels about Google’s DMA compliance measures that ended preferencing intermediaries like Booking.com and harming the end business users like hotels).

A lack of a DMA vision from the Commission. I’ve struggled to see a clear, coherent vision from the Commission of what they are trying to achieve. The Commission argued that Meta’s data scale creates barriers to entry, but their remedy seems to be based on an implicit hope that if they weaken Meta, a new, better advertising player will magically appear. I am deeply skeptical that this will happen, given the immense challenges of achieving scale and the significant barriers for any new data-driven player created by the GDPR itself.

I’ve characterized the Commission’s stance as a Pontius Pilate perspective. They seem to be washing their hands of the economic consequences of their actions, content to simply apply the DMA as they interpret it and say it’s the legislators fault if it goes wrong. You cannot properly interpret complex provisions like Article 5(2) DMA without a teleological approach—an understanding of the ultimate goal—which I believe is missing. A discussion of the basic economic effects on business users is completely absent from the Commission’s decision.

Path to compliance: what is expected of Meta? From a practical standpoint, the path to compliance for Meta is shrouded in ambiguity. The Commission’s demands seem to be shifting the goalposts. Their one clear complaint about choices Meta currently offers to users was about a preselected default option, which Meta already changed, yet apparently the Commission remains unsatisfied.

Engaging in some amateur kremlinology, I suspect the Commission is pushing for a maximalist interpretation: a scenario where the default option for users is a free service with less personalized, contextual ads. However, since contextual advertising performs much worse for advertisers, this model is likely not economically viable to Meta. If many users choose such a default, Meta might need more frequent non‑skippable ads to cover costs—but the Commission could then say such degradations aren’t allowed if they don’t flow from “less personalization” itself.

The Commission seems unconcerned, as their position is that they don’t care about economic consequences on gatekeepers unless a formal application is made under Article 9(1) DMA, arguing that the company’s economic viability is threatened. It looks as if they are trying to deliberately lead Meta into this exact situation.

Given this difficult dialogue, I’ve wondered if Meta might eventually consider a “nuclear” option: completely dismantling Meta Ads and creating separate, integrated ad backends for Facebook and Instagram. This might allow them to argue, under Recital 36 DMA, that there is no cross-use or combination of data between different services, thus avoiding Article 5(2) DMA.

However, I still speculated that some sort of settlement between Meta and the Commission is likely.

The concept of “necessity.” Finally, this whole debate hinges on the unresolved concept of “necessity,” which I see as a difficult axis of this conversation. I believe you cannot realistically separate the business aspect of the funding of a service from the service itself. If you make a service less personalized, it will generate less revenue, so it just has to cost more in a different way to remain viable. This is a business necessity. The problem is that the Court of Justice’s reasoning seems incoherent on this point. It appears to dismiss business necessity from consideration, but in the very same judgment, it says a platform may charge an  appropriate fee if necessary. This creates a contradiction: what kind of necessity allows for a fee if not a business one? This legal ambiguity will not leave us for a long time.

Simonetta, in particular, presented a strong counter-narrative to my points.

On the goal of the DMA (promoting user choice vs promoting contestability). While I focused on the negative impact on business users, Simonetta argued that the decision is fundamentally about giving choice to users. She believes the DMA should be used to experiment with new business models beyond targeted advertising, stating that this is not the only type of innovation that we are looking forward to in the EU. When I raised the concerns of advertisers, she dismissed them, questioning how independent those business users are from the gatekeepers and arguing that the current model is a threat to our democracy.

On the Commission’s role. My skepticism about the Commission’s vision was met with Simonetta’s call for strong enforcement. She stated that she doesn’t expect any settlement between Meta and the Commission, and that she wants the Commission to enforce the DMA “religiously” to clarify the rules and prevent Meta from gaming the system as she believes they have with the GDPR.

On Meta’s ability to comply: I suggested that the Commission’s demands were unclear, but Simonetta disagreed, stating that Meta is not in the dark about what they are supposed to do. She believes the legal principles are pretty clear and that Meta has been given enough indication of what a compliant solution looks like. She views Meta’s pay or consent model as a clear attempt to find a way of gaming the DMA.

Marco’s view was more of a middle ground. He agreed with my assessment that the Commission’s decision lacks clarity on what a compliant future model looks like. He also predicted that the parties will eventually find a compromise or settle, viewing the current decision as an interim step in a long regulatory dialogue.