The Digital Markets Act and What It Changed for Advertisers

Under the Digital Markets Act, no finding of abuse is required and nobody has to be shown to have been harmed before the obligations bite. That single design choice separates it from every competition regime that preceded it, and it explains why gatekeeper behaviour in Europe changed within months rather than after the years an antitrust case takes. For advertisers the relevant provision is not the app store fight that dominated coverage. It is Article 5(2), which exists specifically to stop gatekeepers accumulating personal data in ways that raise entry barriers in online advertising — which makes the DMA, among other things, a piece of advertising market structure regulation.

A regulatory analysis from Digital, Ecommerce & Performance Marketing in Europe. For the channel-level effects see Google Ads after the DMA and Meta after pay-or-consent. Not legal advice. Last reviewed August 2026.

1. Ex-ante regulation and why it changes everything

Traditional competition law is ex-post: a regulator investigates conduct that has already occurred, proves abuse of a dominant position, and imposes a remedy years later. The DMA departs from that model entirely.

Based on published analysis describing the DMA’s departure from the ex-post competition model to an ex-ante model, in which no finding of abuse is required and no company needs to be shown to have harmed anyone before the duties apply. Legal characterisations are general.

This is why European digital marketing changed faster than any American antitrust case has ever changed anything. Nobody had to prove a harm first. The obligations simply started.

2. Who is designated, and how

Designation under Article 3 combines quantitative thresholds with a qualitative judgement, and three conditions must be met cumulatively: significant impact on the internal market, provision of a core platform service acting as an important gateway between businesses and end users, and an entrenched and durable position.

Sources: European Commission designations covering seven gatekeepers and 23 core platform services, with obligations binding from March 2024. Amazon Web Services and Microsoft Azure were reported as facing preliminary designation as of June 2026, which would extend the perimeter into cloud infrastructure.

The list matters for European performance marketing because six of the seven either sell advertising or control an environment in which advertising is measured. A regulation aimed at platform contestability inevitably becomes a regulation of your media supply.

3. Article 5(2) is the advertising article

If you read one provision of the DMA, read this one. Article 5(2) requires gatekeepers to obtain end user consent where personal data is combined or used across different core platform services.

The Commission’s stated purpose is explicit and directly commercial: to prevent the creation of market entry barriers through the accumulation of personal data for the provision of online advertising services by the gatekeeper, and to give users genuine free choice — which exists only if using the service does not depend on consenting to data processing, and if a comparable alternative is offered where consent is refused.

Based on the Commission’s explanation of Article 5(2) in its March 2024 non-compliance proceedings and the April 2025 decision. This is the provision under which Meta was fined €200 million.

Read that purpose clause again. The DMA is treating accumulated personal data as a barrier to entry in the advertising market. That is a structural claim about how European ecommerce marketing gets bought and sold, not a privacy argument.

4. The enforcement record so far

Sources: European Commission decisions and press materials; PwC Legal and Steptoe analyses of the April 2025 non-compliance decisions; Commission DMA review report COM(2026) 178 final of 28 April 2026; reporting on the July 2026 Google specification decision and the General Court judgments. A Google self-preferencing case under Article 6(5) was reported during 2026 as potentially producing the largest DMA fine to date — check whether a decision has since been published.

Both April 2025 decisions gave the companies 60 days to comply or face periodic penalty payments. Apple contested its decision, and Meta has described the finding as incorrect and unlawful while pursuing an appeal.

5. It is not a one-way ratchet

This is the part most commentary omits, and it matters for judging how the regime will develop. The Commission has not simply designated everything and won everything.

Sources: reporting on Commission designation decisions and General Court judgments including case T-1078/23. The picture is mixed by design — a regime that never lost a case would be a poor sign about the quality of its decisions.

6. The fines are not the deterrent

The DMA permits fines of up to 10% of a company’s annual worldwide sales. What has actually been imposed sits nowhere near that ceiling.

Based on reported analysis calculating Apple’s €500 million fine as approximately 0.13% of worldwide annual sales of roughly €345 billion, well below the 10% maximum available under the DMA.

For anyone forecasting how gatekeepers will behave, this is the useful conclusion: the financial penalties are not what changes conduct. The specification decisions are — the orders requiring Google to restructure results and share Search data, or Apple to open iOS connectivity by the end of 2026.

7. Does it actually work?

The Commission’s own review, published in April 2026, found that the DMA had effectively contributed to making EU digital markets fairer and more contestable, achieving a tangible positive impact within just over two years of full applicability.

Two caveats belong immediately alongside that. The Commission is assessing its own regulation. And the review itself notes that because most gatekeeper behavioural change was only visible from March 2024, the assessment period was in practice limited to about two years.

Sources as described. The NBER finding is the most useful single data point here because it is independent of both the regulator and the regulated: choice screens measurably shifted some users while leaving the underlying market share broadly intact.

Six million users moved and the market share barely moved. That is the honest summary of what remedies at this level achieve, and it should temper expectations in both directions.

8. What advertisers should do with all this

The DMA is not a compliance obligation for advertisers — it binds gatekeepers, not their customers. It is a source of structural change in your media supply, and it should be watched the way you would watch a supplier’s roadmap.

Operational assessment. Note that the Article 6(11) search data sharing decision was reported with eligibility criteria and deadlines running to 2027, so the competitive effect is prospective rather than immediate.

The single most useful habit is to date your baselines. European performance marketing operates on platforms whose interfaces are being altered by regulatory decision, and a performance comparison spanning one of those changes is comparing two different markets.

9. What this page does not cover

Scope statement. This page describes how the DMA affects advertising conditions in Europe. It is not legal advice and does not attempt to adjudicate the policy.

10. The 90-day plan

Indicative sequencing. Dating baselines comes first because it is the only step that makes every later change detectable.

11. Mistakes to avoid

Recurring errors in interpreting the DMA commercially; illustrative.

12. What changes next

A Google self-preferencing decision was expected. Reporting during 2026 indicated an Article 6(5) case that could produce the largest DMA fine to date, alongside a coalition of eighteen groups urging the Commission to act on search non-compliance, citing travel-sector click data.

Search data sharing has deadlines running to 2027. The Article 6(11) specification decision came with eligibility criteria and a timetable, so the competitive effect on European ecommerce marketing arrives gradually rather than at once.

Joint DMA-GDPR guidance is coming. The Commission and the European Data Protection Board intended to adopt joint guidelines in 2026, and a public consultation drew over a hundred responses with direct Article 5(2) consent implications for ad targeting.

Simplification pressure runs the other way. The DMA has been named among possible targets in the Commission’s digital fitness check, so the trajectory is not uniformly toward more intervention.

Key Takeaways

• The DMA requires no finding of abuse and no proof of harm. Obligations start on designation, which is why behaviour changed in months.
• Article 5(2) is the advertising provision, aimed explicitly at data accumulation raising entry barriers in online advertising.
• Seven gatekeepers, 23 core platform services, with cloud infrastructure reportedly facing preliminary designation in 2026.
• Apple’s €500m fine was about 0.13% of worldwide sales against a 10% ceiling. The remedies matter more than the money.
• It is not a one-way ratchet — X was not designated, Facebook Marketplace’s designation was revoked, and one Apple case closed without a finding.
• Independent evidence is mixed: choice screens gained Firefox around 6 million users while Google remained roughly 90% dominant.
• Date every baseline. Your platforms are being altered by regulatory decision.

Frequently Asked Questions

Conclusion

The Digital Markets Act is the most consequential piece of European digital marketing infrastructure that most marketers have never read, largely because it was reported as a fight about app stores and fines. The fights are real and the fines are trivial — 0.13% of worldwide sales in the largest case so far. What actually matters sits in the specification decisions: how search results are structured, what data gatekeepers must share, which environments open up and when.

For advertisers the practical stance is neither enthusiasm nor alarm. It is to recognise that your media supply is now shaped by regulatory decisions arriving on their own schedule, to watch those decisions the way you would watch a major supplier’s roadmap, and to date every baseline you hold so that when the next one lands you can tell what it did. The independent evidence suggests the effects are real but smaller than either side claims — six million users moved, and the market barely did.