Regulatory
Intelligence
at Scale

VI. Remedy Framework: Operational Design for Global Commitments

To maximize clearance probability across the United States, European Union, and United Kingdom, Amazon must offer a comprehensive, enforceable, and monitorable remedy package. Drawing on precedents (Google/Fitbit, Microsoft/Activision) and addressing authorities’ theories of harm, the following framework provides operational detail and implementation guidance.

1. Data Firewalls and Non-Use Commitments

Commitment: “Amazon commits that it will not use, have access to, or combine eBay non-public seller data (including but not limited to seller inventory, pricing, sales volumes, product development plans, and customer data) for the benefit of Amazon Retail, Amazon private brands, or any Amazon-owned or -operated business that competes with eBay sellers. This commitment applies globally and has a duration of 10 years from closing.”

Technical Implementation: Separate data storage with RBAC access controls. All data access logged and subject to quarterly trustee review. Annual compliance certification by Amazon’s Chief Compliance Officer with personal liability for false certification.^[8]

Breach Remedies: Immediate corrective action; $10 million per verified breach; escalation to FTC, DG COMP, CMA for potential fines or structural remedies. Precedent basis: Google/Fitbit data silo commitment.^[8]

2. Algorithmic Neutrality and Buy Box Commitments

Commitment: “Amazon commits to apply neutral, non-discriminatory ranking algorithms to all marketplace listings on Amazon Marketplace and eBay. Amazon will not preference Amazon Retail products, Amazon private brands, or Amazon-originated inventory in Buy Box allocation, search results, or advertising placements.”

• Transparent ranking factors: Publish all factors used in ranking algorithms (price, seller rating, fulfilment method, shipping speed, customer reviews).
• No hidden self-preferencing: Algorithms will not include parameters that systematically favour Amazon-owned inventory.
• Equal Buy Box access: When multiple sellers offer identical products at identical prices and service levels, allocation will rotate or randomize rather than defaulting to Amazon Retail.^[2]

Precedent basis: Amazon’s existing EU commitments (2022); Microsoft/Activision behavioural commitment models.^[12]

3. Prime and Logistics Parity

Commitment: “Amazon commits to provide eBay sellers with equal access to Prime badging, Prime-eligible fulfilment options, and Amazon logistics services on non-discriminatory terms. eBay sellers who meet objective performance criteria will qualify for Prime badging regardless of whether they use FBA or third-party logistics providers.”

Implementation: Seller-Fulfilled Prime (SFP) for eBay sellers meeting performance standards. No FBA bundling with other services. Transparent Prime eligibility criteria (2-day delivery >95%, satisfaction >4.5/5) published and applied uniformly. Quarterly reporting to trustee on seller participation.^[2]

4. MFN and Fee Transparency Commitments

Commitment: “Amazon commits that it will not impose, enforce, or maintain any most-favoured-nation clauses, pricing parity requirements, or availability restrictions that prevent sellers from offering more favourable terms on competing marketplaces (eBay, Walmart Marketplace, Etsy, or others).”

• Remove existing MFN clauses from seller agreements within 180 days of closing.
• Publish standardised fee schedule (referral fees, FBA fees, advertising costs) accessible to all sellers.
• Non-retaliation: no ranking, Buy Box, or fee penalties for offering better terms on competing platforms.

5. Interoperability, Portability, and API Access

Commitment: “Amazon commits to provide third-party logistics providers, advertising platforms, payment processors, and analytics tools with open API access to Amazon Marketplace and eBay on non-discriminatory, commercially reasonable terms.”

Implementation: Public API documentation for logistics, advertising, payment, and analytics integration. Data portability in standardised formats (CSV, JSON). Non-discriminatory access fees and SLAs equivalent for Amazon-owned and third-party providers. Precedent basis: Google/Fitbit Web API commitments; DMA Article 6 interoperability obligations.^[8][1]

6. No Tying, Bundling, or Ecosystem Funnelling

Commitment: “Amazon commits that it will not tie, bundle, or condition access to Amazon Marketplace, eBay, Prime, FBA, Amazon Advertising, or Amazon Pay in a manner that forecloses competition from independent providers.”

No defaults favouring Amazon in eBay search results. Voluntary opt-in for FBA, Amazon Advertising, and Amazon Pay. Cross-promotional offers must be opt-in with equivalent offers for competing services. Precedent basis: Booking/eTraveli ecosystem funnelling concerns; DMA Article 5 non-discrimination obligations.^[10][1]

7. Monitoring Trustee and Enforcement Mechanisms

Trustee Structure: Independent third-party firm appointed jointly by Amazon and enforcement authorities (FTC, DG COMP, CMA). Unrestricted access to Amazon’s systems, data, contracts, algorithms, and personnel. Quarterly compliance reports filed publicly.

• Financial penalties: $5-10 million per verified breach, escalating for repeat violations.
• Corrective action: Trustee can require immediate algorithm changes, access restoration, or data deletion.
• Escalation: Serious violations escalate to authorities who can impose additional fines, revoke clearance, or order structural divestitures.

Duration: All commitments remain in force for 10 years from closing, with option for authorities to extend if competition concerns persist. Precedent basis: Google/Fitbit 10-year monitoring trustee; Microsoft/Activision UILs.^[8][12]

Strategic Attack on Element 1: Fragmenting the ‘Overall Plan’

The non-settling party should:

• Map each alleged component of the SCI and identify which components it is alleged to have directly participated in versus those attributed to it solely through the SCI umbrella.
• Challenge the nexus between components, arguing that they reflect separate commercial contexts, different participants, different time periods, or different competitive dynamics.
• Exploit temporal gaps in participation. Intermittent or period-limited involvement undermines the continuity requirement and may break the chain of the single infringement.^[1]

The non-settling party should systematically challenge each piece of evidence, demanding that the Commission prove its case on the basis of contemporaneous documentary evidence rather than ex post admissions by interested parties.

The structural prejudice argument is intellectually compelling but strategically secondary. The courts have not yet annulled a decision solely on Article 41/48 grounds. These pleas are essential insurance for CJEU appeal, but the SCI substantive challenge is where the evidence is most exposed and where the General Court has the clearest track record of intervention.^[2]

I. Executive Summary

Ecosystem theories of harm have become a central analytical framework in digital-markets enforcement, especially in the EU and increasingly in the UK, with growing parallel influence in the U.S. and other major jurisdictions. They move beyond single-market dominance analysis by focusing on how the interconnections between operating systems, app stores, hardware, data assets, and advertising infrastructure can generate advantages that may be leveraged into adjacent or nascent markets.^[1]

These concerns now arise across merger control, single-firm enforcement, and ex ante digital regulation. The European Commission has been the most important institutional driver of this development, with the UK CMA and U.S. agencies generating significant parallel authorities. The cases surveyed below are the leading authorities as of March 2026.

A. European Union - Article 102 TFEU

Google Android (T-604/18) is a leading judicial endorsement of ecosystem-style abuse analysis under Article 102 TFEU. The General Court upheld the Commission’s finding that Google abused its dominant position in licensable mobile operating systems by tying Search and Chrome to Android and using anti-fragmentation agreements to restrict competing Android forks.^[2]

Apple - Mobile Payments (AT.40452) concerned Apple’s restriction of NFC access for rival mobile-wallet providers; the Commission accepted commitments in July 2024, making this a major EU example of ecosystem-based access theory.^[3]

Apple - App Store Practices (AT.40652), closed in November 2024, addressed Apple’s App Store rules for e-books and audiobooks and is another relevant example of app-distribution and platform-governance concerns.^[4]

Microsoft Teams (AT.40721; AT.40873) is a major enterprise-software tying/interoperability case that culminated in commitments in September 2025.^[5]

More recent matters - including Google AI and Data-related Practices (AT.40983) and Exclusion of AI competitors from WhatsApp (AT.41034) - may indicate that ecosystem-style analysis is beginning to extend into AI-related platform competition, but these developments are still recent and should be described with care.

B. Selected Global Developments beyond the EU, UK and U.S.

• South Korea (KFTC): In 2021, the KFTC fined Google roughly KRW 207 billion for restricting Android forks and reinforcing Google Search/Chrome distribution, making South Korea one of the clearest non-EU examples of ecosystem-style leveraging analysis.^[6]
• India (CCI): The CCI’s 2022 Android and Play Store decisions are among the most important G20 precedents applying competition law to platform ecosystems spanning operating systems, app distribution, billing, and developer access.^[7]
• Japan (JFTC / METI): Japan has emphasized ecosystem governance primarily through regulation, including the Transparency Act effective from February 2021, alongside competition scrutiny of major digital-platform practices.^[8]
• Germany (Bundeskartellamt): Section 19a GWB is one of the strongest ecosystem-style competition tools globally, targeting firms of paramount significance for competition across markets, including Alphabet/Google, Amazon, Apple, and Meta.^[9]