Self-preferencing review for ranking and recommendation systems
self-preferencing-prohibitionDomain: competitionType: processDescription
A self-preferencing review program documents that the platform's ranking, recommendation, and visibility surfaces apply uniform criteria to first-party and third-party offerings on the same surface, and surfaces any deviation for competition-law review before it ships. The components are written ranking criteria with the inputs and their weights named, written justification for every platform-controlled input that could plausibly favor first-party listings (boost factors, eligibility filters, default sorts, exclusive placement slots), an audit trail of ranking-system changes with reviewer identity and date, and a competition-law sign-off path for product changes that touch ranking on surfaces where first-party offerings appear. The legal frame splits across three regimes. EU TFEU Article 102 dominance theory is the case-by-case channel that produced Google Shopping (€2.4B fine, upheld on appeal in 2024); the same framework remains operative for non-gatekeeper platforms. DMA Article 6(5) overlays an ex-ante prohibition on the designated gatekeepers, removing the dominance analysis and converting the conduct into a structural obligation. US Sherman Act §2 monopolization theory and FTC Act §5 unfair-methods analysis have raised self-preferencing claims across the Google, Amazon, Meta, and Apple matters, with judicial reception varying meaningfully by court. The UK runs Competition Act 1998 Chapter II abuse-of-dominance analysis alongside the DMCC Act 2024 SMS conduct-requirement framework, which makes uniform treatment an explicit obligation for designated firms. The defensible posture is that ranking is driven by user-relevance criteria applied uniformly to every listing; the indefensible posture is that ranking is driven in any part by whether the listing is the platform's own product. The line between those is sometimes subtle. An integration-availability boost (own products surface better because they integrate with the platform's wallet, identity, or fulfillment) can be characterized either as a relevance signal or as a self-preference, depending on how the boost was designed and what alternatives third parties had to access the same integration. That is where the documented-justification piece earns its keep. Evidence formats that hold up include the published ranking-criteria documentation, the ranking-change audit log showing reviewer identity and competition-law sign-off, the uniform-treatment test results comparing first-party and third-party listings on equivalent inputs, and the internal ranking-fairness audit conducted at a documented cadence.
Required by (4 regulations)
- US Antitrust (Platforms)
US framework: self-preferencing analyzed primarily under Sherman Act §2 (monopolization / attempted monopolization) and FTC Act §5 (unfair methods of competition). DOJ and FTC platform-conduct cases (Google, Amazon, Meta, Apple) have raised self-preferencing theories, with judicial reception varying by court.
Sherman Act, 15 U.S.C. §2; FTC Act, 15 U.S.C. §45
- UK Competition (Platforms)
UK framework: self-preferencing analyzed under Competition Act 1998 Chapter II (abuse of dominance). The CMA's mobile ecosystems and cloud-services market investigations have surfaced self-preferencing concerns; the SMS conduct-requirement framework under the DMCC Act 2024 makes this an explicit obligation for designated firms.
Competition Act 1998 c.41 Chapter II; DMCC Act 2024 c.13
- EU Competition (Platforms)
EU framework: self-preferencing analyzed under TFEU Article 102 dominance theory per Google Shopping (€2.4B fine, upheld 2024). For DMA-designated gatekeepers, Article 6(5) DMA imposes the prohibition ex ante regardless of dominance. The Article 102 framework remains operative for non-gatekeeper platforms.
TFEU Article 102; Commission Decision C(2017) 4444 final in Case AT.39740
- EU DMA
DMA Article 6(5) — ex-ante prohibition on treating own services and products more favourably in ranking and related indexing and crawling than similar third-party services; structurally distinct from the case-by-case Article 102 self-preferencing analysis.
Regulation (EU) 2022/1925
Fulfilled by (1)
- In-house build · full · high effortRanking-system review is product-engineering work coordinated with competition counsel. There is no commodity vendor for this; the Control is operationalized in-house against the company's specific ranking architecture.
Magist does not accept payment from vendors. Methodology.
Evidence formats
- documented ranking-system criteria
- ranking-change audit log
- competition-law sign-off on ranking changes that touch first-party offerings
- uniform-treatment test results
- internal ranking-fairness audit