Algorithmic Trust Intermediaries: Dominance Issues

Introduction

Algorithmic trust intermediaries are digital platforms or systems that create, verify, rank, or enforce “trust” between otherwise separated market participants through algorithms. Examples include online marketplaces, app stores, search engines, payment networks, rating/reputation systems, travel-booking platforms, digital identity providers, credit-scoring systems, and advertising exchanges.

Their competitive significance arises because the intermediary may control the algorithmic gateway through which businesses obtain visibility, customers, authentication, reputation, payments, or access to data. Once users and suppliers become dependent on that gateway, the intermediary may acquire substantial market power.

The central competition-law question is therefore not simply whether an algorithm is technically efficient, but whether a dominant intermediary uses its control over the trust mechanism to exclude rivals, discriminate between users, self-preference its own services, impose unfair conditions, or make switching prohibitively difficult.

In the EU, this issue is increasingly addressed through both Article 102 TFEU and the Digital Markets Act (DMA). The European Commission currently designates services such as Google Shopping, Google Play, Amazon Marketplace, Apple App Store and Booking.com as core platform/intermediation services subject to DMA obligations.

1. Meaning of an Algorithmic Trust Intermediary

A conventional intermediary connects two sides of a market:

Supplier → Intermediary → Consumer

An algorithmic trust intermediary goes further. It determines who is visible, credible, accessible, recommended, verified or transactable.

For example:

• Amazon's ranking algorithm influences which seller consumers see first.
• Booking.com's algorithm determines which accommodation providers receive visibility.
• Google determines how search results are displayed.
• Apple and Google control app-distribution mechanisms.
• Payment networks determine whether transactions can be completed.
• Rating algorithms determine a seller's reputation.
• Digital identity systems determine whether a participant is authenticated.
• Advertising exchanges determine which advertiser obtains access to a particular impression.

Thus, the intermediary is not merely providing infrastructure. It may become a market-governance mechanism.

2. Why Algorithmic Trust Creates Dominance

A. Network effects

The more consumers use the platform, the more suppliers join it.

The more suppliers join, the more useful the platform becomes to consumers.

This creates:

More users → more suppliers → more transactions → more data → better algorithms → more users

A successful intermediary can therefore develop substantial entry barriers.

B. Reputation concentration

Trust may become concentrated in the intermediary's rating or verification system.

A supplier that has accumulated thousands of reviews on one platform may find it difficult to move elsewhere.

This produces a form of reputational switching cost.

The platform therefore controls not merely customer access but also the supplier's accumulated digital reputation.

C. Data advantage

The intermediary can collect information concerning:

• consumer preferences;
• transaction frequency;
• supplier performance;
• conversion rates;
• prices;
• cancellations;
• consumer complaints;
• search behaviour;
• rankings;
• transaction histories.

The resulting data advantage can improve the intermediary's algorithm and reinforce its market position.

3. Relevant Market Definition

The first legal question is usually whether the intermediary possesses dominance.

Several markets may need to be considered separately.

Possible relevant markets

1. Intermediation services
2. Search and discovery services
3. Marketplace services
4. App distribution
5. Payment-network services
6. Online advertising intermediation
7. Reputation/rating services
8. Digital identity or authentication services

The two-sided nature of the intermediary is particularly important.

The CJEU's decision in Booking.com, Case C-264/23 confirmed that market definition involving an online hotel-intermediation platform requires concrete consideration of substitutability between online intermediation and other sales channels.

This prevents authorities from mechanically treating every side of a platform as an ordinary standalone market.

4. Forms of Algorithmic Dominance

A. Algorithmic self-preferencing

A platform may give its own products or affiliated services better rankings than competing services.

Example:

Platform owns marketplace + competing seller

The algorithm could theoretically:

• place its own products higher;
• provide them better recommendations;
• give them greater visibility;
• reduce rivals' ranking;
• provide preferential search filters.

This was central to the Google Shopping litigation.

B. Algorithmic discrimination

A dominant intermediary may give different treatment to similarly situated businesses.

Examples include:

• different ranking;
• different commission rates;
• different access to data;
• different search visibility;
• different verification requirements;
• different recommendation probabilities.

Discrimination becomes particularly problematic where the intermediary cannot provide objective, transparent criteria for the distinction.

C. Algorithmic exclusion

An intermediary can potentially use algorithms to make rivals effectively invisible.

A competitor technically remains on the platform but receives:

• very low ranking;
• reduced recommendations;
• limited advertising access;
• delayed verification;
• restricted APIs;
• reduced consumer exposure.

This may constitute de facto exclusion without formal exclusion.

5. Algorithmic Lock-In

A major concern is the transformation of trust into a switching barrier.

Suppose a seller has:

• 50,000 reviews;
• a high reputation score;
• transaction history;
• verified status;
• customer-following;
• algorithmic ranking.

If these assets cannot be transferred to another intermediary, the supplier may be commercially dependent on the incumbent.

This produces:

Trust accumulation → switching cost → dependency → bargaining asymmetry → potential dominance

Therefore, data portability and reputation portability can become competition-law issues.

6. Algorithmic Control of Access

A trust intermediary may control access to the market itself.

For example:

Seller → platform verification → platform ranking → consumer access

If verification is essential to obtaining customers, arbitrary or discriminatory verification can become an exclusionary mechanism.

The same reasoning applies to:

• API access;
• payment authorization;
• app approval;
• marketplace listing;
• search indexing;
• identity verification;
• advertising access.

7. Important Case Laws

1. Google and Alphabet v Commission — Google Shopping

Case C-48/22 P, CJEU, 10 September 2024

This is one of the most important authorities for algorithmic intermediary dominance.

Google operated the dominant general-search service and gave preferential display treatment to its own comparison-shopping service. Rival comparison-shopping services were disadvantaged through the manner in which Google's search algorithms and presentation mechanisms operated.

The CJEU upheld the €2.4 billion fine and rejected Google's appeal.

Principle

A dominant digital intermediary cannot necessarily escape Article 102 scrutiny merely by characterising its conduct as an improvement to its own algorithmic service.

The case is particularly significant for:

• algorithmic ranking;
• self-preferencing;
• leveraging;
• search neutrality;
• visibility discrimination;
• foreclosure of competing intermediaries.

Relevance

It establishes an important proposition:

Control over algorithmic visibility can constitute a means of exercising market power.

2. Booking.com BV v 25hours Hotel Company Berlin GmbH

Case C-264/23, CJEU, 19 September 2024

Booking.com operated an online accommodation-intermediation platform.

The dispute concerned price-parity clauses, under which hotels were restricted from offering lower prices through certain alternative channels.

The CJEU held that such parity clauses cannot, in principle, automatically be treated as ancillary restraints. It also emphasised the need to examine substitutability between online intermediation and other sales channels.

Principle

Platform contractual conditions must be assessed in light of:

• the platform's intermediary role;
• market structure;
• substitutability;
• competitive effects;
• the relationship between platform and business users.

Relevance

This case demonstrates that algorithmic trust intermediaries can exercise market power not only through algorithms but also through contractual rules governing access to their ecosystem.

3. Ohio v American Express Co.

585 U.S. 529 (2018)

The U.S. Supreme Court addressed the economics of two-sided transaction platforms.

American Express operated a payment network connecting merchants and cardholders.

The Court treated the credit-card network as a two-sided transaction platform, recognising that the two sides interact through a common platform and that competitive effects can arise across both sides.

Principle

Two-sided platforms may need to be analysed as integrated platforms rather than by examining one side in isolation.

Relevance to trust intermediaries

An algorithmic intermediary may simultaneously serve:

• consumers;
• suppliers;
• advertisers;
• payment providers;
• content creators.

Conduct affecting one side can change participation and incentives on the other.

Therefore:

Platform dominance analysis must consider cross-side network effects.

4. Epic Games, Inc. v Google LLC

N.D. Cal., No. 3:20-cv-05671

Epic Games challenged Google's practices concerning Android app distribution and Google Play billing.

In December 2023, a jury found Google liable for unlawfully maintaining monopoly power in relevant Android app-distribution markets. The U.S. Department of Justice has subsequently referred to the case as a jury verdict finding Google illegally monopolized app distribution on Android phones.

Principle

Control over an app-distribution gateway can create substantial competitive power.

Relevance

An app store functions as a trust intermediary because it:

• verifies applications;
• controls distribution;
• controls consumer discovery;
• provides payment mechanisms;
• establishes technical rules;
• controls developer access.

Thus, algorithmic approval and ranking systems may become important instruments of market power.

5. Epic Games v Apple Inc.

559 F. Supp. 3d 898 (N.D. Cal. 2021)

The case concerned Apple's App Store ecosystem and Apple's rules governing developers.

The court examined Apple's control over:

• app distribution;
• payment processing;
• developer access;
• consumer discovery;
• commission arrangements.

Although the court did not find Apple liable on every federal antitrust claim asserted by Epic, it found Apple's anti-steering restrictions unlawful under California's Unfair Competition Law.

Principle

A platform may possess substantial control over the commercial relationship between businesses and consumers even where competitors technically exist outside the platform.

Relevance

This illustrates the importance of steering restrictions.

If an intermediary prevents a supplier from telling consumers:

"You can buy this service more cheaply directly from us"

the intermediary may be protecting its position as the unavoidable gateway.

6. Intel Corp. v Commission

Case C-413/14 P, CJEU, 6 September 2017

Intel concerned loyalty rebates imposed by a dominant undertaking.

The CJEU required consideration of the circumstances surrounding rebate schemes and their potential ability to foreclose equally efficient competitors.

Relevance to algorithmic intermediaries

The principle can extend conceptually to algorithmically administered incentives.

For example, a dominant intermediary could design algorithmic incentives that reward suppliers for:

• exclusivity;
• maintaining particular transaction shares;
• avoiding rival platforms;
• meeting platform-specific targets.

The legal inquiry would focus on whether the system has exclusionary effects rather than merely whether it is technologically automated.

8. Indian Competition-Law Relevance

India provides particularly important examples because the Competition Commission of India has extensively examined digital-platform dominance.

Google Android / Matrimony.com

In Matrimony.com Ltd. v Google LLC & Others, CCI considered Google's position in online general search and related digital markets. The CCI's 2018 order is recorded under Case Nos. 07 and 30 of 2012.

The CCI subsequently examined Google's Android ecosystem and imposed a monetary penalty of approximately ₹1,337.76 crore in 2022 for anti-competitive practices concerning Android mobile devices.

Relevance

These proceedings demonstrate how an ecosystem provider can become a critical gateway between:

Users ↔ operating system ↔ app developers ↔ search services ↔ payment services

The intermediary's control over multiple connected layers can strengthen network effects and increase switching costs.

9. Algorithmic Trust and Essential-Facility-Type Arguments

A particularly difficult question arises when a dominant intermediary becomes practically indispensable.

Suppose:

Platform X controls 90% of consumer transactions.

A rival requests:

• API access;
• ranking access;
• verification;
• interoperability;
• transaction data;
• reputation portability.

If X refuses access, competition law may ask whether the refusal is exclusionary.

However, mere importance is not automatically enough.

Authorities generally need to examine issues such as:

1. dominance;
2. indispensability;
3. objective justification;
4. feasibility of access;
5. competitive harm;
6. effects on consumers;
7. whether access would undermine legitimate security or privacy interests.

10. Algorithmic Transparency

Transparency is increasingly relevant because the intermediary often possesses information unavailable to its business users.

The platform knows:

• why a seller was demoted;
• why an application was rejected;
• how rankings are calculated;
• which consumers were targeted;
• how commission levels are determined;
• which signals affect reputation.

The business user may know none of these things.

This produces an information asymmetry.

Competition concerns become stronger where the intermediary simultaneously:

designs the algorithm + controls the data + applies the rules + competes against the businesses governed by those rules.

11. Self-Preferencing as a Structural Problem

Consider:

Platform → ranking algorithm → consumer visibility

The platform then introduces its own competing product:

Platform's own product → same algorithm → preferred ranking

The intermediary has effectively become:

1. market operator;
2. rule-maker;
3. data collector;
4. algorithm designer;
5. competitor.

That combination creates a structural conflict of interest.

The Google Shopping judgment is especially significant because the CJEU confirmed that Google's preferential treatment of its own specialised search results could constitute abuse of dominance.

12. Algorithmic Reputation Manipulation

Another possible abuse involves manipulation of trust scores.

For example, a dominant marketplace might:

• suppress negative reviews of its own products;
• amplify negative reviews of competitors;
• change seller scores;
• manipulate recommendation probabilities;
• alter verification status;
• downgrade sellers using competing services.

Such conduct could transform reputation management into exclusionary conduct.

The critical evidence would include:

• algorithmic logs;
• ranking changes;
• internal communications;
• A/B testing;
• consumer conversion data;
• comparative treatment;
• effects on rival suppliers.

13. Data Advantage and Algorithmic Dominance

Data can reinforce dominance through a feedback loop:

Large user base

↓

More transactions

↓

More behavioural data

↓

Better algorithm

↓

Better recommendations / trust assessment

↓

More users

This is a form of algorithmic network-effect reinforcement.

Competition authorities therefore increasingly consider whether data access and interoperability can prevent a dominant intermediary from converting its information advantage into durable exclusion.

14. Anti-Steering and Trust Intermediaries

Anti-steering restrictions are particularly important.

A platform may tell businesses:

You may sell through our platform, but you cannot direct customers to your own website.

The consequence is:

Consumer discovery → intermediary → transaction → intermediary commission

The business cannot escape the intermediary even when the actual transaction could occur outside the platform.

The European Commission has treated anti-steering as a specific DMA obligation. Apple was found in 2025 to have breached the DMA's Article 5(4) anti-steering obligation concerning the App Store.

The Commission subsequently identified Google's Play anti-steering practices as a DMA enforcement issue as well.

15. Competition Concerns Created by Algorithmic Trust Intermediaries

16. Regulatory Approach

A comprehensive competition-law analysis should generally proceed through the following sequence:

Step 1 — Identify the intermediary

Determine precisely what the platform intermediates:

Consumers + suppliers + data + payments + reputation

Step 2 — Define the relevant market

Examine:

• direct competitors;
• alternative platforms;
• offline alternatives;
• direct sales;
• multi-homing;
• switching costs.

Step 3 — Establish dominance

Consider:

• market share;
• network effects;
• data advantages;
• entry barriers;
• switching costs;
• ecosystem control.

Step 4 — Identify algorithmic conduct

Ask whether the algorithm:

• ranks;
• recommends;
• excludes;
• discriminates;
• verifies;
• prices;
• allocates access.

Step 5 — Examine competitive effects

Determine whether the conduct:

• forecloses rivals;
• raises rivals' costs;
• reduces innovation;
• increases supplier dependency;
• reduces consumer choice.

Step 6 — Examine justification

Possible legitimate justifications include:

• cybersecurity;
• fraud prevention;
• privacy;
• technical reliability;
• consumer protection;
• quality assurance.

Step 7 — Consider remedies

Potential remedies include:

• non-discriminatory access;
• interoperability;
• data portability;
• reputation portability;
• transparent ranking criteria;
• anti-self-preferencing obligations;
• restrictions on exclusivity;
• steering rights;
• structural separation in exceptional circumstances.

17. Relationship with the Digital Markets Act

The DMA represents a shift from purely ex-post dominance analysis toward ex-ante regulation of powerful gatekeepers.

The EU currently identifies seven gatekeepers—Alphabet, Amazon, Apple, Booking, ByteDance, Meta and Microsoft—with designated core platform services including online-intermediation services such as Google Shopping, Google Play, Amazon Marketplace, Apple App Store and Booking.com.

This is especially relevant to algorithmic trust intermediaries because many of their competitive advantages arise before a conventional antitrust infringement can be conclusively demonstrated.

18. Key Legal Principles Emerging from the Case Law

The cases collectively demonstrate several important propositions:

Principle 1

Digital intermediaries can possess market power because of network effects and gateway control.

Principle 2

Algorithmic self-preferencing can fall within traditional abuse-of-dominance analysis.

The Google Shopping judgment is particularly important here.

Principle 3

Two-sided platforms require cross-market analysis.

Ohio v American Express illustrates the importance of understanding both sides of a transaction platform.

Principle 4

Contractual restrictions imposed by intermediaries can reinforce platform dominance.

Booking.com demonstrates the relevance of parity clauses and alternative distribution channels.

Principle 5

Control over an essential digital gateway may become a competition-law concern even where the intermediary does not manufacture the underlying product.

This is especially relevant to app stores, search engines, marketplaces and payment networks.

Principle 6

Algorithmic automation does not immunise conduct from competition law.

An algorithm is simply the mechanism through which the conduct is implemented.

19. Conclusion

Algorithmic trust intermediaries occupy a particularly powerful position in modern digital markets because they do not merely connect buyers and sellers—they determine whom market participants can trust, find, access and transact with.

Their dominance can therefore arise from a combination of:

network effects + data accumulation + reputation concentration + algorithmic ranking + switching costs + gateway control.

The most important competition risks are self-preferencing, discriminatory ranking, exclusionary verification, anti-steering restrictions, data exploitation, reputation lock-in, interoperability restrictions and algorithmically reinforced supplier dependency.

The leading authorities—particularly Google Shopping, Booking.com, Ohio v American Express, Epic Games v Google, Epic Games v Apple, Intel v Commission, together with the Indian CCI's Matrimony.com/Google and Google Android proceedings—show how competition law is adapting to intermediaries whose market power is exercised through algorithms rather than traditional physical infrastructure.

The central legal proposition can therefore be stated as: