Competition Law and Competition Governance in Data-Intensive Industries

1. Introduction

Data-intensive industries are sectors in which large-scale collection, processing, aggregation, analysis, sharing, and monetisation of data are central to the production or delivery of goods and services. Examples include:

• digital platforms and social media;
• search engines and online advertising;
• e-commerce and marketplaces;
• cloud computing;
• financial technology and digital banking;
• healthcare and pharmaceutical information services;
• telecommunications;
• artificial intelligence and machine learning;
• connected vehicles and IoT;
• insurance and credit scoring;
• data brokerage and analytics;
• digital payments.

Competition law traditionally focused on price, output, market share, and physical infrastructure. Data-intensive markets require a broader approach because a service may be supplied at a monetary price of zero while the undertaking competes through user attention, data, algorithms, quality, privacy, interoperability and innovation.

The central competition-law question is therefore not simply who has the most data, but:

Whether control over data creates, preserves, or reinforces market power and whether that power is being used to exclude competitors or distort competitive conditions.

Modern competition governance increasingly combines conventional antitrust law with data-access rules, interoperability obligations, privacy regulation, merger control, platform regulation, algorithmic governance and sector-specific regulation.

2. Meaning of Data-Intensive Industries

A data-intensive industry generally has five characteristics:

A. Large-scale data collection

Companies collect:

• personal data;
• transaction data;
• location data;
• search histories;
• behavioural data;
• device data;
• financial information;
• health information;
• industrial data;
• metadata.

B. Data-driven network effects

More users can generate more data, which can improve the service, which attracts more users, generating still more data.

This creates a potentially self-reinforcing cycle:

Users → Data → Better algorithms → Better service → More users → More data

C. Economies of scale and scope

Once an undertaking possesses a large data infrastructure, the same data may be used across several markets.

For example:

Search data → advertising → consumer profiling → AI → recommendation services

D. Low marginal cost

Digital services can often serve additional users at relatively low marginal cost, making rapid expansion possible.

E. High switching and ecosystem effects

Data accumulation may be combined with:

• account history;
• recommendation profiles;
• stored payment information;
• APIs;
• proprietary standards;
• loyalty programmes;
• cloud infrastructure;
• software ecosystems.

This can make switching more difficult.

3. Competition Law Framework

Competition governance in data-intensive industries generally operates through four principal areas.

I. Anti-competitive agreements

Authorities may examine:

• data-sharing agreements;
• information exchanges;
• data pooling;
• algorithmic coordination;
• exclusivity;
• non-compete arrangements;
• parity clauses;
• restrictions on data portability;
• agreements restricting access to APIs.

The relevant provisions will depend upon the jurisdiction.

For example:

• EU: Articles 101 and 102 TFEU;
• US: Sherman Act §§1–2 and Clayton Act;
• UK: Competition Act 1998;
• India: Competition Act 2002;
• Germany: GWB, including §19a.

4. Abuse of Dominance

Data can become relevant to dominance in two different ways.

First: data may be a source of market power

A company may possess:

• unique datasets;
• massive datasets;
• real-time datasets;
• proprietary behavioural data;
• data generated by an extensive ecosystem.

Second: data may be an instrument of exclusion

A dominant undertaking may:

• deny competitors access to essential data;
• degrade interoperability;
• combine datasets in exclusionary ways;
• impose discriminatory access conditions;
• use competitor data to compete against those competitors;
• restrict data portability;
• tie data access to another service;
• use data to favour its own downstream products.

5. Data as a Potential Competitive Advantage

Possession of data does not automatically establish dominance.

Competition authorities normally need to examine:

1. whether the data is commercially important;
2. whether it is difficult to replicate;
3. whether competitors can obtain substitutes;
4. whether the data is timely or historical;
5. whether competitors require the data to compete;
6. whether access is technically feasible;
7. whether the data produces significant network effects;
8. whether the data creates switching costs;
9. whether alternative datasets exist;
10. whether the advantage can be maintained over time.

Thus:

Large data holdings ≠ automatically dominant position.

The competitive significance depends on the characteristics of the dataset and the relevant market.

6. Data as an Essential Facility

A particularly important issue is whether certain datasets can constitute an essential input.

The traditional essential-facilities doctrine generally requires stringent conditions, particularly where access would interfere with property rights or investment incentives.

The landmark case is:

Case 1 — IMS Health GmbH & Co. OHG v NDC Health, C-418/01

The case concerned pharmaceutical sales data organised according to the "1860 brick" structure used in Germany.

The Court of Justice examined when refusal by a dominant undertaking to license a protected structure could constitute abuse. It identified exceptional circumstances, including indispensability, elimination of effective competition and prevention of a new product for which there was consumer demand.

Competition-law significance

IMS Health demonstrates that:

• commercially important data structures can have competitive significance;
• interoperability and industry standards can create dependency;
• not every refusal to license data is abusive;
• compulsory access requires stringent conditions.

Modern application

The reasoning is relevant to:

• healthcare datasets;
• financial datasets;
• mapping databases;
• technical standards;
• industrial IoT datasets;
• proprietary AI training datasets.

7. Interoperability and Data Access

Data-intensive competition frequently overlaps with interoperability.

Case 2 — Microsoft Corp. v Commission, T-201/04

The EU Microsoft litigation concerned Microsoft's refusal to provide interoperability information necessary for competing work-group server operating systems.

The General Court upheld the Commission's finding concerning Microsoft's refusal to provide interoperability information and confirmed the importance of enabling competitors to remain viable in the relevant market.

Significance

The case establishes an important principle for data-intensive markets:

Control over information necessary for interoperability can have competitive consequences.

The concept extends conceptually to:

• APIs;
• technical protocols;
• data formats;
• cloud interoperability;
• connected-device interfaces;
• digital identity systems.

However, compulsory access remains legally sensitive and must be justified under the applicable legal test.

8. Data-Driven Self-Preferencing

A platform can collect data from businesses operating on its platform and then potentially use that information to favour its own products.

This creates a vertical conflict:

Platform operator + infrastructure provider + competitor

For example:

Third-party sellers → platform data → platform's retail business

This raises questions about:

• self-preferencing;
• discriminatory rankings;
• access to data;
• use of non-public business information;
• algorithmic ranking.

9. Google Shopping

Case 3 — Google Search (Shopping)

The European Commission's Google Shopping decision concerned Google's treatment of its own comparison-shopping service within general search results.

The Commission concluded that Google positioned and displayed its own comparison-shopping service more favourably than competing services, with traffic diversion forming a central part of the competitive analysis.

Importance for data-intensive industries

Although the case was not simply a "data case", it illustrates how:

• search algorithms;
• user data;
• ranking;
• traffic;
• platform position;

can interact to reinforce market power.

The broader lesson is that competition analysis must examine how control of a data-rich gateway affects downstream competitors.

10. Facebook/Meta and Combining Data

Case 4 — Bundeskartellamt v Facebook/Meta

Germany's Facebook proceeding is one of the most important examples of the interaction between competition law and data governance.

The Bundeskartellamt prohibited Facebook from combining user data from different sources without the required voluntary consent. The case involved data from Facebook and other services and raised the question of whether Facebook's data-collection practices were connected with its dominant position.

The case subsequently involved litigation concerning the relationship between competition law and GDPR principles.

Competition significance

The case demonstrates that:

Privacy conditions can have competitive relevance.

A deterioration in privacy conditions may potentially be relevant to competition where:

• a dominant platform imposes the condition;
• users lack meaningful alternatives;
• data accumulation strengthens market power;
• competitors cannot reproduce the same data advantage.

It therefore helped develop the concept of non-price competition.

11. Amazon Marketplace and Seller Data

Case 5 — CMA Investigation into Amazon Marketplace

The UK's Competition and Markets Authority investigated Amazon concerning its use of non-public third-party seller data, the selection of offers for the Buy Box, and eligibility for Amazon's Prime label.

The CMA ultimately accepted commitments from Amazon in November 2023.

Competitive concern

The underlying structural concern can be represented as:

Independent sellers → platform data → Amazon obtains information → Amazon also competes as retailer

This raises the possibility of dual-role conflicts.

A marketplace operator can potentially act simultaneously as:

1. infrastructure provider;
2. data collector;
3. algorithmic gatekeeper;
4. retailer;
5. logistics provider.

Governance response

Competition governance can therefore require:

• restrictions on use of non-public seller information;
• transparent ranking criteria;
• fair Buy Box procedures;
• non-discriminatory access;
• monitoring commitments.

12. Meta and Advertising Data in the UK

Case 6 — CMA Investigation into Meta's Use of Data

The CMA investigated Meta's use of data obtained from digital advertising services and its potential competitive significance for Facebook Marketplace.

The investigation examined whether Meta's collection and use of advertising data could provide it with an advantage over competitors. The CMA ultimately accepted commitments, with the investigation closing in November 2023; the commitments were subsequently varied in 2024.

Importance

This illustrates an increasingly important competition problem:

Data generated in Market A → used to strengthen Market B

This is sometimes called data leveraging or cross-market data advantage.

13. Facebook/FTC — Data, APIs and Acquisitions

Case 7 — FTC v Facebook/Meta

The US Federal Trade Commission's litigation against Facebook alleged that Facebook maintained monopoly power through a course of conduct involving acquisitions and restrictions imposed on software developers.

The FTC's complaint alleged, among other matters, that Facebook restricted API access in ways that could disadvantage competitive threats. The FTC's case remains a major US competition proceeding concerning Meta.

Competition relevance

The case illustrates how data-intensive ecosystems can combine:

• network effects;
• data advantages;
• API control;
• platform access;
• acquisitions;
• interoperability restrictions.

The acquisition dimension is especially important because data-rich firms may acquire emerging competitors before those competitors develop comparable scale.

14. Data and Merger Control

Data-intensive industries create particular merger-control problems.

Traditional merger analysis asks:

What happens to prices and output?

Data-driven merger analysis may additionally ask:

• Will the transaction combine uniquely valuable datasets?
• Will competitors lose access to important data?
• Will the merged firm increase network effects?
• Will data interoperability decline?
• Will the transaction increase switching costs?
• Will an emerging competitor disappear?
• Will the merged firm gain an advantage in AI training?
• Will the transaction facilitate behavioural profiling?

15. Killer Acquisitions

A large data platform may acquire a smaller company whose current revenues are modest but whose:

• technology;
• dataset;
• user base;
• algorithm;
• innovation pipeline;

could become competitively important.

Consequently, competition authorities increasingly examine acquisitions based on future competitive significance, not merely present turnover.

This has implications for:

• AI startups;
• health-tech;
• fintech;
• cybersecurity;
• adtech;
• social media;
• cloud services.

16. Data Portability as a Competition Remedy

Data portability can reduce switching costs.

A user may hesitate to switch because the existing platform possesses:

• contacts;
• transaction history;
• playlists;
• photographs;
• professional connections;
• health information;
• financial records;
• preferences.

If users can transfer their data efficiently, competitors may find it easier to attract customers.

Therefore:

Data portability → lower switching costs → greater contestability

But portability must be balanced against:

• privacy;
• cybersecurity;
• confidentiality;
• intellectual property;
• third-party rights.

17. Data Sharing

Competition governance must distinguish between pro-competitive data sharing and anti-competitive information exchange.

Potentially pro-competitive

Data sharing may improve:

• safety;
• research;
• fraud detection;
• interoperability;
• infrastructure efficiency;
• healthcare outcomes;
• environmental monitoring.

Potentially anti-competitive

Data sharing can facilitate:

• price coordination;
• output coordination;
• customer allocation;
• exclusion;
• discriminatory treatment;
• algorithmic collusion.

Thus:

Data sharing is not inherently pro-competitive or anti-competitive. Its competitive effect depends upon the information exchanged, participants, timing, purpose and market structure.

18. Algorithmic Competition

Data-intensive industries frequently use algorithms to determine:

• prices;
• rankings;
• recommendations;
• credit decisions;
• advertising;
• inventory;
• delivery allocation.

Algorithms can create competition concerns even without an explicit human agreement.

Potential mechanisms include:

A. Algorithmic implementation

Competitors expressly agree on a strategy and algorithms implement it.

B. Algorithmic monitoring

Algorithms monitor competitors and facilitate rapid reactions.

C. Hub-and-spoke coordination

A common platform or algorithm may influence multiple competing firms.

D. Autonomous coordination

Algorithms may independently respond to market information in ways that produce stable coordination.

Competition law must therefore investigate human conduct, algorithm design, data inputs and resulting market effects.

19. Data and Market Definition

Traditional market-definition techniques can become difficult where services are supplied for zero monetary prices.

For example:

Consumer → free search service

but:

Consumer → attention/data → platform → advertiser revenue

Consequently, authorities may consider:

• quality;
• privacy;
• data collection;
• advertising;
• user attention;
• functionality;
• switching costs.

The relevant competitive relationship may exist even where consumers pay nothing.

20. Multi-Sided Markets

Data-intensive platforms frequently connect several groups.

For example:

Users ↔ Platform ↔ Advertisers

or:

Consumers ↔ Marketplace ↔ Sellers

or:

Patients ↔ Healthcare platform ↔ Providers

Competition analysis therefore needs to examine interdependent sides of the platform rather than treating each side as an isolated market.

A practice benefiting one side may harm another.

21. Network Effects and Data Feedback Loops

Data can produce powerful feedback mechanisms.

Direct network effect

More users → more value.

Data network effect

More users → more data → improved product → more users.

Cross-side effect

More users → more advertisers → more revenue → better service → more users.

These mechanisms can create entry barriers even when technological entry appears easy.

22. Data Concentration and Entry Barriers

A new entrant may be technologically capable of creating a competing service but unable to obtain comparable data.

For example:

Incumbent

10 billion behavioural observations
↓
Better model
↓
Better recommendations
↓
More users
↓
More data

Entrant

Small user base
↓
Limited data
↓
Less accurate model
↓
Difficulty attracting users

This is sometimes described as a data advantage feedback loop.

Competition authorities should determine whether the advantage reflects legitimate competition or exclusionary conduct.

23. Privacy as a Dimension of Competition

Data-intensive competition increasingly involves quality competition.

Quality may include:

• privacy;
• security;
• reliability;
• transparency;
• data minimisation;
• user control.

Where consumers cannot meaningfully choose between different privacy conditions because a dominant undertaking controls the relevant market, privacy may become relevant to competition analysis.

The Meta/Facebook litigation in Germany is particularly significant in this respect.

24. Digital Gatekeeper Regulation

Traditional ex-post competition law can sometimes intervene only after substantial competitive harm has developed.

Modern regulation therefore increasingly uses ex-ante obligations.

The EU Digital Markets Act is an important example.

The European Commission designated major companies including Alphabet, Amazon, Apple, ByteDance, Meta and Microsoft as DMA gatekeepers in 2023.

The DMA framework includes obligations concerning areas such as:

• data combination;
• interoperability;
• access;
• self-preferencing;
• user choice;
• advertising transparency.

In 2026, the Commission adopted measures concerning Google's sharing of anonymised Search data with eligible competing search engines under Article 6(11) DMA.

25. Germany's Section 19a GWB

Germany has adopted a particularly important ex-ante competition-governance mechanism through Section 19a of the GWB.

The provision permits enhanced abuse control for companies of paramount significance for competition across markets.

The Bundeskartellamt has applied this framework to companies including:

• Meta;
• Alphabet/Google;
• Amazon;
• Apple;
• Microsoft.

This represents a shift from:

Traditional model:
Wait for established abuse → investigate → remedy.

towards:

Digital governance model:
Identify systemic market power → impose targeted behavioural restrictions earlier.

26. Data Access and Competition Governance

An effective data-governance framework should distinguish several forms of access:

27. Remedies

Competition authorities may employ several remedies.

Structural remedies

• divestiture;
• separation of businesses;
• restrictions on acquisitions.

Behavioural remedies

• non-discrimination;
• data-access obligations;
• interoperability;
• API access;
• restrictions on data combination;
• transparency obligations.

Data remedies

• data portability;
• data sharing;
• data silos;
• restrictions on cross-use;
• deletion requirements;
• independent data governance.

Algorithmic remedies

• audit requirements;
• ranking transparency;
• monitoring;
• non-discrimination requirements;
• independent technical oversight.

28. Role of Data Protection Law

Competition law and data protection law are distinct but increasingly interconnected.

Competition law asks:

Does conduct restrict competition?

Data protection law asks:

Is personal data processed lawfully and fairly?

Consumer protection asks:

Are consumers being deceived or exploited?

Digital regulation asks:

Are systemic platforms complying with ex-ante obligations?

The same conduct can potentially engage several regulatory regimes.

29. Competition Governance Model

A comprehensive governance framework can be represented as:

Data Collection
↓
Data Aggregation
↓
Market Power Assessment
↓
Competition Risk Analysis
↓
Privacy/Data Protection Review
↓
Interoperability & Access Assessment
↓
Merger/Conduct Investigation
↓
Remedy Selection
↓
Continuous Monitoring

30. Six Core Competition Risks

1. Data foreclosure

A dominant firm denies competitors access to indispensable information.

2. Data leveraging

Data obtained in one market is used to strengthen another market.

3. Data self-preferencing

A platform uses its data advantage to favour its own products.

4. Data aggregation

Combining datasets may create an advantage competitors cannot replicate.

5. Data exclusion

A dominant undertaking prevents rivals from collecting comparable data.

6. Data-enabled coordination

Competitors use common datasets or algorithms to facilitate coordinated behaviour.

31. Important Case-Law Principles

The Amazon and Meta UK investigations resulted in commitments rather than findings of infringement; this distinction is important when using them as case-law examples. Likewise, the FTC's Facebook case concerns allegations and ongoing litigation rather than a final judicial determination of all allegations.

32. Emerging Issues

A. Artificial Intelligence

AI increases the competitive importance of:

• training datasets;
• compute;
• inference data;
• user feedback;
• model outputs;
• proprietary APIs.

Large datasets may become an important source of AI competitive advantage.

B. Cloud Computing

Competition concerns include:

• data portability;
• interoperability;
• egress costs;
• switching barriers;
• cloud credits;
• technical lock-in.

C. Healthcare

Important issues include:

• electronic health records;
• genomic databases;
• medical imaging;
• pharmaceutical sales data;
• AI diagnostic datasets.

D. Financial Services

Data access can determine competition in:

• credit scoring;
• open banking;
• digital payments;
• insurance;
• fraud detection.

E. Connected Vehicles

Vehicle-generated data may become strategically important for:

• repair markets;
• insurance;
• navigation;
• charging;
• autonomous driving;
• aftermarket services.

33. Challenges for Competition Authorities

1. Rapid technological change

By the time an investigation finishes, the market may have changed substantially.

2. Difficult market definition

Zero-price services and multi-sided platforms complicate conventional analysis.

3. Measuring data value

The quantity of data does not necessarily correspond to its competitive value.

4. Privacy-competition interaction

Authorities must avoid treating privacy law and competition law as interchangeable.

5. Algorithmic opacity

The authority may need technical expertise to understand ranking or pricing systems.

6. Cross-border data flows

Digital businesses operate across multiple jurisdictions.

7. Dynamic competition

Today's dominant dataset may become obsolete through technological change.

34. Principles for Effective Competition Governance

A modern regulatory system should incorporate:

Principle 1 — Data neutrality

Control over data should not automatically translate into control over adjacent markets.

Principle 2 — Contestability

New entrants should have realistic opportunities to compete.

Principle 3 — Interoperability

Technical barriers should not unnecessarily prevent switching or multi-homing.

Principle 4 — Non-discrimination

Platform operators should not unfairly discriminate between their own services and dependent businesses.

Principle 5 — Proportionality

Data-access obligations should not unnecessarily destroy legitimate incentives to innovate.

Principle 6 — Privacy compatibility

Competition remedies involving personal data must comply with applicable privacy law.

Principle 7 — Transparency

Important algorithmic and data-governance decisions should be sufficiently reviewable.

Principle 8 — Dynamic assessment

Authorities should consider innovation and future competition, not merely present market shares.

35. Conclusion

Competition law in data-intensive industries is evolving from a narrow concern with price and output toward a broader analysis of data, algorithms, access, interoperability, privacy, innovation, network effects and ecosystem power.

The major cases demonstrate different dimensions of this development:

• IMS Health demonstrates the exceptional circumstances surrounding access to indispensable data-related structures.
• Microsoft demonstrates the importance of interoperability information.
• Google Shopping demonstrates the competitive significance of algorithmic self-preferencing.
• Facebook/Meta in Germany demonstrates the interaction between data aggregation and dominance.
• Amazon Marketplace demonstrates the risks associated with a platform simultaneously possessing seller data and competing against sellers.
• Meta advertising-data proceedings demonstrate the potential leveraging of data across activities.
• FTC v Facebook/Meta demonstrates the importance of APIs, network effects and acquisitions in digital ecosystems.

The emerging model is therefore:

Competition governance in data-intensive industries is not simply about preventing monopolies; it is about maintaining contestable markets in which data, algorithms and digital infrastructure do not become mechanisms for permanently insulating incumbents from competitive pressure.