Competition Law And Competition Implications Of Recommendation Concentration

1. Introduction

Recommendation concentration refers to a situation in which a limited number of digital platforms, search engines, marketplaces, app stores, social-media services, streaming platforms, or other intermediaries control a substantial portion of the mechanisms through which consumers discover products, services, sellers, applications, content, or information.

In traditional markets, competition often depended on:

physical distribution;

shop location;

advertising;

sales networks;

product availability.

In digital markets, competition increasingly depends on:

search rankings;

recommendation systems;

personalised feeds;

"suggested for you" systems;

product rankings;

app recommendations;

video recommendations;

map rankings;

marketplace visibility;

advertising placement.

Therefore, recommendation systems can become a gateway to consumer attention and market access.

The basic economic chain is:

Data → Algorithm → Recommendation → Consumer Attention → Traffic → Sales → More Data → Better Recommendations

When recommendation power becomes concentrated, the undertaking controlling the recommendation mechanism may acquire significant influence over which businesses succeed in reaching consumers.

2. Meaning Of Recommendation Concentration

Recommendation concentration exists when a relatively small number of undertakings control a substantial proportion of the channels through which consumers receive recommendations.

For example:

one search engine controls a large proportion of search-based discovery;

one marketplace controls product recommendations;

one app store controls application discovery;

one social platform controls content recommendations;

a small number of streaming services control entertainment discovery.

The important point is that recommendation concentration is not necessarily the same as market-share concentration in the underlying product market.

A company may not sell the majority of products itself but may nevertheless control a large proportion of the consumer's pathway to those products.

3. Recommendation As A Competitive Gateway

Modern consumers frequently do not examine every available product.

Instead, they rely upon:

rankings;

recommendations;

personalised suggestions;

search results;

ratings;

automated feeds.

Consequently:

Visibility can become a competitive asset.

If a seller is not recommended, ranked, or displayed prominently, the seller may receive significantly less consumer traffic.

Thus, recommendation systems can operate as a gatekeeper between competitors and consumers.

4. Recommendation Concentration And Competition Law

Competition law does not generally require platforms to recommend every competitor equally.

A platform may legitimately develop algorithms that:

improve relevance;

personalise results;

reduce search costs;

improve consumer experience;

promote high-quality products.

Competition concerns arise where a firm with substantial market power uses control over recommendations to:

favour its own products;

exclude competitors;

discriminate against rivals;

impose unfair access conditions;

manipulate rankings;

tie recommendation access to another service;

exploit business dependence;

facilitate exclusionary strategies.

Therefore:

Recommendation concentration itself is not unlawful; abusive use of concentrated recommendation power may be.

5. Why Recommendations Matter Economically

Recommendations affect several important competitive variables.

A. Consumer attention

Consumers have limited time and attention.

B. Search costs

Recommendations reduce the amount of searching consumers need to perform.

C. Traffic

Higher recommendation placement can produce more visits.

D. Sales

Greater visibility can increase transactions.

E. Data accumulation

More transactions generate more data.

F. Network effects

More users can make the recommendation system more effective.

This produces the following feedback loop:

More Users → More Data → Better Recommendations → More Consumer Engagement → More Sellers → More Transactions → More Data

This can make an incumbent's position increasingly difficult to challenge.

6. Forms Of Recommendation Concentration

6.1 Search Recommendation Concentration

A search engine may determine:

which websites appear first;

which products appear prominently;

which services receive visibility.

This gives the search intermediary significant influence over consumer discovery.

6.2 Marketplace Recommendation Concentration

An online marketplace may determine:

preferred products;

sponsored listings;

recommended sellers;

"best match" results;

default product rankings.

Sellers may become dependent upon the platform's recommendation algorithm.

6.3 App Recommendation Concentration

An app store can influence:

which applications are displayed;

featured applications;

search ranking;

recommendation categories.

This can affect developers' ability to acquire users.

6.4 Social-Media Recommendation Concentration

Algorithms may determine:

which posts appear in feeds;

which accounts are recommended;

which videos become visible;

which creators receive distribution.

This creates an important competitive gateway.

6.5 Streaming Recommendation Concentration

Streaming platforms may determine:

which movies are recommended;

which music is promoted;

which creators receive visibility.

Where a platform also owns content, conflicts of interest may arise.

7. Recommendation Concentration And Market Power

Recommendation control may contribute to market power through:

network effects;

data advantages;

consumer dependence;

switching costs;

economies of scale;

ecosystem integration.

However, recommendation power should not automatically be treated as dominance.

Competition authorities must examine:

relevant market;

market shares;

alternatives;

user behaviour;

switching costs;

barriers to entry;

network effects;

data advantages;

competitive effects.

8. Recommendation Concentration And Self-Preferencing

One of the most significant concerns is self-preferencing.

Suppose a marketplace:

operates a platform;

recommends products to consumers;

sells its own competing products.

The platform may have an incentive to recommend its own products more prominently.

This creates a potential conflict:

Platform operator + Recommendation gatekeeper + Competitor

The competition-law question is whether the recommendation practice amounts to exclusionary conduct under the applicable legal framework.

9. Google Shopping And Recommendation Concentration

Case

Google and Alphabet v Commission, Case C-48/22 P

This is one of the most important authorities for analysing digital recommendation and ranking power.

The case concerned Google's treatment of its comparison-shopping service within general search results.

The European Commission found that Google had favoured its own comparison-shopping service and disadvantaged competing comparison-shopping services.

The Court of Justice upheld the relevant finding of abuse.

Relevance to recommendation concentration

Search rankings are effectively a recommendation mechanism.

Consumers frequently rely heavily upon prominent search results.

Therefore, control over search visibility can affect:

consumer traffic;

competitor access;

advertising opportunities;

commercial success.

Principle

A dominant digital intermediary's control over an important discovery mechanism can raise abuse-of-dominance concerns when it is used in a manner that favours its own service and disadvantages competing services.

10. Microsoft v Commission

Case

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

The case involved Microsoft's conduct concerning interoperability information and tying.

Relevance

Modern recommendation systems are often embedded within larger technological ecosystems.

For example:

Operating system → Search → App store → Recommendation → Consumer

If a dominant undertaking controls several layers, it may potentially use one layer to reinforce another.

Principle

Control over an important technological ecosystem can create opportunities for leveraging and foreclosure of competitors in related markets.

11. Google Android

Case

Google and Alphabet v Commission, Case T-604/18

The case concerned Google's contractual practices concerning the Android ecosystem.

The General Court examined arrangements involving:

app stores;

search services;

mobile devices;

distribution;

ecosystem incentives.

Relevance to recommendation concentration

Mobile ecosystems influence how users:

discover applications;

search for information;

access services;

receive recommendations.

If access to one important gateway is tied to another service, competitors may face disadvantages.

Principle

Digital ecosystems can allow a dominant undertaking to use contractual and technological arrangements to reinforce its position across connected markets.

12. Microsoft Internet Explorer Case

Case

Microsoft Corp. v Commission, T-201/04

The broader Microsoft proceedings are also important for understanding the relationship between dominant technological platforms and adjacent services.

Recommendation relevance

A platform can control the technological environment through which users access competing services.

Where the platform also influences defaults, visibility or user access, competitive opportunities can be affected.

Principle

Control over an important platform can affect the competitive conditions of adjacent markets.

13. Bronner v Mediaprint

Case

Oscar Bronner GmbH & Co KG v Mediaprint, Case C-7/97

The case concerned access to a newspaper home-delivery distribution system.

The Court applied a strict approach to refusal-to-deal/essential-facilities claims.

Relevance

Recommendation systems may become critical gateways for reaching consumers.

However, the fact that competitors consider a platform highly important does not automatically mean the platform has a legal obligation to provide access.

Principle

Competition law does not generally require a dominant undertaking to share every commercially important resource; exceptional conditions apply to compulsory access.

This principle is particularly relevant to arguments that a platform should be required to open its recommendation infrastructure.

14. Magill

Cases

RTE and ITP v Commission, Joined Cases C-241/91 P and C-242/91 P

The cases concerned the refusal to license television programme information.

The Court recognised exceptional circumstances under which refusal to license intellectual property could constitute an abuse.

Relevance

Recommendation systems rely heavily on:

information;

databases;

content;

data;

proprietary technology.

A platform may control valuable information used in creating recommendations.

However, Magill demonstrates that compulsory access to protected information is exceptional.

Principle

Control over information can become a competition issue in exceptional circumstances, but competition law does not automatically require information sharing.

15. IMS Health

Case

IMS Health GmbH & Co OHG v NDC Health GmbH & Co KG, Case C-418/01

IMS Health concerned a protected information structure and refusal to license.

Relevance

Recommendation systems may depend upon proprietary:

datasets;

classification systems;

databases;

information architectures.

The case is relevant when competitors argue that access to a particular information structure is indispensable.

Principle

The exceptional conditions governing compulsory access to protected information must be carefully satisfied.

16. Intel

Case

Intel Corp. v Commission, Case C-413/14 P

Intel concerned conditional rebates and exclusionary effects.

Relevance to recommendation concentration

Platforms may use commercial incentives to influence:

seller participation;

advertising;

distribution;

platform visibility;

exclusivity.

If businesses become dependent on a recommendation platform, contractual incentives may potentially reinforce that dependence.

Principle

Where exclusionary conduct is alleged, its ability to foreclose competition and its actual or potential effects may be relevant to the analysis.

17. Deutsche Telekom

Case

Deutsche Telekom AG v Commission, Case C-280/08 P

The case concerned margin squeeze in telecommunications.

Relevance

Recommendation concentration can become more significant when the undertaking controls several vertical levels:

Infrastructure → Platform → Recommendation → Consumer

Vertical integration may create opportunities to disadvantage downstream rivals.

Principle

A dominant undertaking controlling an important upstream input may face competition-law scrutiny where its conduct restricts effective downstream competition.

18. Slovak Telekom

Cases

Slovak Telekom a.s. v Commission, Cases C-165/19 P and C-166/19 P

The cases concerned access to telecommunications infrastructure and exclusionary conduct.

Relevance

Digital recommendation systems depend upon underlying infrastructure such as:

networks;

data centres;

cloud services;

application programming interfaces;

computing systems.

Infrastructure and recommendation control may therefore interact.

Principle

Control over infrastructure can influence downstream competitive conditions when used in an exclusionary manner.

19. United Brands

Case

United Brands Company v Commission, Case 27/76

United Brands is a foundational EU authority concerning dominance and abuse.

Relevance

Recommendation concentration should ultimately be analysed through the broader concept of market power.

A company controlling recommendations may possess significant influence, but the legal question remains whether it has a dominant position in the relevant market and whether the conduct constitutes abuse.

Principle

Dominance is an economic and legal concept requiring analysis of the relevant market and competitive conditions.

20. Recommendation Concentration And Entry Barriers

A new platform may face a difficult problem.

It needs:

users;

data;

interactions;

feedback;

transactions.

But without users, its recommendation system may initially be less effective.

This produces:

Few users → Little data → Weak recommendations → Low visibility → Few users

Meanwhile, an incumbent may experience:

Many users → Large data → Better recommendations → Greater engagement → More users

This is sometimes described as a data-network-effect feedback loop.

21. Recommendation Concentration And Data Advantage

Recommendation quality depends partly upon data.

Platforms may collect:

clicks;

searches;

purchases;

viewing history;

product preferences;

dwell time;

consumer interactions.

More data may permit more sophisticated predictions.

Therefore:

Recommendation concentration can become data concentration.

Data concentration can subsequently reinforce market power.

22. Recommendation Concentration And Consumer Lock-In

Consumers may become accustomed to a particular platform's recommendations.

Examples include:

personalised shopping;

music recommendations;

video recommendations;

search preferences;

travel recommendations.

Moving to another platform may mean losing:

history;

personalised results;

saved preferences;

reputation;

recommendations.

This can increase switching costs.

23. Recommendation Concentration And Small Businesses

Small businesses increasingly depend upon digital recommendation systems.

For example:

A restaurant may depend on:

search ranking;

map ranking;

food-delivery recommendations.

A seller may depend on:

marketplace ranking;

recommended products;

sponsored placement.

A developer may depend upon:

app-store ranking;

featured applications.

Therefore, recommendation concentration may create commercial dependency.

24. Recommendation Concentration And Foreclosure

Foreclosure occurs when conduct reduces competitors' ability to compete effectively.

Recommendation-based foreclosure may occur where a dominant platform:

systematically demotes competitors;

gives its own products preferential visibility;

imposes discriminatory ranking conditions;

restricts rivals' access to recommendation channels;

combines recommendation access with exclusivity;

uses proprietary data to disadvantage rivals.

The legal analysis must establish more than the mere fact that some competitor receives lower ranking.

25. Recommendation Concentration And Algorithmic Discrimination

Recommendation systems can provide different visibility to different businesses.

For example:

Different treatment is not automatically unlawful.

Competition authorities would need to determine:

whether the firm is dominant;

whether the treatment is discriminatory;

whether it disadvantages competition;

whether objective justification exists;

whether consumers are harmed.

26. Recommendation Concentration And Algorithmic Transparency

A competition authority may need to understand:

ranking criteria;

recommendation inputs;

treatment of sponsored content;

treatment of affiliated products;

changes in algorithms;

effects on competitors.

However, competition law does not necessarily require platforms to disclose every algorithm or trade secret.

The focus should be on determining whether the conduct produces unlawful competitive foreclosure.

27. Recommendation Concentration And Tacit Coordination

Recommendation algorithms may also affect competition between sellers.

Suppose competing firms can observe:

competitor prices;

rankings;

consumer demand;

inventory;

promotions.

Algorithms may rapidly adjust competitive strategies.

This can potentially increase the risk of coordination.

However:

Similar algorithmic outcomes do not by themselves establish an illegal cartel.

There must be a legally relevant basis for liability under the applicable competition law.

28. Recommendation Concentration And Advertising

Recommendation systems and advertising systems are increasingly interconnected.

A platform may control:

organic ranking;

paid placement;

sponsored recommendations;

advertising auctions;

consumer data.

This creates possible conflicts of interest when the platform simultaneously acts as:

intermediary;

recommender;

advertiser;

seller;

data controller.

Competition concerns may arise if these roles are used to disadvantage competing businesses.

29. Recommendation Concentration And Vertical Integration

Vertical integration may produce efficiencies.

For example:

Marketplace + Recommendation System + Logistics

may reduce costs.

But vertical integration may also create incentives to favour affiliated businesses.

Competition analysis should therefore distinguish:

Legitimate integration

Integration improves:

efficiency;

quality;

delivery;

consumer experience.

Potential foreclosure

Integration is used to:

disadvantage rivals;

restrict access;

favour affiliated products;

raise rivals' costs.

30. Recommendation Concentration And Merger Control

Recommendation concentration can increase through mergers and acquisitions.

A major platform may acquire:

recommendation technology;

AI startups;

data analytics companies;

search businesses;

content platforms;

marketplace competitors.

Competition authorities may consider whether the transaction:

increases data concentration;

eliminates an emerging competitor;

strengthens network effects;

combines complementary datasets;

reinforces recommendation power.

31. Killer Acquisitions

Recommendation platforms may identify emerging competitors through their data.

A large platform may therefore identify a startup that is developing:

a new recommendation technology;

a competing search service;

an alternative marketplace;

an AI discovery tool.

Acquiring such a business before it becomes a major competitor may reduce future competitive pressure.

This is why merger analysis may need to consider potential competition, not merely current market shares.

32. Recommendation Concentration And Innovation

Recommendation concentration can have two opposite effects.

Positive effects

It can:

reward high-quality products;

reduce search costs;

help innovative products find consumers;

reduce advertising expenses;

improve matching.

Negative effects

If recommendations are controlled by dominant intermediaries, they may:

favour established businesses;

disadvantage new entrants;

reduce experimentation;

discourage innovation;

make market access dependent on platform algorithms.

Thus competition law should consider dynamic competition.

33. Recommendation Concentration And Consumer Welfare

Consumer welfare in recommendation markets is not limited to price.

Relevant factors include:

quality;

choice;

innovation;

privacy;

convenience;

accuracy;

diversity of suppliers.

A recommendation system may be free for consumers but still affect competition through reduced:

choice;

quality;

innovation.

34. India: Competition-Law Framework

In India, the Competition Act, 2002 provides the principal framework.

Section 3

Section 3 concerns anti-competitive agreements.

Recommendation algorithms may become relevant where competing undertakings use them to facilitate:

price coordination;

market allocation;

information exchange;

other prohibited coordination.

Section 4

Section 4 addresses abuse of dominant position.

Potential recommendation-related conduct can involve:

discriminatory conditions;

denial of market access;

leveraging;

tying;

exclusionary conduct.

Section 5

Section 5 concerns combinations.

Mergers involving platforms, data, recommendation technology and emerging competitors may therefore attract competition scrutiny where statutory thresholds and other requirements are met.

35. Recommendation Concentration And Market Access Under Indian Law

For many digital businesses, access to consumers is the critical competitive resource.

A platform can potentially become a gateway through which businesses obtain:

customers;

traffic;

visibility;

transactions.

Where a dominant platform controls this gateway, conduct that denies or restricts meaningful market access may become relevant under the abuse-of-dominance framework.

36. Recommendation Concentration And Essential Facilities

A difficult legal question is whether a dominant recommendation platform should be required to provide competitors with access to its recommendation system.

The answer is not automatically yes.

Cases such as:

Bronner;

Magill;

IMS Health

demonstrate that compulsory access is generally exceptional.

The claimant may need to demonstrate the legally required conditions, including circumstances relating to indispensability and the competitive consequences of denial.

37. Recommendation Concentration And Self-Preferencing: Legal Test

A useful analytical framework is:

Step 1

Is the undertaking dominant?

Step 2

Does it control an important recommendation or discovery channel?

Step 3

Does it favour its own products or services?

Step 4

Are competitors disadvantaged?

Step 5

Is there actual or potential foreclosure?

Step 6

Are there legitimate objective justifications or efficiencies?

Step 7

Is the conduct proportionate to those objectives?

This avoids the simplistic assumption that every recommendation preference is unlawful.

38. Recommendation Concentration And Consumer Dependence

A business may become economically dependent on a recommendation platform even without a formal exclusive contract.

For example:

Platform → Ranking → Customer Traffic → Sales

If the platform controls a substantial portion of customer traffic, a change in ranking may significantly affect the business.

This raises questions concerning:

bargaining power;

platform dependency;

access;

discrimination;

transparency.

39. Recommendation Concentration And Competition Between Platforms

Recommendation concentration can also create competition between ecosystems.

For example:

Platform A: search + shopping + advertising + recommendations

versus

Platform B: search + shopping + advertising + recommendations.

If one ecosystem becomes substantially larger, network effects may make it increasingly difficult for another platform to compete.

Competition authorities may therefore examine:

multi-homing;

switching costs;

interoperability;

data portability;

ecosystem effects.

40. Key Distinction: Recommendation Quality Vs Recommendation Manipulation

This distinction is important.

Legitimate recommendation optimisation

A platform changes its algorithm to:

improve relevance;

improve quality;

reduce spam;

increase user satisfaction.

Potentially problematic manipulation

A dominant platform deliberately changes recommendations to:

disadvantage competitors;

favour its own services;

exclude emerging rivals;

punish businesses that use competing services.

The legal assessment depends upon the facts, market power and applicable competition rules.

41. Competition Implications In Summary

42. Ten Important Case Laws At A Glance

43. Most Important Case: Google Shopping

For examination purposes, Google Shopping is particularly important because it demonstrates how a digital intermediary's control over a major consumer-discovery mechanism can become a competition-law issue.

The case illustrates the connection:

Search Engine → Ranking → Visibility → Consumer Traffic → Competitive Opportunity

This is very close to the economic structure of recommendation concentration.

However, the case should not be reduced to the proposition that "algorithmic ranking is unlawful." The legal finding depended upon the specific circumstances, Google's dominance, the nature of its practices and their effects on competition.

44. Legal Principles To Remember

Principle 1

Recommendation concentration is not itself illegal.

Principle 2

A recommendation system can become an important gateway to market access.

Principle 3

Data and network effects can reinforce recommendation power.

Principle 4

Self-preferencing requires careful competition-law analysis; it is not automatically unlawful in every jurisdiction.

Principle 5

Refusal to provide access to recommendation infrastructure is not automatically an abuse.

Principle 6

Algorithmic parallel behaviour is not automatically a cartel.

Principle 7

Competition law protects the competitive process, not every individual competitor's visibility.

Principle 8

Innovation and better recommendation technology may legitimately produce market success.

Principle 9

The important issue is whether market power is used to artificially restrict effective competition.

45. Short Exam Answer

Recommendation concentration refers to a situation where a small number of digital platforms control a substantial portion of the recommendation, ranking and discovery mechanisms through which consumers find products, services, sellers or content.

It has important competition-law implications because recommendations influence consumer attention, traffic and market access. Concentrated recommendation power can be reinforced by data advantages, network effects, switching costs and ecosystem integration.

Potential competition concerns include self-preferencing, discriminatory ranking, exclusionary conduct, tying, bundling, foreclosure, refusal to provide access, algorithmic coordination and anti-competitive acquisitions. However, recommendation concentration itself is not unlawful. A competition-law violation generally requires the relevant legal conditions concerning dominance, anti-competitive agreement, merger control or exclusionary conduct to be established.

Important cases include United Brands, Bronner, Magill, IMS Health, Microsoft, Intel, Google Shopping, Google Android, Deutsche Telekom and Slovak Telekom.

The most directly relevant authority is Google Shopping, because it demonstrates how control over digital search and ranking can affect competitors' access to consumers.

46. Conclusion

Recommendation systems are becoming one of the most important forms of digital market infrastructure.

In traditional markets:

Distribution controlled access to consumers.

In digital markets:

Recommendation can control consumer attention and discovery.

This creates a potentially powerful competitive resource.

The central economic cycle is:

Data → Algorithm → Recommendation → Visibility → Consumer Traffic → Sales → More Data → Better Recommendations → Greater Market Power

When this cycle results from innovation and better service, it can be entirely legitimate and pro-competitive.

The competition-law concern arises where a dominant undertaking uses recommendation power to:

favour its own products;

disadvantage rivals;

restrict market access;

impose exclusionary conditions;

reinforce ecosystem dominance;

prevent effective entry;

reduce innovation or consumer choice.

Accordingly, recommendation concentration should be analysed not simply by asking who controls the algorithm, but by asking how that control affects market access, competitive opportunities, entry, innovation and consumer choice.

Quick Revision Formula

Recommendation Concentration → Data → Algorithms → Consumer Attention → Visibility → Market Access → Network Effects → Market Power → Potential Self-Preferencing/Foreclosure → Competition-Law Scrutiny

One-Line Revision Point

Recommendation concentration becomes a competition-law concern when concentrated control over digital discovery gives a firm durable market power that is used to restrict effective competition, rather than merely reflecting legitimate innovation or superior service.