Hyper-Personalized Persuasion Ecosystems and Dominance Theory

1. Introduction

Hyper-personalized persuasion ecosystems are digital environments in which platforms use large quantities of personal, behavioural, contextual, transactional, and inferred data to tailor the information, advertising, recommendations, prices, interfaces, and messages presented to individual users.

The competition-law concern is not personalization by itself. The concern arises when a dominant undertaking controls the infrastructure, data, algorithms, attention, and commercial interfaces through which persuasion occurs, thereby enabling it to influence users and business customers in ways that reinforce its market power.

The traditional dominance model asks whether a firm can behave independently of competitors, customers, or consumers. Hyper-personalization complicates this analysis because market power may be exercised through choice architecture, algorithmic ranking, individualized targeting, switching costs, data accumulation, and manipulation of attention, rather than merely through higher prices.

2. Meaning of Hyper-Personalized Persuasion

Hyper-personalization goes beyond ordinary targeted advertising.

A platform may construct an individual-level profile from:

• search history;
• browsing behaviour;
• location;
• purchase history;
• app usage;
• social connections;
• device characteristics;
• demographic information;
• inferred interests;
• emotional or behavioural signals;
• responses to previous advertisements;
• engagement patterns;
• transaction history; and
• predicted willingness to purchase.

AI systems can then determine:

what information a particular user sees, when the user sees it, how it is presented, and what commercial action is most likely to follow.

This creates a persuasion ecosystem, rather than merely an advertising service.

3. From Personalization to Persuasion

The important distinction is:

Personalization → Prediction → Influence → Behavioural steering → Commercial exploitation

For example, a platform may predict that a particular consumer:

1. is likely to purchase a product;
2. is particularly sensitive to scarcity messaging;
3. is unlikely to compare competing products;
4. has a high switching cost;
5. will respond to a particular price;
6. is unlikely to leave the platform; and
7. can therefore be exposed to a highly individualized commercial strategy.

The competitive concern becomes substantially greater when the same platform also controls the marketplace or distribution channel.

4. Dominance Theory

Under competition law, dominance generally concerns the ability of an undertaking to act to a significant extent independently of competitive constraints.

The classic formulation comes from United Brands v Commission, where dominance was understood as a position of economic strength enabling an undertaking to behave to an appreciable extent independently of competitors, customers and consumers.

Hyper-personalization can strengthen this independence through several mechanisms.

A. Data advantage

A dominant platform may possess datasets unavailable to rivals.

B. Feedback loops

More users generate more behavioural data.

More data improve personalization.

Better personalization attracts or retains more users.

More users generate still more data.

This creates a data-network-effect loop.

C. Attention control

The platform may control the user's informational environment.

D. Switching costs

Personalized recommendations, histories, preferences and accumulated profiles may make migration to another platform less attractive.

E. Algorithmic opacity

Competitors may not know how rankings, recommendations or individualized offers are generated.

F. Cross-market leveraging

Data collected in one market may strengthen the platform in another.

5. The Economic Structure of the Problem

A simplified model is:

Users → Data → AI profiling → Personalized persuasion → Transactions → More data → Greater market power

The platform therefore potentially controls:

data + computation + prediction + attention + distribution + transaction

This combination can create a form of dominance that traditional price-based analysis may underestimate.

6. Hyper-Personalization as a Barrier to Entry

A new competitor may technically be able to build a competing application.

But it may lack:

• comparable behavioural data;
• historical interaction data;
• advertiser relationships;
• recommendation infrastructure;
• user trust;
• identity graphs;
• machine-learning feedback;
• installed user base; and
• real-time behavioural signals.

Consequently, the relevant barrier is not merely technological entry.

It is informational and behavioural entry.

A potential entrant might have an equally good algorithm but still perform worse because it does not have comparable data to train and continuously improve that algorithm.

7. Personalization and Consumer Choice

Competition law traditionally assumes that consumers exercise meaningful choice.

Hyper-personalization can challenge that assumption.

Instead of presenting the same marketplace to everyone, a platform can effectively create:

one marketplace for every consumer.

Two consumers searching for the same product may receive:

• different rankings;
• different advertisements;
• different recommendations;
• different promotional messages;
• different bundles;
• different default options; and potentially
• different prices.

This can reduce price transparency and comparability.

8. Dominance Through Choice Architecture

Dominance may be strengthened where the platform determines the architecture through which users make decisions.

Examples include:

• default selections;
• personalized recommendations;
• auto-renewal prompts;
• individualized discounts;
• scarcity messages;
• personalized notifications;
• ranking manipulation;
• customized search results;
• tailored subscription offers; and
• individualized cancellation friction.

The competition-law question becomes:

Is the dominant undertaking merely responding to consumer preferences, or is it using its market power to shape those preferences and restrict competitive alternatives?

9. Self-Preferencing and Hyper-Personalized Ranking

Suppose a dominant marketplace operates both:

1. the marketplace infrastructure; and
2. competing products on that marketplace.

It can potentially use personalization to favour its own products.

The preference need not be identical for every user.

Instead, the algorithm could determine:

“Which competing product should this particular user be prevented from seeing?”

This creates a more sophisticated form of self-preferencing.

Traditional self-preferencing:

Platform → own product ranked higher

Hyper-personalized self-preferencing:

Platform → identify user's preferences → selectively suppress competing alternatives → promote platform's own product

10. Data Advantage and Exploitative Conduct

Hyper-personalization may also raise concerns about exploitation.

A dominant platform could potentially use extensive knowledge about consumers to:

• identify willingness to pay;
• target vulnerable consumers;
• intensify purchasing pressure;
• personalize contractual terms;
• discriminate between consumers;
• manipulate renewal decisions; or
• reduce meaningful comparison.

The competition-law difficulty is distinguishing legitimate personalization from abusive exploitation of market power.

11. Relevant Case Laws

1. United Brands v Commission (Case 27/76)

Principle: Dominance involves a position of economic strength enabling an undertaking to behave independently of competitive pressures.

Relevance

This remains a foundational case for dominance theory.

Applied to hyper-personalized ecosystems, the question becomes whether control over:

• data,
• users,
• attention,
• algorithms, and
• distribution

allows a platform to behave independently of competitors and customers.

The case therefore supplies the conceptual foundation for treating non-price forms of market power as relevant to dominance.

12. Hoffmann-La Roche v Commission (Case 85/76)

The Court emphasized the concept of dominance and identified conduct capable of weakening competitive structures as particularly problematic.

Relevance to hyper-personalization

A dominant platform may use individualized targeting to strengthen customer dependence.

Instead of imposing uniform exclusionary conditions, the platform could potentially employ individualized incentives or recommendations that make competing services progressively less attractive.

The important lesson is that dominance analysis should examine the structure and effects of the conduct, rather than merely asking whether every individual consumer was expressly prevented from switching.

13. Google Search (Shopping) – Google v Commission

The Google Shopping litigation concerned the preferential positioning and display of Google's comparison-shopping service within general search results.

Relevance

This case is highly significant for hyper-personalized persuasion ecosystems because search ranking is fundamentally an information-allocation mechanism.

A dominant intermediary controls what consumers encounter first.

With hyper-personalization, this power becomes even more significant because rankings can potentially be individualized.

The theoretical progression is:

Search dominance → ranking control → personalized ranking → individualized commercial visibility.

Thus, competition law must consider not merely whether an algorithm technically ranks results, but whether the dominant platform uses control over information presentation to distort competition.

14. Google Android – Google and Alphabet v Commission

The Android case involved Google's conduct concerning mobile operating systems, search, browsers, and distribution arrangements.

Relevance

The case demonstrates how dominance can extend across an interconnected digital ecosystem.

Hyper-personalization can intensify the ecosystem effect:

Operating system → user data → search → app usage → behavioural profile → personalized recommendation → advertising → further data.

A dominant undertaking operating across multiple technological layers may therefore possess advantages unavailable to a single-market competitor.

15. Google AdSense – Google and Alphabet v Commission

The AdSense case concerned Google's practices relating to online search advertising intermediation.

Relevance

Advertising intermediaries occupy a crucial position between advertisers and users.

Hyper-personalization increases the importance of that intermediary position because the platform may control:

• user information;
• advertising inventory;
• targeting;
• ranking;
• measurement;
• attribution; and
• advertiser access.

Consequently, a dominant advertising ecosystem can potentially create competitive advantages through control over the entire persuasion chain.

16. Facebook/Meta Data-Combination Proceedings – German Bundeskartellamt

The German competition authority's proceedings concerning Facebook's combination of user data from Facebook and other services are particularly relevant.

The case raised the relationship between:

• market power;
• data collection;
• privacy conditions; and
• exploitative conduct.

Relevance

It demonstrates that data practices can become competition-law relevant when exercised by an undertaking possessing substantial market power.

For hyper-personalization, the implication is important:

The competitive significance of personal data may increase when data aggregation substantially improves the dominant firm's ability to profile and influence users.

Data therefore cannot always be treated as commercially neutral.

17. Amazon Marketplace – European Commission

The European Commission's Amazon proceedings concerning marketplace data and competition between Amazon and independent sellers provide another important analogy.

Relevance

Amazon's dual role as:

1. marketplace operator; and
2. seller,

creates potential conflicts concerning commercially valuable information.

In a hyper-personalized marketplace, the concern could become broader.

A platform could potentially observe:

• consumer preferences;
• seller performance;
• conversion rates;
• product searches;
• demand patterns; and
• individualized purchasing behaviour.

It could then use such information in competing with sellers.

This demonstrates the competitive importance of platform-generated data and intermediary position.

18. Intel v Commission (C-413/14 P)

The Intel litigation is important for the assessment of exclusionary conduct and effects.

Relevance

The case reinforces the importance of examining whether conduct is capable of restricting competition rather than relying mechanically on formal categories.

Applied to hyper-personalization, authorities may need to investigate:

• actual or potential foreclosure;
• duration;
• coverage;
• targeting intensity;
• switching behaviour;
• competitive responses; and
• effects on rivals.

An individualized algorithmic strategy may therefore require an effects-oriented analysis.

19. Theories of Harm

Several theories of harm can arise simultaneously.

1. Exclusionary personalization

The platform gives rivals systematically less visibility.

2. Self-preferencing

The dominant undertaking's products receive personalized advantages.

3. Data leveraging

Data acquired in one market are used to strengthen another market.

4. Exploitative personalization

Consumers are subjected to individually optimized commercial pressure.

5. Switching-cost reinforcement

Personalization makes departure from the ecosystem progressively more costly.

6. Predatory personalization

A dominant undertaking could potentially provide highly individualized incentives designed to eliminate particular competitors or customer segments.

7. Algorithmic foreclosure

Competitors remain formally available but are effectively hidden from relevant consumers.

20. Network Effects and the Persuasion Flywheel

Hyper-personalized platforms may exhibit a distinctive feedback mechanism:

More users

↓

More behavioural data

↓

Better predictions

↓

More effective personalization

↓

Higher engagement/conversion

↓

More advertisers and sellers

↓

More transactions

↓

More data

↓

Greater competitive advantage

This can create a persuasion flywheel.

The concern is that the advantage may become self-reinforcing even without traditional exclusionary contracts.

21. Market Definition Problems

Traditional market definition becomes difficult.

Possible relevant markets include:

• social networking;
• online advertising;
• search;
• digital intermediation;
• recommendation services;
• personal-data services;
• attention markets;
• retail marketplaces; and
• AI personalization services.

A platform may simultaneously participate in several of these.

Moreover, a consumer may pay zero monetary price while providing valuable data and attention.

Therefore:

Zero monetary price does not mean zero economic value.

22. The Role of Non-Price Competition

Competition authorities should examine:

• privacy;
• data portability;
• transparency;
• recommendation quality;
• user autonomy;
• interoperability;
• advertising exposure;
• switching costs;
• algorithmic neutrality; and
• quality of service.

A dominant platform could theoretically worsen one of these dimensions without raising monetary prices.

Thus, a purely price-centric dominance analysis may miss the competitive harm.

23. Privacy and Competition Interactions

Privacy and competition law remain distinct legal regimes, but they can intersect.

Where a dominant undertaking offers consumers a service on increasingly intrusive data-collection terms, authorities may ask whether the deterioration in privacy quality reflects exploitation of market power.

The important analytical question is not:

“Is every privacy violation an antitrust violation?”

It is:

“Does the undertaking's market power enable it to impose materially worse data conditions than competitive constraints would permit?”

24. Dark Patterns and Hyper-Personalization

Dark patterns become particularly powerful when combined with individual-level data.

For example:

Generic dark pattern:

“Continue” button is visually prominent.

Hyper-personalized dark pattern:

The platform predicts that a particular user is likely to cancel and changes the interface specifically to make cancellation less likely.

The latter raises more complex questions concerning:

• intentionality;
• algorithmic optimization;
• consumer vulnerability;
• measurable effects; and
• abuse of dominance.

25. Individualized Pricing

Hyper-personalized persuasion can also interact with personalized pricing.

Suppose an algorithm predicts:

Consumer A has a high willingness to pay.

and

Consumer B is highly price-sensitive.

The platform could potentially present different commercial propositions.

Competition authorities would need to distinguish legitimate price differentiation from conduct that exploits dominance or facilitates exclusion.

The mere existence of personalized pricing is not automatically abusive.

The market power, purpose, mechanism, and competitive effects remain critical.

26. AI Agents and the Next Stage

The emergence of autonomous AI agents could intensify the issue.

Instead of a human directly interacting with a marketplace, an AI agent may:

1. search products;
2. compare prices;
3. evaluate recommendations;
4. negotiate;
5. purchase goods; and
6. learn from previous transactions.

A dominant platform controlling the agent's recommendation infrastructure could influence not only consumers but also the AI systems making decisions on their behalf.

This potentially transforms:

consumer persuasion

into

machine-mediated persuasion.

27. Dominance Through the Personalization Stack

The most significant future concern may involve control of multiple layers:

A firm controlling several layers can potentially create stack dominance.

28. What Competition Authorities Should Examine

A sophisticated investigation should examine:

Market power

• market shares;
• user dependence;
• advertiser dependence;
• data advantages;
• entry barriers.

Data

• scope of data collection;
• uniqueness of datasets;
• data portability;
• data combination;
• data access.

Algorithms

• ranking criteria;
• personalization parameters;
• experimentation;
• targeting rules;
• treatment of competitors.

Consumer effects

• switching;
• search costs;
• price comparison;
• choice diversity;
• privacy quality.

Competitor effects

• visibility;
• conversion;
• customer acquisition;
• access to data;
• interoperability.

29. Possible Remedies

Competition authorities may consider:

Structural remedies

Separation of marketplace and competing commercial operations.

Behavioural remedies

Restrictions on self-preferencing or discriminatory ranking.

Data remedies

Data portability, interoperability or controlled data access.

Transparency remedies

Disclosure of meaningful information concerning ranking and recommendation systems.

Algorithmic auditing

Independent evaluation of potentially discriminatory or exclusionary systems.

Choice remedies

Meaningful alternatives to default recommendations.

Interoperability

Allowing rivals to compete for users without requiring complete migration.

30. Key Legal Principle

The central principle can be stated as follows:

A dominant undertaking should not be permitted to convert informational superiority into durable exclusionary power merely because the mechanism is implemented through personalization algorithms rather than traditional contractual restrictions.

Hyper-personalization does not create a new category of dominance automatically.

Rather, it creates a new mechanism through which existing dominance can be acquired, maintained, leveraged, or abused.

31. Case-Law Synthesis

32. Conclusion

Hyper-personalized persuasion ecosystems represent a significant evolution in dominance theory.

Traditional dominance focuses heavily on control over supply, prices, distribution, and contractual relationships. Digital platforms can exercise market power through something less visible:

control over what each individual user sees, understands, compares, and ultimately chooses.

The combination of data accumulation, AI prediction, behavioural profiling, algorithmic ranking, attention control, and ecosystem integration can create powerful feedback loops.

The central competition-law challenge is therefore to determine when personalization changes from a legitimate improvement in service quality into a mechanism for:

• exclusion;
• self-preferencing;
• leveraging;
• exploitation;
• foreclosure;
• consumer lock-in; or
• reinforcement of durable dominance.

The most important conceptual shift is from “Who controls the price?” to “Who controls the decision environment in which the price and product are chosen?”

That shift is likely to become increasingly important in competition law as AI-driven personalization moves from advertising and recommendations into search, commerce, finance, employment, healthcare, procurement, and autonomous AI-agent decision-making.