Identity Stitching Algorithms And Data Consolidation Power

1. Introduction

Identity stitching algorithms are technological systems that determine that multiple records, devices, accounts, identifiers, or behavioural traces belong to the same underlying person, household, business, or entity. They can combine information such as email addresses, telephone numbers, device identifiers, cookies, IP addresses, advertising IDs, purchase histories, location signals, browsing behaviour, login credentials, loyalty-program records, and inferred attributes.

The resulting identity graph can become a major source of economic power. A firm that can accurately connect otherwise fragmented datasets may obtain a significant informational advantage over competitors that possess only isolated datasets.

From a competition-law perspective, the central concern is therefore not merely the possession of personal data. It is the ability to consolidate, match, enrich, and commercially exploit data across otherwise separate ecosystems.

2. Meaning of Identity Stitching

Identity stitching generally involves five stages:

Data collection → Identifier matching → Entity resolution → Profile enrichment → Commercial exploitation

For example:

Device A → cookie X → email address Y → loyalty account Z → payment record Q → inferred consumer profile P.

The algorithm concludes that these records probably relate to the same person.

Common techniques

1. Deterministic matching
   • Exact email matches
   • Telephone numbers
   • Login credentials
   • Customer IDs
2. Probabilistic matching
   • IP address
   • Device characteristics
   • Behavioural similarity
   • Location patterns
   • Timing of activity
3. Graph-based matching
   • Connections among devices, accounts and users are represented as an identity graph.
4. Machine-learning entity resolution
   • Algorithms assign probabilities to possible identity matches.
5. Cross-device identification
   • Links smartphones, computers, tablets, smart TVs and connected vehicles to one individual or household.

3. What Is Data Consolidation Power?

Data consolidation power exists where a company can combine datasets from multiple sources in a way that competitors cannot easily reproduce.

The relevant competitive advantage may arise from:

• scale of data;
• diversity of data;
• frequency of collection;
• historical depth;
• cross-service integration;
• identity-matching accuracy;
• exclusive access;
• network effects;
• machine-learning feedback loops.

A company may therefore possess relatively ordinary datasets individually but acquire substantial market power by combining them.

Example

A search engine may know:

• search queries.

A social network may know:

• social relationships.

An advertising platform may know:

• browsing behaviour.

A payments platform may know:

• purchasing behaviour.

An identity-stitching system may connect these datasets into a unified behavioural profile.

The competitive significance comes from the combination, rather than necessarily from any individual dataset.

4. Why Identity Stitching Can Create Market Power

A. Data network effects

More users generate more data.

More data improves identity resolution.

Better identity resolution improves targeting and services.

Better services attract more users.

This creates a reinforcing cycle:

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

B. Economies of scope

A dominant platform operating several services can combine datasets across them.

For example:

Search + video + maps + payments + cloud + advertising

may produce substantially richer identity information than a standalone competitor can obtain.

C. Switching costs

Identity consolidation can create substantial switching costs.

A consumer may have accumulated:

• purchase history;
• preferences;
• reputation;
• authentication credentials;
• social connections;
• loyalty status;
• recommendations;
• account history.

Moving to another provider may mean losing this accumulated identity infrastructure.

D. Entry barriers

A new entrant may be unable to reproduce the incumbent's:

• historical data;
• cross-platform identifiers;
• behavioural profiles;
• identity graph;
• matching accuracy;
• advertiser relationships.

Consequently, data consolidation can become an entry barrier even where the underlying data is theoretically obtainable elsewhere.

5. Identity Stitching and Market Definition

Competition authorities may need to ask whether identity resolution constitutes a distinct relevant capability or whether it is merely an input into another market.

Potential markets include:

1. Online advertising

Identity stitching can improve:

• targeting;
• attribution;
• audience segmentation;
• conversion measurement.

2. Ad-tech identity services

Identity graphs themselves may constitute an important technological input.

3. Data analytics

Consolidated identity information can improve predictive analytics.

4. Authentication and identity services

Identity matching can form part of digital identity infrastructure.

5. Platform ecosystems

Identity data may connect multiple services and increase ecosystem-wide dependency.

6. Abuse Theories

A. Leveraging

A dominant platform may use identity information obtained in one market to strengthen its position in another.

Example:

Dominant search engine → identity data → advertising dominance.

This may constitute leveraging where the necessary dominance and exclusionary effects are established.

B. Self-preferencing

A platform controlling identity infrastructure might give its own advertising or analytics services preferential access to identity information.

C. Refusal to provide interoperability

A dominant provider might prevent competitors from accessing:

• identity APIs;
• authentication interfaces;
• portability mechanisms;
• interoperability standards.

The competition issue becomes particularly serious where the identity infrastructure is indispensable for effective competition.

D. Data tying

Access to one service might be conditioned upon allowing the provider to combine data across several services.

This can create concerns under tying or exploitative-abuse theories depending on the jurisdiction.

E. Exclusivity

Platforms may contractually prevent publishers, advertisers or partners from sharing identity signals with competing identity providers.

7. Privacy and Competition Are Interconnected

Identity stitching demonstrates why privacy and competition law increasingly overlap.

A company may offer a service at a nominal monetary price while obtaining extensive data rights.

The competitive concern can involve:

• reduced privacy;
• reduced consumer choice;
• increased surveillance;
• weaker data portability;
• reduced contestability.

A deterioration in privacy can therefore potentially constitute a non-price dimension of competition.

The important qualification is that competition authorities must establish a legally relevant connection between the conduct and competitive harm rather than treating every privacy violation automatically as an antitrust violation.

8. Six Major Case Laws

1. Facebook/Meta – German Facebook Case

Bundeskartellamt, Facebook (B6-22/16, 2019)

This is one of the most important cases for identity stitching and data consolidation.

The German Bundeskartellamt found that Facebook's collection and combination of data from:

• Facebook;
• WhatsApp;
• Instagram;
• third-party websites;

raised competition concerns.

The case was significant because it considered whether a dominant platform could combine data from different sources as a condition of using its social-networking service.

Relevance

It demonstrates that:

Data combination across ecosystems can itself become relevant to the assessment of market power and abusive conduct.

The case is particularly important for identity stitching because cross-service data aggregation can produce a much richer identity profile than information collected within a single service.

2. Bundeskartellamt v Facebook – German Federal Court of Justice

Bundesgerichtshof, KVR 69/19, 23 March 2021

The German Federal Court of Justice upheld the Bundeskartellamt's approach at the preliminary stage concerning Facebook's terms and data combination practices.

The Court accepted that Facebook's market position and its data-processing practices could be assessed together.

Relevance to identity stitching

The case illustrates a crucial principle:

Data-processing conditions imposed by a dominant digital platform can have competitive significance when they reinforce the platform's market position.

Identity stitching can therefore be relevant not simply as a privacy practice but as a mechanism capable of strengthening entrenched dominance.

3. Google Search (Shopping)

Google Search (Shopping), Case AT.39740, General Court, 2021

The European Commission found that Google had abused its dominant position by favouring its own comparison-shopping service in general search results.

The General Court substantially upheld the Commission's finding.

Relevance

Although the case was not directly about identity stitching, it establishes the broader principle that a dominant digital platform can use control over a crucial digital infrastructure to favour its own downstream service.

Identity consolidation can produce a comparable leveraging structure:

Identity infrastructure → information advantage → downstream competitive advantage.

The case therefore helps explain why identity data can become strategically significant when controlled by a vertically integrated platform.

4. Google Android

Google Android, Case AT.40099, European Commission, 2018

The European Commission found several forms of abuse involving Google's Android ecosystem, including tying and contractual restrictions.

Relevance

The Android ecosystem illustrates how control over a technological ecosystem can allow a dominant firm to reinforce its position across related markets.

Identity stitching can operate similarly:

operating system → device identity → account identity → app behaviour → advertising profile.

Thus, control over a foundational identity layer can potentially facilitate expansion into adjacent markets.

5. IMS Health v NDC Health

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

The Court of Justice considered refusal-to-license issues concerning a pharmaceutical-sales data structure.

The case is fundamental to the essential-facilities/refusal-to-supply doctrine.

Relevance

Identity graphs can raise a similar question where a dominant undertaking controls a uniquely valuable data infrastructure.

However, the strict conditions governing refusal-to-supply cases mean that:

mere possession of valuable data does not automatically create an obligation to share it.

The data must satisfy the applicable legal requirements for intervention.

6. Microsoft v Commission

Case T-201/04, Microsoft Corp. v Commission, General Court, 2007

Microsoft concerned interoperability information and Microsoft's control over important software interfaces.

The General Court upheld major aspects of the Commission's abuse finding.

Relevance

Identity stitching increasingly depends upon interoperability among:

• platforms;
• authentication systems;
• APIs;
• devices;
• advertising systems.

Microsoft therefore provides an important analogy for understanding how control over an important technological interface can become a competition concern.

9. Additional Important Case Laws

7. Google and Alphabet – AdSense

Google AdSense for Search, Case AT.40411

The Commission addressed contractual restrictions affecting third-party websites' use of competing search advertising services.

Relevance

It demonstrates how contractual control over data and digital distribution channels can restrict competing advertising systems.

8. Amazon Marketplace

European Commission – Amazon Marketplace investigation

The Commission investigated Amazon's use of marketplace seller data and its competitive relationship with sellers.

Relevance

The case demonstrates the competitive significance of a platform possessing extensive commercially sensitive information generated by users of its ecosystem.

Identity stitching adds another layer because a platform can potentially connect information about:

sellers + customers + transactions + browsing + advertising interactions.

10. Identity Stitching as a Feedback Loop

One of the most significant theoretical concerns is the identity-data feedback loop.

Stage 1

The platform acquires users.

Stage 2

It collects identity signals.

Stage 3

Algorithms connect those signals.

Stage 4

The resulting identity graph improves targeting.

Stage 5

Better targeting attracts advertisers.

Stage 6

More advertisers increase platform revenue.

Stage 7

Revenue finances further acquisition and technological development.

Thus:

Scale → identity data → matching accuracy → monetisation → scale

This can produce dynamic market power even where conventional static market-share analysis initially understates the platform's position.

11. Identity Stitching and Killer Acquisitions

Identity consolidation can also arise through mergers and acquisitions.

A dominant platform may acquire:

• authentication providers;
• advertising identity companies;
• customer-data platforms;
• loyalty platforms;
• analytics companies;
• data brokers.

The acquisition can eliminate an independent identity-data source and allow the acquirer to integrate another dataset into its existing graph.

Competition concern

The authority may therefore ask:

1. Does the target possess uniquely valuable identity data?
2. Can the data be combined with the acquirer's existing datasets?
3. Will the transaction increase matching accuracy?
4. Will rivals lose access to an important identity provider?
5. Will entry barriers increase?
6. Will interoperability decline?

12. Data Consolidation and German Competition Law

German competition law is especially relevant because Section 19a GWB allows intervention concerning certain undertakings of paramount significance across markets.

Large digital ecosystems may have:

• enormous datasets;
• multiple interconnected services;
• ecosystem-wide network effects;
• substantial financial resources;
• technological advantages.

Identity stitching can contribute to this ecosystem-wide significance because it allows information obtained in one market to improve competitive positioning in another.

The Facebook case is therefore particularly important in understanding the interaction between:

dominance + data collection + data combination + ecosystem power.

13. Identity Stitching and Data Portability

Data portability can potentially reduce identity-related lock-in.

Under the GDPR, Article 20 provides a right to data portability in specified circumstances.

From a competition perspective, portability may:

• lower switching costs;
• facilitate multi-homing;
• help new entrants;
• reduce informational advantages;
• improve contestability.

However, portability does not necessarily require a dominant firm to transfer every inferred or proprietary identity graph.

A crucial distinction exists between:

User-provided/observed data

and

Proprietary inferences or derived identity scores.

The latter may present much harder competition-law and regulatory questions.

14. Identity Graphs as Strategic Assets

An identity graph can have several layers:

The more layers a platform successfully integrates, the greater its potential data-consolidation advantage.

15. Competition Risks

1. Entrenchment

Existing dominance becomes harder to challenge.

2. Entry barriers

New firms lack equivalent historical identity data.

3. Cross-market leveraging

Data collected in one market strengthens another market.

4. Reduced multi-homing

Users become increasingly dependent on a single identity.

5. Reduced interoperability

Competitors cannot replicate the identity infrastructure.

6. Surveillance advantages

The dominant firm obtains superior behavioural visibility.

7. Advertising concentration

Better identity resolution can increase advertising effectiveness and reinforce advertiser dependence.

8. Algorithmic discrimination

Highly granular identity profiles may permit differentiated treatment of consumers.

16. Possible Remedies

Competition authorities could consider:

Structural remedies

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

Behavioural remedies

• data silos;
• restrictions on cross-use;
• non-discrimination obligations;
• interoperability;
• API access.

Consumer-oriented remedies

• meaningful consent;
• data portability;
• deletion rights;
• transparent identity-linking practices.

Competition-oriented remedies

• prohibition of tying;
• restrictions on exclusivity;
• interoperability requirements;
• access to essential interfaces where legally justified.

17. Key Legal Test

A competition authority assessing identity stitching should ask:

1. Is the undertaking dominant?

↓

2. Does it control an unusually valuable identity/data infrastructure?

↓

3. Can the data be combined across markets or services?

↓

4. Does identity stitching create a substantial competitive advantage?

↓

5. Are rivals unable to reproduce that advantage reasonably?

↓

6. Is there exclusionary or exploitative conduct?

↓

7. Is there demonstrable harm to competition, innovation, consumer choice, privacy, or contestability?

↓

8. Is a competition-law remedy proportionate and legally available?

18. Important Distinction: Data Volume vs Data Consolidation Power

Possessing a huge dataset does not automatically mean possessing market power.

The critical question is often:

What can the undertaking do with the data that competitors cannot reasonably replicate?

Identity stitching can transform fragmented information into a strategically valuable asset.

Thus:

Raw data ≠ identity graph

and

identity graph ≠ automatically market power

but:

unique identity graph + network effects + ecosystem integration + barriers to replication + exclusionary conduct = potentially significant competition concern.

19. Overall Assessment

Identity stitching algorithms represent a new dimension of digital-market power because they allow firms to transform fragmented information into persistent, cross-platform economic identities.

The most important competition-law concern is not simply that a firm has more data. It is that the firm may obtain an increasingly difficult-to-replicate informational infrastructure capable of:

• lowering customer acquisition costs;
• improving prediction;
• increasing advertising effectiveness;
• strengthening network effects;
• increasing switching costs;
• facilitating cross-market leveraging;
• raising barriers to entry;
• reinforcing ecosystem dominance.

The Facebook data-combination litigation is particularly important because it demonstrates how combining information from multiple services can become relevant to the assessment of dominance and abuse. IMS Health and Microsoft provide important frameworks for thinking about access to indispensable information and interoperability, while Google Shopping and Google Android illustrate how control over digital infrastructure can be leveraged into adjacent markets.

Core proposition

Identity stitching converts fragmented data into an integrated identity infrastructure. Where that infrastructure is difficult for rivals to reproduce and is used to reinforce an already powerful digital ecosystem, identity resolution can become a source—and potentially a mechanism—of durable market power.