Alternative Data Scoring Platforms: Competition Issues

Introduction

Alternative data scoring platforms use non-traditional datasets—such as transaction histories, mobile-phone activity, e-commerce behaviour, utility payments, geolocation, social-media activity, device information, app usage, employment indicators and other behavioural data—to generate credit, risk, fraud, insurance, investment, marketing or eligibility scores.

From a competition-law perspective, the central concern is not simply the possession of data. The problem arises where a platform's control over unique datasets, scoring algorithms, data-processing infrastructure, distribution channels or feedback loops allows it to restrict competitors, disadvantage customers, engage in self-preferencing, impose unfair conditions, or make entry practically difficult.

The principal competition issues can arise under abuse-of-dominance rules, merger control, restrictive-agreement provisions, essential-facility principles, unfair trading conditions and digital-platform regulation.

1. Relevant Market

An alternative-data scoring ecosystem can contain several potentially distinct markets:

1. Alternative-data collection
2. Data aggregation and enrichment
3. Data analytics
4. Scoring/risk-assessment services
5. Credit-information services
6. Fraud-detection and identity-verification services
7. Insurance-risk scoring
8. Investment and financial analytics
9. Platform/API access to scoring information

A competition authority should therefore avoid automatically treating "data" as one single market.

For example, the European Commission's examination of S&P Global/IHS Markit considered particular financial-data and analytics products separately, including company credit-risk analytics data.

2. Data Concentration and Market Power

The first major issue is concentration of unique data.

An alternative-data scoring platform may possess:

• millions of consumer records;
• transaction histories;
• behavioural profiles;
• repayment information;
• purchasing patterns;
• device-level information;
• real-time behavioural signals;
• proprietary risk models;
• historical scoring outcomes.

The competitive advantage may become self-reinforcing:

More users → more data → better model → better scores → more customers → more data.

This creates a data-network effect.

A new entrant may technically be able to create a scoring algorithm but still be unable to reproduce the incumbent's dataset.

The Google cases demonstrate the broader competition-law significance of data-driven advantages. In Google Shopping, the General Court recognised the importance of Google's general search infrastructure and the difficulty competitors could face where access to traffic could not effectively be replaced.

3. Data as a Barrier to Entry

Traditional barriers to entry include:

• capital requirements;
• intellectual property;
• distribution networks;
• regulation.

Alternative-data markets add another barrier:

access to sufficiently large and sufficiently differentiated datasets.

A scoring competitor may need years of historical information before its model becomes commercially competitive.

The incumbent can consequently possess a data advantage that cannot easily be replicated through investment alone.

This is particularly important where the data are:

• proprietary;
• continuously generated;
• highly granular;
• difficult to purchase;
• protected by contractual restrictions;
• obtained through an ecosystem;
• subject to strong network effects.

4. Data Exclusivity Agreements

A dominant scoring platform might contract with:

• banks;
• retailers;
• telecom companies;
• payment processors;
• insurers;
• employers;
• online marketplaces.

It may require these businesses to provide data exclusively to the platform.

Such arrangements can create foreclosure.

Example

Suppose Platform A obtains exclusive access to five years of transaction data from a major payment network.

Platform B cannot obtain equivalent information.

If Platform A subsequently becomes the principal provider of alternative credit scores, the exclusivity arrangement may prevent Platform B from competing effectively.

Competition authorities would examine:

• duration;
• market coverage;
• exclusivity percentage;
• availability of alternative datasets;
• switching possibilities;
• foreclosure effects;
• legitimate commercial justification.

5. Refusal to Provide Data or API Access

Another major issue is data-access foreclosure.

A dominant scoring platform may refuse competitors access to:

• APIs;
• historical data;
• risk scores;
• verification information;
• data feeds;
• interoperability interfaces.

The legal question is whether the requested information constitutes an input whose denial can substantially eliminate effective competition.

The classic European authority is:

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

The Court of Justice held that refusal to license protected information can constitute abuse in exceptional circumstances, including where refusal prevents the emergence of a new product, lacks objective justification and eliminates competition.

This is highly relevant to alternative-data platforms because a proprietary dataset may become competitively significant where rivals cannot realistically reproduce it.

6. Self-Preferencing

A platform may simultaneously:

1. provide scoring services to third parties; and
2. compete against those same customers.

This creates a potential conflict of interest.

For example:

Data platform → provides consumer scores to lenders

while simultaneously operating:

Data platform → its own lending marketplace.

It could theoretically:

• provide better scores to its own lending operation;
• delay competitors' access;
• rank its own financial products higher;
• provide competitors with inferior API functionality;
• use competitors' data to improve its own competing service.

The Google Shopping judgment is an important analogy because the General Court upheld findings concerning preferential treatment of Google's own specialised service over competing services.

7. Algorithmic Discrimination Against Competitors

A scoring platform controls the algorithm through which users and businesses are evaluated.

Potentially problematic practices include:

• reducing competitors' visibility;
• systematically lowering competitors' scores;
• altering ranking criteria;
• imposing discriminatory access conditions;
• withholding high-quality data;
• giving the platform's affiliated businesses preferential treatment.

The issue becomes especially serious when the scoring algorithm is simultaneously a market-access mechanism.

In such circumstances, the algorithm may function like a private regulatory system.

8. Data Tying and Bundling

A dominant platform could condition access to one service upon purchasing another.

For example:

"To obtain our alternative credit score, you must also purchase our fraud-detection service."

Or:

"Access to our transaction dataset is available only if you use our payment-processing service."

Potential competition concerns include:

• leveraging dominance from one market into another;
• foreclosure of specialist competitors;
• increased switching costs;
• reduced interoperability.

The assessment would depend upon market power, market definition, foreclosure, indispensability and possible efficiencies.

9. Excessive or Unfair Data-Access Conditions

Competition concerns are not limited to monetary prices.

A dominant platform may impose:

• excessive licensing fees;
• discriminatory API charges;
• unreasonable minimum-volume requirements;
• restrictive data-use licences;
• non-compete provisions;
• excessive audit requirements;
• contractual restrictions preventing customers from using competing scoring services.

The quality and terms of access can therefore be competitively important even where the nominal monetary price is zero.

The European Commission's approach to digital markets recognises that free services can nevertheless involve dominance and significant barriers to switching and entry.

10. Exploitative Data Collection

An alternative-data platform may collect extensive personal information because users have little practical alternative.

This creates a connection between data protection and competition law.

A major case is:

Facebook/Meta – German Competition Authority, B6-22/16

The Bundeskartellamt found that Facebook's dominant position and its ability to combine data from Facebook, WhatsApp, Instagram and third-party websites created competition concerns. The authority regarded the imposition of extensive data-combination conditions as exploitative abuse and linked the data advantage to competitors' ability to compete.

This is particularly relevant to alternative scoring because combining datasets can produce significantly more valuable behavioural profiles than any individual dataset.

11. Data Feedback Loops

Alternative-data scoring platforms can create a powerful feedback loop:

More customers

↓

More behavioural data

↓

More accurate scoring

↓

More commercial adoption

↓

More transactions

↓

Even more data

This can produce an entrenched competitive advantage.

The competition concern becomes stronger where competitors cannot obtain equivalent data even if they offer superior algorithms.

12. Switching Costs and Lock-In

Businesses using an alternative scoring platform may become dependent upon:

• proprietary APIs;
• historical scores;
• proprietary risk categories;
• integrated software;
• customer databases;
• automated underwriting systems;
• proprietary documentation.

Switching can then involve:

• rebuilding models;
• recalibrating historical scores;
• regulatory validation;
• customer migration;
• technical integration;
• loss of historical information.

Consequently, technical switching costs can operate as barriers to entry and expansion.

13. Merger-Control Concerns

A particularly important issue is the acquisition of one data-rich company by another.

Consider:

Major scoring platform + major alternative-data provider

The combined firm could obtain:

• more datasets;
• better predictive models;
• broader customer access;
• stronger network effects;
• greater vertical integration.

Authorities may therefore investigate:

Horizontal effects

Both companies provide scoring services.

Vertical effects

One provides data while the other provides scoring infrastructure.

Conglomerate effects

The merged company can bundle unrelated data and analytics products.

Data-driven effects

The merged database becomes difficult for competitors to reproduce.

14. S&P Global/IHS Markit

United States v S&P Global Inc. et al.

The S&P Global/IHS Markit transaction is a particularly useful modern precedent for data-intensive financial markets.

The U.S. Department of Justice required divestitures addressing competition concerns, including businesses providing commodity price-reporting information.

The European Commission likewise examined data and analytics markets in the transaction, including company credit-risk analytics and discrete financial-data content sets.

The case demonstrates that specialised information assets can themselves be central to merger analysis.

15. Six Key Case Laws

1. IMS Health GmbH & Co. OHG v NDC Health GmbH & Co. KG

Case C-418/01, CJEU, 2004

Principle

A dominant undertaking's refusal to license a competitively indispensable information asset may constitute abuse in exceptional circumstances.

Relevance

A proprietary alternative-data dataset could raise analogous issues where:

• competitors cannot realistically reproduce it;
• access is indispensable;
• refusal excludes competition;
• a new or improved service depends upon the information.

2. Google and Alphabet v European Commission (Google Shopping)

Case T-612/17, General Court, 2021

Principle

A dominant digital platform can abuse its position by favouring its own specialised service through preferential ranking and display.

Relevance

An alternative-data platform could potentially discriminate between:

• its own scoring products; and
• rival scoring products.

The case is particularly relevant to algorithmic self-preferencing and access to platform infrastructure.

3. Facebook/Meta – Bundeskartellamt

B6-22/16, Germany, 2019

Principle

The German competition authority examined extensive combination of data from Facebook, affiliated services and third-party sources as an abuse of dominance.

Relevance

It demonstrates how data accumulation itself can have competitive significance, particularly where data advantages reinforce market power.

4. S&P Global/IHS Markit

Principle

Financial-information and analytics markets can contain highly specialised datasets that require separate competition analysis.

The authorities examined price assessments, financial information and credit-risk analytics and imposed/accepted structural remedies concerning particular information businesses.

Relevance

This is directly analogous to an alternative-data scoring platform whose database contains unique financial or behavioural information.

5. Nielsen Consumer LLC v Circana Group / NPD-IRI

Principle

The dispute illustrates the competitive significance of proprietary consumer-information datasets and the relationship between data-sharing arrangements, confidentiality and competition.

NPD and IRI operated in consumer analytics, while Nielsen was a major competitor; the litigation concerned disclosure of protected data following the NPD-IRI transaction.

Relevance

It illustrates how data ownership, confidentiality and consolidation can become strategically important in information-intensive markets.

6. Google – Search Data Access under the Digital Markets Act

The European Commission's 2026 proceedings concerning Google's search-data access provide an important modern regulatory development.

The Commission required Google to implement measures concerning access to anonymised search-ranking, query, click and view data under Article 6(11) DMA.

Relevance

Although this is a DMA proceeding rather than a traditional Article 102 judgment, it demonstrates the emerging regulatory recognition that access to large-scale data can be necessary to maintain competitive conditions in digital markets.

16. Additional Relevant Authorities

Several broader competition cases are also useful by analogy:

United Brands v Commission

Case 27/76

Relevant to dominance, barriers to entry and the special responsibilities of dominant firms.

Hoffmann-La Roche v Commission

Case 85/76

Important for exclusionary conduct and the special responsibility of dominant undertakings.

Slovak Telekom v Commission

Joined Cases C-165/19 P and C-166/19 P

Relevant to exclusionary access practices and the relationship between dominant infrastructure and downstream competition.

Bronner v Mediaprint

Case C-7/97

Important for determining when refusal of access to an infrastructure or facility can amount to abuse.

Microsoft v Commission

Case T-201/04

Important for interoperability, refusal to supply information and leveraging of technological dominance.

These authorities can be used to construct the legal framework for data-access and interoperability disputes even where the factual technology differs.

17. Algorithmic Collusion Risks

Alternative-data platforms can also create risks of coordinated behaviour.

Suppose several competitors purchase the same scoring service.

The platform's algorithm continuously observes:

• prices;
• demand;
• customer behaviour;
• transactions;
• competitor activity.

If the scoring platform supplies identical optimisation recommendations to competing businesses, the technology could potentially facilitate:

• price alignment;
• customer allocation;
• output coordination;
• risk-selection coordination;
• common pricing parameters.

The competition-law question is whether the conduct reflects independent algorithmic optimisation or some form of communication, coordination or facilitation among competitors.

18. Common Data Standards and Homogenisation

A further concern is algorithmic homogenisation.

If virtually every lender uses the same:

• dataset;
• scoring methodology;
• risk thresholds;
• predictive variables;

then competitive differentiation may decline.

For example:

Platform score = 780

could become the standard market signal.

If virtually all lenders rely on the same platform, competitors may cease developing independent risk-assessment models.

This may reduce:

• innovation;
• methodological diversity;
• price competition;
• product differentiation.

19. Predatory Data Acquisition

A powerful platform might acquire data-producing companies not primarily for their customers but for their datasets.

This raises questions about:

• nascent competition;
• killer acquisitions;
• elimination of potential competitors;
• accumulation of unique datasets;
• data-driven entry barriers.

Merger authorities therefore increasingly consider whether an apparently small acquisition has strategic data significance.

20. Data Portability and Interoperability

Competition may be improved by allowing businesses to transfer:

• historical scoring data;
• transaction information;
• customer records;
• model outputs;
• risk histories.

Interoperability can reduce:

• switching costs;
• dependency;
• lock-in;
• entry barriers.

The European Commission's 2026 Google Search-data proceedings provide a contemporary example of regulatory intervention designed to facilitate data access on fair, reasonable and non-discriminatory terms.

21. Essential-Facility Analysis

A proprietary dataset is not automatically an essential facility.

The exceptional nature of the doctrine means authorities normally examine:

1. Whether the undertaking is dominant;
2. Whether the information is genuinely indispensable;
3. Whether competitors can reasonably reproduce it;
4. Whether refusal eliminates effective competition;
5. Whether access is technically feasible;
6. Whether refusal has an objective justification;
7. Whether access can be provided without disproportionate harm.

IMS Health and Bronner are therefore particularly important when analysing this issue.

22. Competition Concerns in India

For an Indian alternative-data scoring platform, the principal framework would potentially include the Competition Act, 2002, particularly:

• Section 4 – abuse of dominant position;
• Section 5 – combinations;
• Section 6 – regulation of combinations;
• Section 3 – anti-competitive agreements.

Potential Section 4 concerns include:

Denial of market access

Refusing competitors access to important data/API infrastructure.

Discriminatory conditions

Providing different data quality or access terms to competing businesses.

Unfair conditions

Imposing unreasonable contractual restrictions on customers.

Leveraging

Using dominance in data aggregation to enter scoring, lending or insurance markets.

Predatory conduct

Using revenue from a data monopoly to subsidise exclusionary downstream services.

Self-preferencing

Giving the platform's own financial or commercial products preferential treatment.

23. Competition-Law Risk Matrix

24. Possible Competition Remedies

Competition authorities may consider:

Structural remedies

• divestiture of datasets or businesses;
• separation of data assets;
• prohibition of certain acquisitions.

Behavioural remedies

• FRAND access;
• non-discriminatory APIs;
• data portability;
• interoperability;
• prohibition of self-preferencing;
• restrictions on exclusive agreements.

Transparency remedies

• explanation of scoring criteria;
• audit mechanisms;
• independent monitoring.

Data remedies

• controlled data access;
• anonymised datasets;
• API access;
• interoperability standards.

The appropriate remedy depends on whether the competitive harm arises primarily from concentration, exclusion, discriminatory access, contractual restrictions or algorithmic conduct.

Conclusion

The competition-law problem surrounding alternative-data scoring platforms is fundamentally the interaction of data concentration + algorithmic capability + network effects + platform dependency.

The most important risks are:

1. Data-driven dominance
2. Exclusive acquisition of alternative data
3. Refusal or discriminatory provision of data/API access
4. Self-preferencing
5. Data tying and bundling
6. Algorithmic foreclosure
7. Customer and competitor lock-in
8. Data-driven mergers
9. Common-algorithm coordination
10. Exploitation of consumers through excessive data collection

The cases of IMS Health, Google Shopping, Facebook/Meta and S&P Global/IHS Markit are particularly useful because they demonstrate different legal approaches to indispensable information, digital self-preferencing, accumulation of user data and concentration of specialised information assets. The newer Google Search-data proceedings further illustrate the movement toward explicit regulatory obligations concerning access to competitively significant data.