Competition Law and Ecosystem Dependency Measurement Techniques

1. Introduction

Ecosystem dependency measurement means identifying and quantifying how strongly consumers, businesses, developers, suppliers, or competitors depend on a particular digital or commercial ecosystem.

Dependency becomes particularly important where an ecosystem has:

strong network effects;

high switching costs;

proprietary data;

interoperability advantages;

dominant infrastructure;

default status;

large installed user bases;

contractual restrictions;

complementary services.

The European Commission's current DMA work expressly treats lock-in, switching costs, interoperability and data portability as important factors in assessing the contestability of digital ecosystems. In June 2026, for example, the Commission preliminarily identified AWS and Azure as potential cloud gatekeepers and referred to their entrenched user bases, lock-in effects, high switching costs and large ecosystems. (Digital Markets Act (DMA))

Core formula

ECOSYSTEM → DEPENDENCY → MEASUREMENT → MARKET POWER → CONDUCT → COMPETITIVE EFFECT → LEGAL ASSESSMENT

2. Meaning of Ecosystem Dependency

Dependency exists where a participant cannot easily operate, compete, or switch without continuing to use a particular ecosystem.

For example:

Developer → App Store → Consumers

A developer may depend on the App Store because consumers are concentrated there.

Similarly:

Business → Cloud Provider → Data + Applications + Infrastructure

A business may depend on a cloud provider because moving applications and data elsewhere is expensive or technically difficult.

Therefore:

Dependency is not simply popularity. It is the degree to which an alternative is realistically available.

3. Why Measure Dependency?

Dependency measurement helps answer several competition-law questions:

Does the ecosystem possess market power?

Can customers realistically switch?

Can competitors enter?

Are business users locked in?

Does the ecosystem control an important gateway?

Can the ecosystem impose disadvantageous terms?

Would users leave if conditions worsened?

Are network effects preventing effective competition?

4. Main Dependency Measurement Techniques

4.1 Switching-Cost Analysis

One of the most direct methods is measuring the cost of leaving the ecosystem.

Costs can include:

financial costs;

technical migration;

data transfer;

retraining;

loss of functionality;

contractual penalties;

loss of customers;

loss of accumulated reputation;

loss of applications;

compatibility problems.

Formula

TOTAL SWITCHING COST = FINANCIAL + TECHNICAL + CONTRACTUAL + DATA + OPPORTUNITY COST

The greater the switching cost, the greater the potential dependency.

5. Switching-Time Measurement

Money is not the only relevant factor.

A business may technically be able to switch but require:

3 months;

12 months;

3 years

to migrate.

Therefore regulators can measure:

Time required to migrate from Ecosystem A → Ecosystem B

Long migration periods may indicate stronger dependency.

6. Data-Portability Measurement

Data portability measures how easily a user can transfer accumulated data to another ecosystem.

Relevant questions include:

Can all data be exported?

Is it machine-readable?

Can historical data be transferred?

Are metadata included?

Can the data be imported by competitors?

Is transfer automatic?

Does transfer involve fees?

Does functionality disappear after transfer?

The DMA specifically requires data portability for end users under Article 6(9), and the Commission has identified improved device-to-device data transfer as an important mechanism for reducing switching barriers between mobile ecosystems. (Digital Markets Act (DMA))

Dependency indicator

Low portability → higher switching cost → potentially greater dependency

7. Multi-Homing Measurement

Multi-homing occurs when users or businesses simultaneously use multiple ecosystems.

Example:

A seller uses:

Amazon;

another marketplace;

its own website.

Measurement

Multi-homing rate = users active on multiple platforms ÷ total relevant users

If:

80% use only Ecosystem A;

20% use A + B,

dependency on A may be greater than where most users actively use several platforms.

However, multi-homing must be analysed carefully because not all multi-homing is equally effective.

8. Single-Homing Rate

The opposite indicator is single-homing.

Formula

Single-homing rate = users relying exclusively on one ecosystem ÷ total users

High single-homing can increase ecosystem power because users have fewer practical alternatives.

For example:

90% single-homing → strong potential dependency

But this remains an economic indicator rather than an automatic finding of unlawful conduct.

9. Customer Concentration

Dependency can also be measured at the business level.

Suppose a seller obtains:

20% of sales from Platform A;

30% from Platform B;

50% from its own website.

The seller is diversified.

But if:

90% of sales come from Platform A,

its dependence may be substantially higher.

Formula

PLATFORM DEPENDENCY = PLATFORM-BASED REVENUE ÷ TOTAL RELEVANT REVENUE

10. Revenue-at-Risk Analysis

A more sophisticated technique measures how much revenue would be lost if access to the ecosystem disappeared.

Example

A developer earns:

₹10 crore total annual revenue

and:

₹7 crore from Ecosystem A

Then:

Revenue dependency = 70%

This can be supplemented by measuring the time required to replace that revenue.

11. User-Dependency Ratio

Another useful measure is the proportion of users who cannot realistically obtain the same service outside the ecosystem.

Formula

DEPENDENT USERS = USERS WITHOUT PRACTICAL ALTERNATIVE ÷ TOTAL RELEVANT USERS

This requires qualitative analysis because an alternative may technically exist but be commercially ineffective.

12. Network-Effect Measurement

Network effects can amplify dependency.

Direct network effect

More users make the service more valuable.

Users ↑ → Value ↑ → Users ↑

Indirect network effect

More users attract developers/sellers.

Users ↑ → Developers ↑ → Applications ↑ → Users ↑

A strong network effect can make switching less attractive because leaving means losing access to the entire network.

13. Ecosystem Completeness Index

Dependency may increase when one ecosystem offers many complementary services.

For example:

Search + Browser + Maps + Cloud + Payments + Advertising + AI

A user may remain because leaving one service means losing integration with many others.

A simplified measurement could be:

ECOSYSTEM COMPLETENESS = NUMBER AND IMPORTANCE OF COMPLEMENTARY SERVICES

This is not a formal legal test but an economic indicator.

14. Interoperability Measurement

Dependency can be measured by asking:

How easily can an external service interact with the ecosystem?

Indicators include:

API availability;

technical compatibility;

functionality;

response times;

access fees;

certification requirements;

data exchange;

hardware access.

The DMA's Article 6(7) specifically seeks to prevent gatekeepers from reserving certain operating-system capabilities exclusively for their own services by requiring access for third parties under the applicable framework. (Digital Markets Act (DMA))

15. API Dependency

An API may become an important ecosystem gateway.

Measure:

percentage of transactions through the API;

alternatives available;

migration cost;

technical compatibility;

API pricing;

rate limits;

access restrictions.

Example

Business → API → Platform

If 95% of the business's transactions depend on the API and there is no practical substitute, API dependency may be substantial.

16. Default-Status Measurement

Default settings can significantly influence dependency.

Measure:

percentage of users retaining the default;

percentage changing the default;

difficulty of changing it;

number of steps required;

frequency of alternative selection.

A default can create behavioural inertia.

Thus:

DEFAULT → USER INERTIA → USAGE → DATA → NETWORK EFFECT → DEPENDENCY

17. Search and Ranking Dependency

For marketplaces and search ecosystems, dependency can be measured through:

traffic originating from the platform;

ranking position;

click-through rate;

conversion rate;

percentage of sales attributable to platform ranking.

Example:

If 85% of a seller's customer acquisition comes from one platform's search ranking, the seller may be highly dependent on that platform.

18. Commission Dependency

A platform's financial terms can also be studied.

Measure:

commission percentage;

effective commission;

advertising charges;

payment fees;

logistics charges;

total platform cost.

Formula

TOTAL PLATFORM BURDEN = COMMISSION + ADVERTISING + PAYMENT + LOGISTICS + OTHER REQUIRED COSTS

A high burden combined with strong dependency can be relevant when analysing exploitation or exclusion.

19. Contractual Dependency

Contracts can create ecosystem dependence through:

exclusivity;

minimum commitments;

termination penalties;

long duration;

automatic renewal;

parity clauses;

non-compete obligations;

restrictions on multi-homing.

Dependency indicator

CONTRACTUAL EXIT COST = TERMINATION COST + REMAINING COMMITMENTS + MIGRATION COST

20. Quality-of-Alternative Analysis

Simply counting alternatives is insufficient.

Suppose there are five competing cloud providers.

If switching from the incumbent requires:

rewriting software;

transferring huge datasets;

losing compatibility;

retraining staff,

those alternatives may not be practically equivalent.

Therefore:

Number of alternatives ≠ effective alternatives.

Competition analysis should examine quality, cost and feasibility of alternatives.

21. Critical Loss and Diversion

Economic techniques can help measure dependency.

Diversion ratio

If a customer leaves Ecosystem A, where does the customer go?

Diversion to B = customers moving from A to B ÷ customers leaving A

High diversion to a particular alternative suggests closer competitive substitution.

Critical-loss analysis

It asks how many customers a platform could lose before a hypothetical price increase became unprofitable.

These tools can assist market definition and competitive-effects analysis.

22. Elasticity Measurement

Price elasticity measures how demand responds to price changes.

Formula

PRICE ELASTICITY = % CHANGE IN QUANTITY DEMANDED ÷ % CHANGE IN PRICE

If users barely reduce usage when price increases, demand may be relatively inelastic.

In an ecosystem, low elasticity may reflect:

lack of alternatives;

high switching costs;

network effects;

strong dependency.

But low elasticity can also have other explanations, so it must be interpreted in context.

23. Dependency Shock Test

A useful practical technique is the shock test.

Ask:

What happens if access to the ecosystem suddenly disappears?

Measure:

lost revenue;

lost users;

lost functionality;

migration time;

replacement costs;

customer losses;

technical disruption.

Formula

DEPENDENCY SHOCK = ECONOMIC LOSS + MIGRATION COST + TIME COST + NETWORK LOSS

The larger the shock, the greater the potential dependency.

24. Case Laws

1. Google Android — C-738/22 P

This is a particularly important modern ecosystem case.

The Court of Justice dismissed Google's appeal in July 2026 and upheld the €4.34 billion Commission fine arising from Google's conduct concerning Android, including pre-installation and licensing arrangements promoting Google Search and Chrome. (curia)

The Court's reasoning expressly recognised that an ecosystem characterised by significant barriers and network effects can make entry or maintenance of an equally efficient competitor practically difficult. (Curia)

Dependency measurement relevance

Relevant indicators include:

Android installed base;

pre-installation;

default status;

user switching;

network effects;

app ecosystem;

access to complementary services.

Principle

Ecosystem dependency can reinforce exclusionary effects when network effects and barriers make effective entry or expansion difficult.

25. Microsoft v Commission — T-201/04

Microsoft involved interoperability between Microsoft's Windows client operating system and work-group server operating systems.

The General Court upheld the Commission's finding concerning Microsoft's refusal to supply interoperability information and the associated exclusionary effects. The case also involved tying Windows with Windows Media Player. (Infocuria)

Dependency measurement

Relevant indicators include:

dependence on Windows;

interoperability requirements;

technical switching costs;

compatibility;

access to protocols.

Principle

Technical dependence can become competition-relevant when control of one ecosystem layer limits effective competition in another.

26. Bronner — C-7/97

Bronner concerned access to a newspaper home-delivery network.

The Court adopted stringent conditions for requiring a dominant undertaking to provide access to infrastructure.

Dependency measurement

The key question was essentially whether there was a realistic alternative.

Relevant indicators include:

availability of alternative distribution networks;

cost of duplication;

technical feasibility;

economic viability.

Principle

Dependency must be genuine and sufficiently strong; mere usefulness is not necessarily enough to establish an obligation to provide access.

27. IMS Health — C-418/01

IMS Health involved access to a pharmaceutical data structure.

The Court established demanding conditions for compulsory access to intellectual-property-protected infrastructure/data structures.

Dependency measurement

Relevant questions include:

Is the input indispensable?

Can competitors develop an alternative?

Is substitution realistically possible?

Would refusal eliminate effective competition?

Principle

Indispensability is stronger than commercial convenience.

This distinction is crucial when measuring ecosystem dependency.

28. Slovak Telekom — C-165/19 P

Slovak Telekom concerned access to telecommunications infrastructure and pricing conditions.

The case illustrates the importance of examining:

infrastructure dependence;

downstream access;

pricing conditions;

alternative infrastructure;

competitive foreclosure.

Dependency technique

A regulator can compare:

cost and feasibility of using incumbent infrastructure

with

cost and feasibility of constructing or accessing alternatives.

The modern EU Article 102 framework continues to treat access and infrastructure cases such as Slovak Telekom as important authorities.

29. Intel — C-413/14 P

Intel concerned rebates granted by a dominant undertaking.

The Court held that where the dominant undertaking submits evidence that its conduct is not capable of restricting competition, the Commission must assess all relevant circumstances in the appropriate cases, including economic factors relevant to foreclosure.

Dependency measurement relevance

The authority can examine:

customer dependence;

percentage of demand covered by rebates;

duration;

market coverage;

rival access;

effective prices.

Principle

High customer dependency can increase the exclusionary potential of commercial incentives, but the competitive effect must still be demonstrated.

30. Post Danmark II — C-23/14

Post Danmark II concerned a dominant postal operator's rebate scheme.

Dependency measurement

Relevant indicators include:

customer loyalty;

proportion of demand subject to rebates;

effective prices;

switching opportunities;

competitor ability to contest customers.

Principle

Customer dependency and rebate structures should be analysed together rather than treating the nominal price alone as decisive.

31. Case-Law Summary Table

32. Dependency Measurement Dashboard

A regulator could construct a multidimensional dashboard:

This is an analytical framework, not a statutory scoring system.

33. Dependency Index

For research purposes, a hypothetical index could be constructed:

EDI = w₁S + w₂D + w₃M + w₄N + w₅I + w₆C

Where:

S = switching cost;

D = data dependency;

M = single/multi-homing;

N = network effects;

I = interoperability dependence;

C = contractual dependence;

w = appropriate analytical weights.

Important

Competition authorities should not automatically convert these variables into a legal score.

The index is useful for:

investigation;

screening;

economic analysis;

market studies;

merger assessment.

34. Dynamic Dependency

Dependency can change over time.

Stage 1

Many alternatives.

↓

Stage 2

Ecosystem gains users.

↓

Stage 3

Network effects strengthen.

↓

Stage 4

Users become dependent.

↓

Stage 5

Switching becomes costly.

↓

Stage 6

Entry becomes harder.

Therefore:

CURRENT MARKET SHARE ≠ COMPLETE MEASURE OF FUTURE DEPENDENCY

35. Dependency and Cloud Ecosystems

Cloud services are an important modern example.

Dependency can be measured through:

data migration costs;

egress charges;

proprietary APIs;

application compatibility;

cloud-specific services;

staff expertise;

contractual commitments;

AI-service integration.

The Commission's 2026 cloud investigation specifically examined interoperability, financial conditions, and contractual/commercial practices, with the stated aim of identifying barriers to switching and strengthening customer choice. (Digital Markets Act (DMA))

36. Dependency and Mobile Ecosystems

Mobile dependency can be measured through:

operating-system share;

app availability;

default settings;

data portability;

application compatibility;

accessory compatibility;

messaging networks;

switching costs.

The Commission's 2026 DMA work identified interoperability and data portability as mechanisms intended to reduce dependence when users move between mobile ecosystems. (Digital Markets Act (DMA))

37. Dependency and Messaging Ecosystems

Messaging is particularly susceptible to network effects.

Formula

Users → Contacts → Network value → Lock-in

The DMA therefore includes interoperability obligations for designated gatekeeper messaging services. The Commission reported in 2026 that Meta's WhatsApp and Messenger were subject to this framework, including staged implementation for one-to-one and group chats. (Digital Markets Act (DMA))

Measurement

number of contacts on platform;

percentage of communications through platform;

availability of interoperable alternatives;

switching cost;

network loss after switching.

38. Dependency vs Dominance

This distinction is essential.

Dependency

Means:

A customer, business or developer relies significantly on an ecosystem.

Dominance

Means:

An undertaking possesses substantial market power under the applicable competition-law framework.

One does not automatically prove the other.

Formula

DEPENDENCY → EVIDENCE

DOMINANCE → LEGAL MARKET POWER

ABUSE → CONDUCT + COMPETITIVE EFFECT

39. Dependency vs Lock-In

They are related but not identical.

Dependency = reliance on the ecosystem.

Lock-in = difficulty of leaving the ecosystem.

An undertaking may depend on a platform because it is highly valuable even when switching is easy.

Conversely, switching may be difficult even where the service itself is not particularly valuable.

Therefore both should be separately measured.

40. Ultra-Short Revision

Ecosystem dependency = degree of reliance on an interconnected platform or infrastructure.

Main indicators:

switching cost;

switching time;

data portability;

single-homing;

multi-homing;

network effects;

revenue concentration;

interoperability;

API dependence;

contractual restrictions.

High dependency can strengthen ecosystem market power.

Dependency is not automatically dominance.

Dominance is not automatically abuse.

Effective alternatives matter more than merely theoretical alternatives.

Google Android demonstrates ecosystem/network dependency.

Microsoft demonstrates technical/interoperability dependence.

Bronner and IMS Health demonstrate the importance of indispensability.

Slovak Telekom demonstrates infrastructure dependence.

Intel and Post Danmark II demonstrate customer/rebate dependence.

DMA tools such as data portability and interoperability directly address certain ecosystem switching barriers. (Digital Markets Act (DMA))

Memory Formula

DEPENDENCY = SWITCHING COST + NETWORK EFFECT + DATA LOCK-IN + SINGLE-HOMING + TECHNICAL/CONTRACTUAL BARRIERS

Final Exam Conclusion

Ecosystem dependency measurement techniques provide competition authorities with a structured way to determine how strongly users, businesses and competitors rely on a particular ecosystem. The most important techniques measure switching costs, switching time, data portability, multi-homing, revenue concentration, interoperability, network effects, contractual restrictions and the availability of realistic alternatives. Cases such as Google Android, Microsoft, Bronner, IMS Health, Slovak Telekom, Intel and Post Danmark II demonstrate different forms of dependency analysis. The central principle is that dependency is an important indicator of competitive power, but it must be connected to the relevant legal test before it can establish dominance, abuse, foreclosure or another competition-law infringement.