Digital Ecosystem Migration Costs And Lock-In Quantification

Introduction

Digital ecosystem migration costs are the economic, technical, contractual, behavioural, and informational costs incurred when a user, business, developer, advertiser, or supplier moves from one digital ecosystem to another. Lock-in exists where these costs become sufficiently substantial that switching becomes commercially unattractive, even when competing alternatives are available.

In competition law, the central issue is not simply whether switching is expensive. The question is whether migration costs materially reduce competitive pressure on an incumbent and thereby enable it to maintain or strengthen market power.

Digital ecosystems can create unusually strong lock-in because several dependencies operate simultaneously:

• data accumulation;
• interoperability and API dependencies;
• cloud infrastructure;
• application compatibility;
• identity and authentication systems;
• accumulated reputation and ratings;
• subscriptions and digital purchases;
• developer tools and software-development kits;
• advertising histories;
• network effects;
• learning effects and user familiarity;
• contractual commitments; and
• ecosystem-specific hardware or complementary products.

Thus, migration should be analysed as a multi-dimensional economic cost, rather than as a simple monetary switching fee.

1. Meaning of Digital Ecosystem Migration Costs

Migration costs are the costs associated with transferring economic activity from Ecosystem A to Ecosystem B.

A simplified formulation is:

MC=F+T+D+C+B+N+L+RMC = F + T + D + C + B + N + L + R

Where:

• F = financial switching costs;
• T = technical migration costs;
• D = data-transfer and data-reconstruction costs;
• C = contractual costs;
• B = behavioural/learning costs;
• N = network-related costs;
• L = loss of accumulated ecosystem benefits;
• R = reputational or relationship costs.

The important point is that many of these costs are implicit rather than directly invoiced.

For example, moving from one cloud provider to another may technically be possible at low contractual cost, but the customer may still face:

• application rewriting;
• database conversion;
• employee retraining;
• API modification;
• security reconfiguration;
• downtime;
• loss of historical analytics;
• changes to monitoring systems; and
• renegotiation with customers and suppliers.

The aggregate migration cost may therefore be much greater than the provider's formal termination fee.

2. Lock-In Versus Ordinary Switching Costs

Not every switching cost constitutes problematic lock-in.

Ordinary switching cost

A consumer may spend a few hours learning a competing application.

Strong lock-in

A business may need months of engineering work to migrate databases, rewrite applications and retrain personnel.

Ecosystem lock-in

The strongest form occurs where leaving one service requires abandoning a cluster of interconnected products.

For example:

Smartphone → operating system → app purchases → cloud storage → payment account → messaging → wearable → identity credentials.

The user may technically be able to switch the smartphone, but switching the entire ecosystem is substantially more expensive.

Competition authorities therefore increasingly need to examine ecosystem-level switching costs, rather than analysing individual products in isolation.

3. Principal Sources of Digital Lock-In

A. Data lock-in

Users may accumulate:

• photographs;
• transaction histories;
• customer records;
• search histories;
• playlists;
• professional contacts;
• advertising data;
• health or fitness records;
• machine-learning datasets.

If data portability is technically limited, the accumulated dataset becomes an important switching barrier.

Competition concern

A dominant platform may obtain competitive insulation because rivals cannot reproduce the user's accumulated information at reasonable cost.

B. Technical lock-in

Technical dependency can arise through:

• proprietary APIs;
• non-standard data formats;
• closed protocols;
• proprietary development environments;
• platform-specific software;
• specialised hardware;
• undocumented interfaces.

Technical incompatibility can make nominally "open" switching economically unrealistic.

4. Cloud Computing Lock-In

Cloud markets provide a particularly important example.

A firm may initially choose a cloud provider because of price or performance. Over time, however, it may become dependent upon:

• proprietary databases;
• serverless functions;
• machine-learning infrastructure;
• storage architectures;
• identity systems;
• monitoring tools;
• proprietary APIs.

The cost of migration therefore increases with the depth of technological integration.

A useful model is:

Cloud Lock ⁣− ⁣in=Migration Cost+Downtime+Retraining+Reengineering+Data TransferCloud\ Lock\!-\!in = Migration\ Cost + Downtime + Retraining + Reengineering + Data\ Transfer

The relevant competition question is whether the provider's ecosystem architecture creates artificially elevated migration costs, rather than merely reflecting legitimate technical investment.

5. Network-Effect Lock-In

Network effects can create a different form of switching cost.

A user does not merely ask:

"How much does it cost me to change platforms?"

The user may instead ask:

"Will the people I need to interact with also move?"

This creates a collective switching problem.

For example:

Individual Switching Value=Alternative Platform Value−Migration Cost−Network LossIndividual\ Switching\ Value = Alternative\ Platform\ Value - Migration\ Cost - Network\ Loss

Even a superior rival may struggle to attract users if those users would lose access to their existing network.

This is particularly significant for:

• social networks;
• messaging;
• professional platforms;
• marketplaces;
• payment networks;
• multiplayer gaming;
• creator platforms.

6. Quantifying Migration Costs

A competition authority can construct a Migration Cost Index (MCI).

For example:

MCI=Total Migration CostAnnual Customer ExpenditureMCI = \frac{Total\ Migration\ Cost}{Annual\ Customer\ Expenditure}

Illustratively:

If annual expenditure on the incumbent's service is £500,000:

MCI=725,000500,000=1.45MCI = \frac{725,000}{500,000}=1.45

Thus, migration costs equal 145% of one year's expenditure.

That does not automatically establish market power, but it is strong evidence that the incumbent's installed customer base may be substantially insulated from competitive switching.

7. Migration-Cost Ratio

Another useful metric is:

MCR=PV(Migration Costs)PV(Expected Benefits From Switching)MCR = \frac{PV(Migration\ Costs)}{PV(Expected\ Benefits\ From\ Switching)}

Where:

• PV = present value;
• migration costs include one-off and recurring costs;
• expected benefits include lower prices, improved quality, better functionality or innovation.

If:

MCR>1MCR > 1

switching may be economically unattractive even where the rival is objectively superior.

The analysis becomes particularly important where the incumbent can impose a small but persistent degradation in quality without losing customers.

8. The "Effective Switching Rate"

Authorities can also examine actual switching behaviour.

ESR=Number of Customers SwitchingEligible Customer BaseESR = \frac{Number\ of\ Customers\ Switching}{Eligible\ Customer\ Base}

A very low switching rate may suggest lock-in, but it must be interpreted carefully.

Low switching can also result from:

• customer satisfaction;
• superior product quality;
• temporary contracts;
• lack of attractive alternatives.

Therefore, switching data should be combined with counterfactual evidence.

9. The SSNIP/Quality-Degradation Connection

Traditional market definition asks whether customers would switch following a small but significant non-transitory increase in price.

Digital ecosystems require a broader approach.

Instead of:

"Would users switch following a 5–10% price increase?"

the authority may ask:

"Would users migrate following a measurable degradation in privacy, interoperability, functionality, service quality or innovation?"

This is particularly relevant for zero-price platforms.

For example:

Effective Switching Pressure=Price Elasticity+Quality Elasticity+Privacy Elasticity+Interoperability ElasticityEffective\ Switching\ Pressure = Price\ Elasticity + Quality\ Elasticity + Privacy\ Elasticity + Interoperability\ Elasticity

A platform may possess significant market power even where monetary prices remain zero.

10. Data Portability as a Lock-In Variable

Data portability can substantially reduce migration costs.

Suppose:

MCbefore=£1,000MC_{before}=£1,000

and effective portability reduces it to:

MCafter=£300MC_{after}=£300

The portability intervention has reduced migration costs by:

70%70\%

This creates a direct competition-law connection between data governance and market contestability.

Portability is therefore not merely a consumer-rights mechanism. It can operate as a competition-enhancing remedy.

11. Interoperability and Migration Costs

Interoperability can similarly reduce lock-in.

For example:

Migration Cost=Technical Integration+Data Conversion+Network LossMigration\ Cost = Technical\ Integration + Data\ Conversion + Network\ Loss

Interoperability can reduce the first two components and, in some circumstances, network-related losses.

This explains why competition authorities may consider:

• API access;
• interoperability obligations;
• common technical standards;
• switching interfaces;
• data export;
• multi-homing;
• identity portability.

12. Six Important Case Laws

1. United States v. Microsoft Corp. (2001)

The Microsoft litigation is foundational for understanding technological lock-in.

Microsoft's control over the Windows operating-system environment and its relationship with complementary software demonstrated how a dominant technology platform can use ecosystem advantages to reinforce its position.

The case is particularly relevant to migration-cost analysis because application compatibility and the installed base of Windows created substantial barriers to moving to alternative operating systems.

Principle

Compatibility and ecosystem dependence can constitute important barriers to entry and competitive displacement.

13. European Commission – Google Android (2018)

The Google Android decision is highly relevant to ecosystem lock-in.

The Commission examined Google's contractual arrangements involving Android devices, Google Search, the Play Store and browser applications.

The underlying competition concern was not merely the price of an individual product. It concerned the way interconnected contractual and technical arrangements could reinforce Google's position across the mobile ecosystem.

Relevance

The case illustrates:

• ecosystem leverage;
• defaults;
• distribution restrictions;
• application ecosystems;
• network effects;
• cross-market reinforcement.

Principle

Competition analysis may need to examine interdependencies between complementary digital products, rather than considering each product separately.

14. European Commission – Google Shopping (2017)

The Google Shopping case demonstrates how a dominant digital platform can advantage its own complementary service through control over a critical digital access point.

The importance for migration-cost analysis lies in the broader concept of platform dependency.

A rival may technically exist but remain unable to obtain sufficient traffic to compete effectively.

Thus:

Rival Entry≠Effective ContestabilityRival\ Entry \neq Effective\ Contestability

where the rival lacks equivalent access to the platform's ecosystem.

Principle

Digital competition can be distorted even where competitors formally remain present in the market.

15. European Commission – Google Search (AdSense) (2019)

The AdSense decision provides another ecosystem-level example.

The Commission examined contractual restrictions affecting third-party websites' ability to source search advertising from competing providers.

The case demonstrates the importance of analysing commercial dependency.

A business dependent upon one platform for advertising revenue may face significant costs if it attempts to migrate to competing infrastructure.

Principle

Contractual arrangements can increase ecosystem dependence and reduce effective multi-homing.

16. Bundeskartellamt – Facebook/Meta Data Case (2019)

The German Facebook decision is particularly important because it connected data collection, platform power and competition law.

The Bundeskartellamt examined Facebook's ability to combine user data from Facebook with data obtained from other services.

The case demonstrates that lock-in can arise not merely from technical barriers but from accumulated data advantages.

Relevance to migration-cost quantification

A user's accumulated data can have:

• economic value;
• personalisation value;
• advertising value;
• social value;
• network value.

Therefore:

Data Lock ⁣− ⁣in=Cost of Losing Historical Data+Cost of Reconstructing DataData\ Lock\!-\!in = Cost\ of\ Losing\ Historical\ Data + Cost\ of\ Reconstructing\ Data

The greater these values, the greater the effective switching barrier.

17. FTC v. Facebook / Meta Litigation

The US litigation concerning Facebook's acquisitions of Instagram and WhatsApp provides another important perspective.

The competition analysis considered network effects, barriers to entry and the difficulty of displacing an established social-network platform.

The relevance to migration costs is particularly strong for social ecosystems.

Users may not switch because the value of a social network depends upon the presence of:

• friends;
• family;
• businesses;
• creators;
• advertisers;
• communities.

Therefore, the switching cost is partly collective rather than individual.

Principle

Network effects can make established digital ecosystems difficult to displace even where users can technically create accounts with rival services.

18. Epic Games v. Apple

The Epic Games litigation concerning Apple's App Store provides a major example of ecosystem control involving:

• app distribution;
• payment systems;
• developer access;
• commissions;
• platform rules;
• alternative distribution channels.

The case is highly relevant to migration-cost analysis because developers may have invested heavily in an ecosystem-specific distribution and monetisation infrastructure.

The relevant cost is therefore not simply:

"Can the developer publish somewhere else?"

Instead:

"What proportion of its users, revenue, technical integration and accumulated investment would the developer lose by moving?"

Principle

Platform governance can affect the practical contestability of digital ecosystems even where alternative technologies technically exist.

19. A Comparative Case-Law Matrix

20. Quantification Through Customer-Level Analysis

Competition authorities should ideally calculate migration costs at the customer level.

For customer ii:

MCi=Ti+Di+Ri+Ki+Ni+LiMC_i = T_i+D_i+R_i+K_i+N_i+L_i

Where:

• TiT_i = technical migration;
• DiD_i = data migration;
• RiR_i = retraining;
• KiK_i = contractual costs;
• NiN_i = network loss;
• LiL_i = loss of accumulated ecosystem benefits.

The authority can then calculate:

Average MC=∑MCiNAverage\ MC = \frac{\sum MC_i}{N}

However, the distribution is often more important than the average.

A platform may have:

• low switching costs for casual users;
• moderate switching costs for ordinary users;
• extremely high switching costs for professional users.

This segmentation can materially change the competition analysis.

21. Time as a Component of Migration Cost

Migration should not be measured solely in monetary terms.

If migration takes six months, the opportunity cost may be substantial.

Total Migration Cost=Direct Cost+Opportunity Cost+Downtime CostTotal\ Migration\ Cost = Direct\ Cost + Opportunity\ Cost + Downtime\ Cost

For a business:

Opportunity Cost=Lost Output×Duration of MigrationOpportunity\ Cost = Lost\ Output \times Duration\ of\ Migration

Consequently, a technically inexpensive migration can nevertheless be economically prohibitive.

22. Ecosystem Depth Index

A useful conceptual measure is an Ecosystem Depth Index (EDI):

EDI=w1I+w2D+w3N+w4C+w5S+w6HEDI = w_1I+w_2D+w_3N+w_4C+w_5S+w_6H

Where:

• II = integration intensity;
• DD = data dependence;
• NN = network dependence;
• CC = contractual dependence;
• SS = service complementarity;
• HH = historical investment.

The weights w1…w6w_1 \ldots w_6 can be determined according to the relevant market.

The higher the EDI, the greater the possibility that the ecosystem operates as a retention structure rather than merely as a collection of independent products.

23. Lock-In Can Be Deliberate or Structural

Competition law should distinguish between two situations.

Structural lock-in

Lock-in results naturally from:

• economies of scale;
• legitimate innovation;
• network effects;
• technical complexity.

This does not automatically constitute unlawful conduct.

Strategic lock-in

The incumbent deliberately increases switching costs through:

• restrictive contracts;
• technical incompatibility;
• discriminatory interoperability;
• excessive data restrictions;
• tying;
• self-preferencing;
• exclusionary defaults;
• artificial degradation of portability.

The second category is considerably more problematic.

24. Migration Costs and Article 102 TFEU

Under Article 102 TFEU, high migration costs can strengthen an assessment of dominance because they reduce the competitive constraint imposed by existing and potential rivals.

Potential theories include:

A. Refusal to interoperate

A dominant platform may make migration technically difficult by denying necessary interoperability.

B. Tying

A dominant service may require customers to adopt complementary products, increasing ecosystem dependency.

C. Exclusive arrangements

Contracts may discourage customers from multi-homing.

D. Self-preferencing

The platform may favour its own complementary products, making rival ecosystems less attractive.

E. Exploitative lock-in

A dominant platform may worsen terms after users have become sufficiently dependent upon the ecosystem.

25. UK Competition-Law Relevance

In the UK, migration-cost analysis is particularly relevant to the assessment of:

• substantial market power;
• barriers to entry;
• barriers to expansion;
• consumer switching;
• multi-homing;
• network effects;
• ecosystem leverage;
• exclusionary abuse.

The Digital Markets, Competition and Consumers Act 2024 also makes ecosystem-level analysis increasingly important for firms designated with strategic market status.

The relevant conceptual question is:

Could a sufficiently large number of users or businesses realistically move away from the ecosystem in response to worsened competitive conditions?

If the answer is no because migration costs are extremely high, the incumbent may face significantly weaker competitive discipline.

26. Quantifying Lock-In Through Counterfactual Analysis

A sophisticated investigation should compare:

Actual migration cost

MCAMC_A

with:

Counterfactual migration cost

MCCMC_C

where the latter represents migration under technically neutral and interoperable conditions.

Then:

Artificial Lock ⁣− ⁣in=MCA−MCCArtificial\ Lock\!-\!in = MC_A-MC_C

For example:

MCA=£900,000MC_A=£900,000 MCC=£400,000MC_C=£400,000

Therefore:

Artificial Lock ⁣− ⁣in=£500,000Artificial\ Lock\!-\!in=£500,000

This approach is particularly useful because it distinguishes legitimate technical complexity from potentially exclusionary design.

27. Migration Cost and Consumer Welfare

High migration costs can harm consumers by reducing their ability to discipline the incumbent.

The effects can include:

• higher prices;
• reduced quality;
• reduced privacy;
• weaker innovation;
• reduced choice;
• lower interoperability;
• slower technological development.

The harm may therefore occur even when prices do not increase.

A digital platform might instead:

increase advertising intensity → reduce privacy → worsen interoperability → impose less favourable contractual terms.

Users remain because migration is too costly.

28. Migration Costs and Multi-Homing

Multi-homing can substantially weaken lock-in.

If users simultaneously maintain accounts with several platforms:

Effective Switching Cost↓Effective\ Switching\ Cost \downarrow

because abandoning one platform does not require abandoning the entire network.

Consequently, authorities should measure:

• percentage of multi-homing users;
• cost of maintaining multiple accounts;
• functionality differences;
• data synchronisation;
• cross-platform interoperability;
• contractual restrictions on multi-homing.

A platform with extensive multi-homing may have lower effective lock-in than headline market share suggests.

29. Important Evidentiary Indicators

Authorities investigating digital ecosystem lock-in should seek:

1. internal switching-cost studies;
2. customer churn data;
3. customer complaints;
4. exit interviews;
5. API documentation;
6. data-export functionality;
7. termination clauses;
8. migration contracts;
9. engineering estimates;
10. customer acquisition costs;
11. multi-homing rates;
12. developer migration statistics;
13. downtime estimates;
14. interoperability records;
15. internal strategy documents.

Particularly revealing evidence may be an internal statement that customers are unlikely to leave because "migration is too difficult."

30. A Practical Lock-In Quantification Framework

A competition authority can use the following sequence:

Step 1 — Identify the ecosystem

Map:

Core Service→Complementary Services→Data→Network→InfrastructureCore\ Service \rightarrow Complementary\ Services \rightarrow Data \rightarrow Network \rightarrow Infrastructure

Step 2 — Identify switching events

Determine what must actually change.

Step 3 — Measure direct costs

Calculate:

• termination fees;
• migration fees;
• implementation costs.

Step 4 — Measure indirect costs

Calculate:

• retraining;
• lost productivity;
• lost data;
• loss of network connections.

Step 5 — Measure time

Calculate migration duration and opportunity costs.

Step 6 — Measure counterfactual migration

Ask how expensive switching would be under effective interoperability.

Step 7 — Test behavioural response

Analyse actual customer switching.

Step 8 — Assess competitive significance

Determine whether migration costs materially weaken rivals' ability to constrain the incumbent.

31. Key Legal Test

The central competition-law test can be expressed as:

Lock ⁣− ⁣in becomes competitively significantwhen migration costs are sufficiently high to prevent effective switching\boxed{ Lock\!-\!in\ becomes\ competitively\ significant when\ migration\ costs\ are\ sufficiently\ high\ to\ prevent\ effective\ switching }

But:

High Migration Costs≠Automatic AbuseHigh\ Migration\ Costs \neq Automatic\ Abuse

The authority must establish additional elements such as:

• dominance or substantial market power;
• exclusionary conduct;
• causation;
• foreclosure;
• competitive harm;
• lack of sufficient countervailing competition.

Conclusion

Digital ecosystem migration costs are increasingly important indicators of market power because digital competition depends not merely upon whether an alternative exists, but whether users can realistically move to it.

The most important insight is that lock-in should be measured across the whole ecosystem:

Lock ⁣− ⁣in=Data+Technology+Network+Contracts+Learning+Accumulated Investment+Complementary Services\boxed{ Lock\!-\!in = Data + Technology + Network + Contracts + Learning + Accumulated\ Investment + Complementary\ Services }

The strongest methodology therefore compares actual migration costs with counterfactual interoperable migration costs. This allows competition authorities to distinguish legitimate ecosystem efficiencies from artificially created switching barriers.

The Microsoft, Google Android, Google Shopping, Google AdSense, Facebook/Meta, FTC–Facebook and Epic Games litigation collectively demonstrate that compatibility, defaults, data accumulation, network effects, contractual dependency and platform governance can all contribute to digital ecosystem entrenchment.