JULY 2026

Contestability Index: Coding Handbook

Tracy Xu, Amr Yakout, Zhiyang Cheng
Knight-Georgetown Institute

Download PDFVersion 4.1 · 60 pages · 1.4 MB

About the Knight-Georgetown Institute

The Knight-Georgetown Institute (KGI) is dedicated to connecting independent research with technology policy and design. KGI serves as a central hub for the growing network of scholarship that seeks to shape how technology is used to produce, disseminate, and access information. KGI is designed to provide practical resources that policymakers, journalists, and private and public sector leaders can use to tackle information and technology issues in real time. Georgetown University and the Knight Foundation came together to launch the institute in 2024. Learn more about KGI at https://kgi.georgetown.edu.

I. Introduction

This Coding Handbook documents the pilot implementation of the Contestability Measurement Framework and the coding records underlying the measures. Section II sets out the De Jure scoring criteria, data methodology, and legal-form classification. Section III specifies the De Facto scoring rules, indicator structure, market-specific adaptation, gatekeeper selection, and evidence standards. Section IV reports the jurisdiction-level De Jure comparison and presents the factor-level scores, rationales, market coverage, and legal bases for each jurisdiction. Section V applies the framework to the pilot gatekeeper–market observations. For each application, it defines the market scope and applicable mechanisms, reports the aligned market-level De Jure and De Facto assessments, and presents Effective Contestability as a derived summary of the two layers.

A dedicated project database tracks the full lifecycle of the underlying data. For each item, it records the source (the issuing authority and official publication), the raw collected document, the processed data used for analysis, and a complete version history recording when entries were collected, reviewed, and revised across snapshots. As a result, every entry reported in this Coding Handbook is fully traceable, from the final score back through the processed text to the original official document. The database contents specific to each measure are described in the corresponding section.

II. De Jure Contestability

This section sets out how De Jure Contestability is coded. It defines the scoring criteria applied to every mechanism–jurisdiction cell, describes the data methodology that produces the underlying record, and sets out the legal-form classification used to keep instruments comparable across very different legal systems. The coded results are reported in Section IV.

A. Scoring criteria

Each mechanism–jurisdiction cell rests on three factors, each scored on the same 0 / 0.5 / 1 grid: Existence (whether the relevant obligation is present in the enacted framework), Scope (the range of firms, services, and activities covered), and Enforceability (whether the provision is backed by credible oversight and sanctions).

Final provision score = MIN(Existence, Scope, Enforceability). The minimum is used because the factors do not compensate for one another: an obligation that does not exist is not helped by strong enforcement machinery; a well-drafted obligation binding a single firm is still not a market-wide rule; a broad rule without credible enforcement offers little protection in practice. Scope is judged against the mechanism’s own economic domain rather than the whole economy — a marketplace-only statute covers the full domain of a marketplace-native mechanism (Scope = 1 there) but only part of the domain of a mechanism that operates economy-wide (Scope = 0.5).

B. Data methodology

The sixteen mechanisms are grouped under the four dimensions shown in Section IV.A. A mechanism is a stable contestability channel rather than a provision of any single statute; where a channel can be secured through more than one route (for example, alternative distribution through either sideloading or third-party app stores, or technical interoperability through either API access or hardware access), the routes are treated as one mechanism and scored on the strongest available channel.

The data were collected in four steps, so that every score can be traced to a named legal instrument and every instrument to its official text. First, for each jurisdiction the digital-competition instruments in force or pending as of June 2026 (statutes, regulations, court remedies, conduct requirements, administrative rules) were registered with citations and links to official sources, together with a mapping from each mechanism to the articles that govern it. Second, for each mechanism the operative text of every instrument bearing on it was copied verbatim from the primary source, with English translations where needed; scoring works from this primary text, not from secondary summaries. Third, each mechanism–jurisdiction pair was scored on the three factors set out in Section II.A. Fourth, the mechanism scores were averaged into dimension scores and then into the De Jure Contestability Index reported in Section IV.A. The records produced at each step are kept in the project database.

For the De Jure layer, the project database described in the introduction will include: (i) the source register — every statute, regulation, court remedy, conduct requirement, and administrative rule bearing on the sixteen mechanisms, with citation, status, official-source link, and the pending/pipeline instruments tracked alongside those in force; (ii) the original documents — the official texts as published (official gazettes, EUR-Lex, e-Gov, court dockets), stored in the original language; (iii) the processed versions — the cleaned machine-readable texts and the verbatim provision extracts, with English translations, that scoring works from; (iv) the assessments — all 144 mechanism × jurisdiction records with the three factor scores, the final score, the rationale, the targeted market, and the legal basis, plus the market-coverage matrix; and (v) the history — the review status of each record and its changes across snapshots as instruments are enacted, amended, activated by orders, or repealed (for example, a pipeline bill moving to in force).

C. Legal-form classification

To keep the factor signatures comparable across very different legal systems, instruments are classified into one of eight mutually exclusive legal forms; each form has a typical factor signature from which its score ceiling follows under the MIN rule. This is why, outside the EU and Germany, most non-zero cells sit at 0.5: they rest on firm-specific orders, sectoral statutes, guidance, or case practice rather than on generally applicable ex ante statutes.

Finally, when a market-level index is computed, a sectoral instrument counts at full strength within its own market (for example, Turkey’s marketplace rules in the marketplace market), even where the jurisdiction-level table shows a Scope-capped 0.5.

III. De Facto Contestability

This section sets out how De Facto Contestability is coded. It states the scoring criteria, specifies the indicator set, sets out how the set is adapted from one market to the next, states how the gatekeeper and market are selected, and describes the evidence standards. These rules are shared across markets. The market applications are reported in Section V.

A. Scoring criteria

Each dimension is decomposed into De Facto indicators drawn from the Methodology Manual and scored on the 0 / 0.5 / 1 grid used throughout the Index. Each indicator is scored twice, once for the pre-DMA baseline and once for the post-DMA state, so that the change isolates the effect of the regulation from the pre-existing level.

Scores are assigned by the following rule.

Before scoring, each indicator is assigned an applicability status. Applicable means the mechanism operates in the market. Not Applicable means it does not, and the indicator is excluded from aggregation rather than scored zero, since non-applicability carries no information about contestability. Applicable but Missing means the mechanism operates but the evidence required to score it has not been obtained; the indicator is excluded from the denominator and the reason is recorded. Missing status is used where a score would otherwise depend on inference, on general knowledge of a platform, or on the absence of an adverse regulatory finding.

Dimension scores are the unweighted arithmetic mean of the applicable indicators scored. The De Facto Contestability Index is the equal-weighted average of the four dimension scores. The De Jure layer takes the minimum of Existence, Scope and Enforceability because a missing obligation is not cured by enforcement machinery. The De Facto layer averages, because partial progress on several channels adds to observed contestability.

Formal availability does not establish adoption. Where an indicator concerns use or market behaviour, platform documentation alone supports no score above 0.5.

Market-share persistence is excluded as a scored dimension. Following Furman et al. (2019), market share is an outcome of contestability rather than a dimension of it, and including it would introduce circularity into the Index.

B. The De Facto indicator set

The indicator set follows the sixteen observable market conditions in the Methodology Manual. Each is the observable counterpart of one contestability mechanism and is assigned to one dimension. Codes are stable across markets.

Each indicator is scored once. Where an observed condition bears on more than one mechanism, it is assigned to the mechanism whose economic function it most directly reflects, and the assignment is documented. The set is not extended with market-specific composites: aggregate measures such as cost of entry, or structural infrastructure costs not imposed by the gatekeeper, are recorded in the evidence column rather than scored.

C. Market-specific variation in the indicator set

The four dimensions and sixteen mechanisms are stable and do not vary. Three things vary across markets.

Applicability. Some mechanisms do not operate in a given market. Cross-platform messaging connectivity applies to number-independent interpersonal communications and to no other market coded here. Parity applies to app stores and intermediation, and reaches search through Shopping, which sits inside the market boundary. Anti-steering reaches any market whose business users are acquired through a Core Platform Service, and so applies in search, though the conduct it observes there differs. Applicability is fixed before scoring and held constant across jurisdictions and periods within a market.

Observation. The same mechanism is observed through different evidence. Default switching is observed in search through choice-screen deployment and the number of steps to change an engine, and in app distribution through the range of reassignable default categories and whether the store is among them.

Competitive weight. The same indicator carries different economic weight across markets. Multi-homing signals little in search, where a user running two engines still sends most queries to one, and signals more in social networking, where accounts on several services coexist and attention divides. Data portability matters more in social networking, where the social graph is the asset that holds users, than in search. Interoperability binds in messaging and operating systems and matters less in marketplaces. Under equal weighting these differences appear in the evidence and in score dispersion, not in the weights. A market-specific weighting scheme requires justification and documentation.

De Facto scores are comparable across jurisdictions and periods within a market, and across gatekeepers within a market where the same indicator set has been applied. They are not comparable across markets. An index of 0.50 in app distribution and 0.50 in search describe different conditions reached through different mechanisms. Cross-market statements should be made at the level of the mechanism.

D. Gatekeeper and market selection

De Facto Contestability is measured for a gatekeeper–market pair. The Methodology Manual requires the gatekeeper to be identified as part of the observation where competitive conditions are platform-specific. An indicator set applies only where the named firm operates the bottleneck the mechanism describes.

A gatekeeper enters a market observation only where it controls a Core Platform Service in that market. Scoring Amazon in online search would be an error of classification: Amazon operates no general search engine, the search indicators would have no object, and the resulting score would measure non-participation rather than low contestability. Alphabet holds the search bottleneck in the European Union. Amazon is the subject in online marketplaces, Meta in social networking and number-independent messaging, and Apple and Alphabet in app distribution and mobile operating systems.

Where a market has more than one gatekeeper, each is coded separately on the same indicator set and reported alongside the other. Averaging them would remove the variation the measure is intended to record. A market-level summary across gatekeepers requires an explicit weighting rule for the relative economic importance of each ecosystem.

Three consequences follow. Applicability is determined against the named gatekeeper’s service. Evidence must concern that gatekeeper; a decision addressed to Apple establishes nothing about Alphabet. And a pre-period score of 0 is distinct from non-participation: Apple scores 0 on alternative distribution in December 2023 because it prohibited the channel, which is a finding about contestability, while a firm operating no service in the market is not scored.

E. Evidence standards

Each pair is scored from Commission non-compliance decisions, preliminary findings and specification decisions; gatekeeper compliance reports, developer documentation and published fee schedules; national competition and data-authority actions; judgments of the Court of Justice; and market data from independent providers. Every cell is traceable from its score to the originating document.

A regulatory penalty is not by itself evidence of a market condition. A gatekeeper may be fined because its conduct did not change, in which case the coding is unchanged. A score moves only on documented change in the observable condition. Where a penalty decision also records implemented changes, those records are the evidence, and the decision is cited for the finding that friction persists.

Announced measures that were not operative during the observation window are recorded in the evidence column and do not affect the score.

Three pairs are coded in this version: Alphabet in online search, Apple in iOS app distribution, and Alphabet in Android app distribution. The two app-distribution observations use one indicator set and are reported together so the ecosystems can be compared.

IV. Jurisdiction-Level De Jure Contestability

This section documents the Jurisdiction-Level De Jure Contestability layer of the Index: the existence, scope, and enforceability of the sixteen regulatory mechanisms defined in the Methodology Manual, assessed across the enacted legal frameworks of nine jurisdictions (the European Union, Germany, the United States, Turkey, Japan, Korea, China, the United Kingdom, and Australia) as a post-period snapshot dated June 2026.

This layer applies the scoring criteria, data methodology, and legal-form classification set out in Section II. It is organized as follows. Section IV.A reports the cross-jurisdiction scores by mechanism, together with the dimension averages and the aggregate De Jure Contestability Index; and Section IV.B presents the full scoring table for each jurisdiction, in sequence by country, with the factor scores, rationales, target markets, and legal bases that constitute the evidence for every cell.

A. Cross-jurisdiction comparison

The table reports the final score of every mechanism–jurisdiction cell, organized with the jurisdictions on the horizontal axis and the mechanisms, grouped by dimension, on the vertical axis. The dimension averages and the De Jure Contestability Index — the equal-weighted average of D1–D4 — close the table as its aggregate rows. Jurisdiction codes: EU = European Union, DE = Germany, US = United States, TR = Turkey, JP = Japan, KR = Korea, CN = China, UK = United Kingdom, AU = Australia.

The same 0.5 can mean different things — a strong law with narrow reach (Turkey’s anti-steering rule, marketplaces only) or broad reach in weak legal form (China’s ranking-neutrality rules) — so the factor scores matter, not just the final score. The per-jurisdiction tables in Section IV.B record the three factors separately, with the rationale, target market, and legal basis for every cell.

B. Jurisdiction scoring tables (scores and evidence)

The tables below present the full scoring record for each jurisdiction, in sequence by country, in the order EU, Germany, United States, Turkey, Japan, Korea, China, United Kingdom, Australia. A note under each country heading states the jurisdiction-wide enforcement machinery. The columns give the mechanism; the final score; the factor scores E = Existence, S = Scope, Enf = Enforceability; the rationale, which records only what deviates from the standard reading of the scores set out in Section II.A (‘—’ indicates that there is no deviation to explain, and for rows scored 0 the absent or pipeline instrument is named in the Legal basis column); the target market or field; and the full name of the legal basis. Dimension averages and the jurisdiction’s De Jure Contestability Index close each table.

1. EU

Enforcement: European Commission (exclusive competence); fines up to 10% of worldwide turnover, 20% for repeat infringements; DMA obligations self-executing on designation.

2. Germany

Carries the EU scores: DMA fully applicable; enforcement by the Commission plus the Federal Cartel Office’s decentralised powers (11th GWB amendment) and active private enforcement (LG Mainz 2025).

3. United States

No ex ante statute: obligations are firm-specific court remedies enforced by the supervising courts (contempt powers; the US v. Google final judgment adds a Technical Committee with source-code access). Rows scored 0 have no instrument on point beyond general antitrust law.

4. Turkey

Enforcement: Ministry of Trade administrative fines plus the ETHS licensing regime (E-Commerce Law 6563 as amended by Law 7416). Rows scored 0 await the draft amendment to Competition Law 4054 (pipeline).

5. Japan

Enforcement: JFTC; surcharges up to 20% of relevant Japanese turnover (30% for repeat violations); MSCA in full force since 18 December 2025.

6. Korea

Enforcement: KCC for the TBA app-market rules; KFTC case practice under the MRFTA; PIPA portability in force since March 2025 (sectoral). Rows scored 0 have no instrument on point.

7. China

Hybrid regime: statutes (AML, AUCL, E-Commerce Law) state high-level duties; ministerial rules (SAMR, CAC, MIIT) supply the operative detail; enforcement is discretionary and campaign-cyclical. Rows scored 0 have no instrument on point.

8. United Kingdom

DMCCA: obligations arise only as conduct requirements (CRs) imposed after SMS designation — designated-but-no-CR cells score 0 (machinery ready, nothing to enforce); CMA penalties up to 10% of global turnover. Three search CRs are in force at the observation date: the publisher conduct requirement, imposed 3 June 2026, and the fair-ranking and data-portability conduct requirements, both imposed 17 June 2026. The User Choice conduct requirement was consulted on 28 January 2026 and has not been imposed. Apple/Google mobile commitments took effect 1 April 2026. The assimilated P2B Regulation (2019/1150) provides an economy-wide business-user floor. DMCCA commitments are binding and CMA-enforceable, classified F2.

9. Australia

Two instruments on point are in force: the Consumer Data Right (CCA Part IVD), enforced by the ACCC/OAIC, and Australian Privacy Principle 6 of the Privacy Act 1988, enforced by the OAIC. All other rows score 0: they are pipeline cells for the proposed Digital Competition Regime, except choice integrity (Unfair Trading Practices Bill 2026, pending) and commercial independence and network interoperability (no instrument on point).

V. Market–Jurisdiction Applications

This section applies the framework market by market. It reports the De Facto Contestability layer of the Index: the observed contestability of specific gatekeeper–market pairs, measured by the extent to which enacted obligations have translated into changes in market conditions in practice. Where the De Jure layer scores the law as written across nine jurisdictions, the De Facto layer scores conduct and outcomes for individual Core Platform Services (CPS), comparing a pre-DMA baseline (December 2023) with the latest observed state (July 2026); the scoring criteria and evidence base it applies are set out in Section III. For each market, the section first characterizes the market itself, then presents the full De Facto scoring table for the gatekeeper–market pair — the sub-indicator scores, the pre/post comparison, and the supporting evidence for every cell — restates the market-level De Jure Contestability score — both as it stands in June 2026 and against a December 2023 pre-DMA baseline, so that the legal layer can be read as a movement rather than a level — and finally combines the layers into the market’s Effective Contestability. Two EU markets are coded in full, across three gatekeeper–market pairs: online search (Google Search), and app distribution (Apple App Store and Google Play). The De Jure baseline is coded for Google Search and the Apple App Store, the two pairs for which a market-level De Jure catalogue is reported.

A. EU — Online Search

Market: online search. Assessed service: Google Search (Alphabet Inc.), designated as a Core Platform Service under the EU Digital Markets Act (DMA) on 6 September 2023.

1. Market scope and applicable mechanisms

The standardized market is online search, assessed through Google Search, a designated Core Platform Service under the DMA. Applicable mechanisms are identified separately for the two layers. At the De Jure layer, all sixteen mechanisms are scored: in the European Union the DMA and the Platform-to-Business Regulation impose a complete set of obligations on the designated service. At the De Facto layer, the set narrows. De Jure scoring credits an obligation wherever it binds Alphabet as a designated gatekeeper, including where its operative locus is app distribution, the operating system or messaging; De Facto scoring observes only conduct in the search service itself, so a mechanism operating in an adjacent market has no object here. Technical interoperability (D4.1) and cross-platform messaging connectivity (D4.2) are therefore Not Applicable, and the remaining fourteen indicators apply. Applicability is fixed before scoring and held constant across both periods (Section III.C). Which of the fourteen could then be scored on the available evidence is a separate question, reported with the scoring table in Section V.A.3.

2. Market-level De Jure Contestability

For the search market, the Existence, Scope, and Enforceability factors are carried over from the jurisdiction-level coding of Section IV.B unchanged; for each mechanism, the governing Digital Markets Act or supporting provision and the rationale as applied to Google Search. Where an obligation directly governs the search service it is justified on that provision; where it binds Alphabet as a designated gatekeeper but operates in an adjacent market — app distribution, the operating system, or messaging — it is marked as such and scored 1 as part of the gatekeeper's full obligation set. The table reports two periods. The June 2026 column states the law as it stands; the December 2023 column scores the same sixteen mechanisms against the pre-existing acquis with the Digital Markets Act excluded, so that what the Act contributed can be separated from what the Union already required. That baseline is a counterfactual rather than a literal legal state: at December 2023 the Act was in force, applicable, and the service had been designated, but the Article 5, 6 and 7 obligations did not bind until 7 March 2024, so scoring the literal state would set Enforceability to zero on every Act-based cell and reach the same result by a longer route. Applicability is held constant across the two periods, as Section III.C requires.

The sixteen mechanism scores aggregate as elsewhere in the Index: within each dimension the mechanism scores are averaged, and the four dimension scores are then averaged with equal weight into the market-level index (Section IV.A). All four dimensions score 1.00 — User Mobility & Switching (D1), from #1–#5; Rival Access & Distribution (D2), from #6–#9; Platform Neutrality & Non-Discrimination (D3), from #10–#12; and Data, Interoperability & Infrastructure Access (D4), from #13–#16 — giving an equal-weighted market-level De Jure Contestability of 1.00.

At the December 2023 baseline the same four dimensions stood at 0.50, 0.38, 0.50 and 0.13, giving an equal-weighted market-level De Jure Contestability of 0.38 — a movement of +0.62. One mechanism scored 1 before the Act: data portability, on Article 20 of the General Data Protection Regulation, which the Act raised to a continuous and real-time standard rather than created. Ten scored 0.5, held there by the Platform-to-Business Regulation's transparency-only character, by the firm-specific reach of an Article 102 remedy, or by the general terms of the data-protection acquis. Five scored 0, and D4 stood at 0.13: nothing in force required interoperability, messaging connectivity, or access to the data generated on the search engine, and cross-service data combination was constrained only by the lawful-basis requirement of Article 6 of the General Data Protection Regulation, as read in Case C-252/21 Meta Platforms v Bundeskartellamt.

3. Market-level De Facto Contestability

Gatekeeper: Alphabet Inc. Core Platform Service: Google Search. Designated 6 September 2023.

Two indicators are Not Applicable: technical and messaging interoperability operate in the device and communications layers. Two are Applicable but Missing. Twelve are scored.

Google's De Facto Contestability Index rises from 0.28 to 0.52 (Δ +0.24). D2 is the highest-scoring dimension in both periods at 0.67. Article 5(4) applies to business users acquired through Search, and the Shopping ads policies impose no restriction on communicating or promoting offers or on concluding contracts off-platform, so merchants could route customers to their own checkout before the DMA and can do so now, though further analysis might be warranted by pulling the full merchant center terms and services. The largest gain is in D4, which moves from 0 to 0.50 on the Article 5(2) consent flows deployed in March 2024 and on the European Search Dataset Licensing Program, which Google has operated since 6 March 2024. Under Article 6(11) the Commission opened specification proceedings in January 2026 and adopted binding measures on 16 July 2026; these take effect from January 2027, after the observation window. D3.1, self-preferencing in results, moves from 0 to 0.5: Google removed the Flights unit and introduced dedicated comparison-site units ahead of the March 2024 deadline, and Mirai measured Google Hotels' share as a source of reservations falling from 13.4% to 8.9%, a 33% relative decline, with online travel agencies gaining visibility within Google against a flat non-EU control. The indicator is held below 1 by the Commission's decision of 23 July 2026 in Case DMA.100193, which found breach of Article 6(5) across nine Alphabet services, dated from March 2024 and ongoing at adoption, and imposed a fine of EUR 460 million with a 60-day remedy order. D1 moves from 0.20 to 0.40 on the EEA choice screens and their randomised, non-pre-selected design, held down by application replacement, which scores 0 in both periods because the Google Search app remains non-removable. Two indicators are excluded as Applicable but Missing: commercial pricing flexibility and protection of business-user information.

4. Effective Contestability

For Google in the EU search market, the market-level De Jure score is 1.00: at the De Jure layer the full catalogue of sixteen mechanisms applies and each scores 1 for the designated Google Search service. The De Facto Contestability Index stands at 0.52 as of July 2026 (Section V.A.3), so Effective Contestability is 0.76. The distance from 1 is an implementation shortfall, which Section V.A.3 identifies: the choice screen and the consent regime have improved user mobility and data conditions, while the structural barrier of a web-scale index and the still-unremedied self-preferencing in ranking hold the realized side down. With the legal ceiling already reached, every future De Facto gain of 0.1 raises the Effective Contestability score by 0.05.

Read against the December 2023 baseline, the same cell is a movement rather than a level. The legal layer stood at 0.38 and the realized layer at 0.28, giving a baseline Effective Contestability of 0.33. The legal layer has since moved +0.62 and the realized layer +0.24, so Effective Contestability has risen +0.43. Almost the whole of the legal movement is attributable to the Act: before it, ten of the sixteen mechanisms rested on transparency duties, on firm-specific Article 102 remedies, or on the general data-protection acquis, five rested on nothing, and only data portability was already complete.

B. EU — iOS App Distribution

Market: iOS app distribution. Assessed service: Apple App Store (Apple Inc.), designated as a Core Platform Service under the EU Digital Markets Act (DMA) on 6 September 2023.

1. Market scope and applicable mechanisms

The standardized market is iOS app distribution, assessed through the Apple App Store; the App Store and iOS are designated Core Platform Services under the DMA. Applicable mechanisms are identified separately for the two layers. At the De Jure layer, all sixteen mechanisms are scored: the DMA imposes a complete set of obligations on the designated services, so the market-level De Jure coding restates the full jurisdiction-level catalogue of Section IV. At the De Facto layer, the market has two gatekeepers. Under Section III.D each is coded separately on the same indicator set rather than averaged: Apple Inc., for the App Store on iOS, and Alphabet Inc., for Google Play on Android. Little is filtered out here — the DMA’s distribution, steering, payment, ranking and interoperability obligations all act on conduct observable in this market — so only cross-platform messaging connectivity (D4.2) is Not Applicable and the remaining fifteen indicators apply to both gatekeepers. Applicability is fixed before scoring and held constant across both periods and both gatekeepers (Section III.C); the evidence available to score each indicator differs between the two, and is reported in Section V.B.3.

2. Market-level De Jure Contestability

For the app-distribution market, the Existence, Scope, and Enforceability factors are carried over from the jurisdiction-level coding of Section IV.B unchanged; for each mechanism, the governing Digital Markets Act or supporting provision and the rationale as applied to the Apple App Store. The Act's obligations govern app distribution directly across almost the whole catalogue; only messaging interoperability operates in an adjacent market, and it is marked as such. The table reports two periods. The June 2026 column states the law as it stands; the December 2023 column scores the same sixteen mechanisms against the pre-existing acquis with the Digital Markets Act excluded, so that what the Act contributed can be separated from what the Union already required. That baseline is a counterfactual rather than a literal legal state: at December 2023 the Act was in force, applicable, and the service had been designated, but the Article 5, 6 and 7 obligations did not bind until 7 March 2024, so scoring the literal state would set Enforceability to zero on every Act-based cell and reach the same result by a longer route. Applicability is held constant across the two periods, as Section III.C requires.

The sixteen mechanism scores aggregate as elsewhere in the Index: within each dimension the mechanism scores are averaged, and the four dimension scores are then averaged with equal weight into the market-level index (Section IV.A). All four dimensions score 1.00 — User Mobility & Switching (D1), from #1–#5; Rival Access & Distribution (D2), from #6–#9; Platform Neutrality & Non-Discrimination (D3), from #10–#12; and Data, Interoperability & Infrastructure Access (D4), from #13–#16 — giving an equal-weighted market-level De Jure Contestability of 1.00.

At the December 2023 baseline the four dimensions stood at 0.40, 0.25, 0.50 and 0.25, giving an equal-weighted market-level De Jure Contestability of 0.35 — a movement of +0.65. No Article 102 decision bound Apple in the Union at that date: the music-streaming decision is dated 4 March 2024 and the mobile-payments commitments 11 July 2024. Apple's baseline therefore rests entirely on instruments of general application, but on more of them than search does, because the App Store is both an online intermediation service under the Platform-to-Business Regulation and an online platform under the Digital Services Act, and an online search engine is neither. Two mechanisms scored 1, seven scored 0.5 — six of them capped by Article 15, which delegates enforcement of that Regulation to Member States, the seventh resting on the lawful-basis requirement of the data-protection acquis — and seven scored 0, precisely the channels the Act was written to open.

3. Market-level De Facto Contestability

EU — iOS App Distribution

Gatekeeper: Apple Inc. Core platform service: iOS operating system and App Store (online intermediation). Designated 6 September 2023. Baseline December 2023; post-period July 2026.

Google Play App Distribution

Gatekeeper: Alphabet Inc. Core Platform Services: Android and Google Play. Designated 6 September 2023.

One indicator is Not Applicable and five are Applicable but Missing. Ten are scored. The market definition and indicator set are identical to those used in Section V.B.

Apple's De Facto Contestability Index rises from 0.11 to 0.53 (Δ +0.42) and Google Play's from 0.14 to 0.48 (Δ +0.34). Apple's pre-period score is the lowest of the three observations because it prohibited alternative distribution, link-outs and removal of pre-installed applications outright, leaving D2 and D3 at 0. Its gain is concentrated in D1, which reaches 0.60, the largest dimension movement in any observation (+0.50), carried by the only indicator to reach 1 through post-DMA change: from iOS 18.2, EU users can fully delete Safari and the App Store. Google Play starts higher on D1 and D2 because Android permitted sideloading behind a per-source permission from Android 8.0 and exposed centralised default-app management before the DMA. Its movement is concentrated in the two channels Android had closed: the restructure of 30 June 2026 separated the service fee, which applies to every payment route, from a billing fee that applies only when Play billing is used, and Article 5(2) consent flows for linking Google services were deployed in March 2024. Alternative distribution does not rise: Android permitted sideloading and alternative stores before the DMA, and the Registered App Stores programme announced on 4 March 2026 was not operative during the observation window. In its decision of 23 July 2026, which fined Google EUR 430 million under Article 5(4), the Commission found that Google prevents developers from freely communicating and promoting offers and concluding contracts in distribution channels of their choice, including third-party app stores. The two indices do not rest on equal evidence: five of Google Play's sixteen indicators are excluded as Applicable but Missing against two for Apple, and four of the five reflect documentation not obtained in this coding round rather than any feature of the market. Google Play's index therefore rests on a thinner evidence base than Apple's, not on a narrower set of applicable mechanisms.

4. Effective Contestability

For Apple in the EU iOS app-distribution market, all sixteen mechanisms operate and each scores 1 for the designated App Store, giving a market-level De Jure score of 1.00. The De Facto Contestability Index stands at 0.53 (Section V.B.3), so Effective Contestability is 0.77 — the higher of the two cells, because the DMA’s distribution obligations act directly on platform-imposed prohibitions whose removal is observable: third-party stores exist and core apps can be uninstalled, even though adoption frictions keep the realized side at half strength.

Against the December 2023 baseline, Apple’s legal layer stood at 0.35 and its realized layer at 0.11, giving a baseline Effective Contestability of 0.23. The legal layer has moved +0.65 and the realized layer +0.42, so Effective Contestability has risen +0.54, the largest movement recorded in the Index. The two markets’ legal layers moved almost identically, +0.62 in search and +0.65 in app distribution, although their pre-existing regimes were differently composed — Alphabet’s baseline rested substantially on firm-specific Article 102 remedies, Apple’s entirely on instruments of general application. The realized layers did not move alike, and the implementation shortfall is roughly three times larger in search. The baselines explain why: Apple’s starting position was one of outright prohibition, and prohibitions are cheap to remove and easy to observe being removed, whereas the search conditions the Act targets were already partially regulated in 2023 by the Shopping and Android decisions and have moved comparatively little since. Google Play has a De Facto baseline but no De Jure baseline, so its cell is reported as a level only.