Competition law still lacks an operational methodology for defining relevant markets for multi-sided platform businesses. Smith and Duke (2021) challenged the accepted dichotomy under which transaction platforms (where the customer groups transact with each other through the platform, e.g. payment-card networks) are assessed in a single market while non-transaction platforms (where each group deals only with the platform, e.g. social networks) are split into separate markets per customer group, arguing that a single market covering all user groups should be defined to assess the effects of conduct engaged in by a non-transaction platform business — but they supplied no operational test methodology, and enforcement practice and the companion literature leave open how courts and authorities should actually run market definition for platforms. The unresolved core is methodological: when a single market versus separate interrelated markets is appropriate; how the hypothetical monopolist test and its SSNIP instrument (a small but significant non-transitory increase in price) should be adapted when substitution patterns differ sharply across sides, one side pays a zero price or a two-part tariff, and indirect network effects (each user group's demand depending on the size of the other) are weak, asymmetric, or hard to quantify; how quality changes on zero-price sides should be measured (SSNDQ-type tests, based on a small but significant non-transitory decrease in quality); and how effects on non-focal customer groups can be incorporated without re-fragmenting the market. The European Commission's 2024 Market Definition Notice deliberately leaves the single-versus-multiple question case-by-case; UK, US, and Indian decisions diverge in opposite directions; and the most recent treatments (2023–2026) are proposals, not settled frameworks. The problem remains open.

Background

In competition (antitrust) law, most merger and abuse-of-dominance analysis begins by defining the relevant market — the arena of effective competition within which market shares are computed and market power presumptions are drawn. The standard economic tool for this is the hypothetical monopolist test (HMT): one asks whether a hypothetical monopolist controlling a candidate set of products could profitably sustain a small but significant non-transitory increase in price (a 'SSNIP', typically 5–10%); if it could, the candidate set is a market, and if not, the closest substitutes are added and the question is repeated. Multi-sided (two-sided) platforms complicate this picture: they serve two or more distinct customer groups — for example readers and advertisers on a news site, or consumers and insurers using a price comparison website — whose demands are linked by indirect network effects, meaning each group values the platform more when the other group is larger, and the platform sets its prices and quality on all sides jointly to maximize total profit. A foundational economics literature (Filistrucchi, Geradin, van Damme and Affeldt 2014) and subsequent practice distinguished transaction platforms, where the customer groups transact with each other through the platform (e.g. payment-card networks), from non-transaction platforms, where each group deals only with the platform (e.g. social networks or media); the received convention was to define a single market for transaction platforms but separate markets for each customer group of a non-transaction platform. The US Supreme Court adopted the transaction-side half of this convention in Ohio v. American Express Co., 585 U.S. 529 (2018), holding that in two-sided transaction markets 'only one market should be defined'. Three features of real platforms strain this framework. First, the two sides often face very different substitution possibilities, so a price test can point to different market boundaries on each side. Second, one side frequently pays a zero monetary price (the platform monetizes the other side), which makes a conventional SSNIP undefined on that side and has pushed authorities toward quality-based variants such as the SSNDQ (small but significant non-transitory decrease in quality) test. Third, indirect network effects can be weak, one-directional, or hard to quantify, making the feedback loops between sides hard to embed in the hypothetical monopolist's profit calculus. Against this background, Smith and Duke (2021) argued that the transaction/non-transaction dichotomy is unpersuasive and that competition authorities and courts should define a single market covering all customer groups whenever assessing the conduct of a platform business, because market definition is purposive, conduct aimed at one side affects the platform's integrated business, and separate side-by-side markets misclassify conduct (for example by reading a free offering as predatory pricing or a high advertising price as evidence of market power when the real constraint operates across the integrated business). Since then, the European Commission's revised Market Definition Notice (2024) stated that the Commission 'may define' either one market encompassing all user groups or separate interrelated markets per side, depending on the facts, and acknowledged both that indirect network effects make the SSNIP test more challenging and that zero monetary prices call for non-price parameters and possibly an SSNDQ; the UK Competition Appeal Tribunal in BGL (Holdings) Ltd v Competition and Markets Authority [2022] CAT 36 defined two separate markets for a price comparison website and crafted a 'special' SSNIP that introduces a positive price where the actual price is zero; and US courts continue to confine single markets to transaction platforms. How — and whether — a workable single-market methodology can be implemented in detailed platform cases is the open question.

Problem Statement

How can market definition for platform businesses be operationalized for use by courts and competition authorities? Smith and Duke (2021) argued that a single market covering all customer groups should be defined to assess platform conduct, including for non-transaction platforms (platforms where each customer group deals only with the platform rather than transacting with the other group through it), but supplied no operational test methodology; whether and how a single-market approach to platforms can be implemented in detailed cases remains open and contested in enforcement practice and the companion literature — the UK Competition Appeal Tribunal observed in BGL (Holdings) Ltd v Competition and Markets Authority [2022] CAT 36 (para 92) that whether a two-sided market is a single market, two separate markets, or circumstance-dependent is a novel question with no established approach, and Antuca, Niels and Ralston-Smith (2023) document the continuing absence of consensus. Specifically, a workable methodology is needed that determines: (i) under what conditions a single market encompassing all user groups should be defined versus separate (though interrelated) markets for each side — including cases where customer groups face very different substitution patterns, where one side pays a zero price or a two-part tariff (a fixed access fee plus a per-use charge), and where indirect network effects (each user group's demand depending on the participation of the other) are weak, asymmetric, or hard to quantify; (ii) how the hypothetical monopolist test should be applied to a platform — in its SSNIP form (a small but significant non-transitory increase in price) or an adapted form — given that the hypothetical monopolist sets prices and quality jointly across sides and that a conventional price increase is undefined or meaningless on a zero-price side; (iii) how changes in quality or other non-price parameters on zero-price sides should be measured and tested (for example through SSNDQ-type constructions, which test a small but significant non-transitory decrease in quality instead of a price increase) in a way that is implementable with real-world evidence; and (iv) how effects on non-focal customer groups are incorporated into the market-definition exercise without reintroducing the per-side fragmentation the single-market approach was meant to avoid. These four components serve one objective: a decision-relevant, evidence-based protocol for drawing market boundaries in platform cases, grounded in both the economics of two-sided markets and the doctrinal constraints of competition law (including the cited decisions and the 2024 Market Definition Notice).

References

  1. Rhonda L. Smith and Arlen Duke (2021), 'Platform businesses and market definition', European Competition Journal 17(1):93–117. DOI: 10.1080/17441056.2020.1851477, https://doi.org/10.1080/17441056.2020.1851477
  2. Lapo Filistrucchi, Damien Geradin, Eric van Damme and Pauline Affeldt (2014), 'Market Definition in Two-Sided Markets: Theory and Practice', Journal of Competition Law & Economics 10(2):293–339. DOI: 10.1093/joclec/nhu007, https://doi.org/10.1093/joclec/nhu007
  3. Gunnar Niels (2019), 'Transaction versus Non-Transaction Platforms: A False Dichotomy in Two-Sided Market Definition', Journal of Competition Law & Economics 15(2–3):327–357. DOI: 10.1093/joclec/nhz016, https://doi.org/10.1093/joclec/nhz016
  4. Ohio v. American Express Co., 585 U.S. 529 (2018). https://www.law.cornell.edu/supremecourt/text/16-1454
  5. Gunnar Niels and Helen Ralston (2021), 'Two-sided market definition: some common misunderstandings', European Competition Journal 17(1):118–133. DOI: 10.1080/17441056.2020.1851478, https://doi.org/10.1080/17441056.2020.1851478
  6. Helen Ralston, Will Leslie, Eoin O'Reilly, Robin Masters and Victoria Heinen (2021), 'Choosing the right filter: Market definition in two-sided markets', Competition Law & Policy Debate 6(3–4):67–73. DOI: 10.4337/clpd.2021.03-04.07, https://doi.org/10.4337/clpd.2021.03-04.07
  7. Lukas Solek (2021), 'Need to Revise or Apply the Concept of Market Definition with a View to "Zero-Price" and Overarching Markets', Journal of European Competition Law & Practice 12(8):593–603. DOI: 10.1093/jeclap/lpab005, https://doi.org/10.1093/jeclap/lpab005
  8. James Mancini and Cristina Volpin (2021), 'Quality Considerations in Digital Zero-Price Markets: OECD Background Paper', SSRN. DOI: 10.2139/ssrn.3862529, https://doi.org/10.2139/ssrn.3862529
  9. Shilpi Bhattacharya and Pankhudi Khandelwal (2022), 'Judging a book by its cover?: analysing the Indian approach to defining platform markets', International Review of Law, Computers & Technology 36(3):330–351. DOI: 10.1080/13600869.2022.2030032, https://doi.org/10.1080/13600869.2022.2030032
  10. BGL (Holdings) Ltd & Others v Competition and Markets Authority [2022] CAT 36, Competition Appeal Tribunal, 8 August 2022. https://www.catribunal.org.uk/sites/default/files/2022-08/20220808%201380%20BGL%20v%20CMA%20Approved%20Judgment%20%5B2022%5D%20CAT%2036%20-%20Website%20%281%29.pdf
  11. Andreea Antuca, Gunnar Niels and Helen Ralston-Smith (2023), 'Compare the Markets: Two-Sided Market Definition in the Comparethemarket Case', CPI Antitrust Chronicle, January 2023. https://www.competitionpolicyinternational.com/wp-content/uploads/2023/01/5-COMPARE-THE-MARKETS-TWO-SIDED-MARKET-DEFINITION-IN-THE-COMPARETHEMARKET-CASE-Andreea-Antuca-Gunnar-Niels-Helen-Ralston-Smith.pdf
  12. European Commission (2024), 'Commission Notice on the definition of the relevant market for the purposes of Union competition law', Official Journal of the European Union C, C/2024/1645, 22.2.2024. Full text (Publications Office Cellar): http://publications.europa.eu/resource/cellar/db5ffcb5-d125-11ee-b9d9-01aa75ed71a1.0006.03/DOC_1, landing page: https://op.europa.eu/en/publication-detail/-/publication/db5ffcb5-d125-11ee-b9d9-01aa75ed71a1/language-en
  13. Niccolò Galli and Pier Luigi Parcu (2025), 'Multi-sided platforms and the relevant antitrust market', in Research Handbook on Competition and Technology, Edward Elgar Publishing. DOI: 10.4337/9781035302642.00010, https://doi.org/10.4337/9781035302642.00010
  14. Anzhelika Gerasymenko and Vasyl Hanovskyi (2025), 'Methodology of Market Definition: Case of Digital Multi-Sided Platforms', Yearbook of Antitrust and Regulatory Studies 18(31). DOI: 10.7172/1689-9024.yars.2025.18.31.1, https://doi.org/10.7172/1689-9024.yars.2025.18.31.1
  15. United States v. Google LLC, No. 1:23-cv-00108-LMB (E.D. Va., Memorandum Opinion, 17 April 2025), Document 1410. https://www.adexchanger.com/wp-content/uploads/2025/04/Judge-Leonie-Brinkema-ad-tech-decision-DOJ-Google.pdf
  16. Emanuela Lecchi (2026), 'Microsoft/Activision Blizzard and Booking/eTraveli: an ecosystem market definition in five takeaways', Journal of Antitrust Enforcement 14(1):34–59. DOI: 10.1093/jaenfo/jnaf029, https://doi.org/10.1093/jaenfo/jnaf029
  17. Helen Ralston-Smith and Gunnar Niels (2026), 'A practical guide to market definition for multi-sided platforms: how can we build on the principles in the 2024 Market Definition Notice?', Journal of Antitrust Enforcement 14(1):154–170. DOI: 10.1093/jaenfo/jnag001, https://doi.org/10.1093/jaenfo/jnag001