The article – which can be found here – seeks to analyse, from both an economic and a legal perspective, how relevant product markets for multi-sided platforms are defined under  European competition law.

It starts from the premise that delimiting the relevant product market in the context of multi-sided platforms is a particularly delicate and complex task. This is because well-established economic approaches for defining product markets were not designed to deal with  multi-sided platforms, which address at least two distinct groups of customers. The paper further argues that multi-sidedness can be considered a question of degree. Therefore, for the purpose of using standard concepts of market definition, it is necessary to distinguish between different types of multi-sided platforms – the stronger the indirect network effects involved, the more the market definition will deviate from traditional approaches.

The paper is structured as follows:

  • Section 2 puts multi-sided platforms in an economic context, building on the literature on network effects. These effects have been repeatedly analysed in these emails: ‘From an economic perspective, network effects can be called positive externalities because joining a platform causes benefits for third parties without these benefits being reflected by the market price. These externalities are internalized by platforms. The business model of platforms, therefore, consists in internalizing externalities by connecting a maximum number of users.’ Multi-sided platforms enable distinct groups of customers to interact with each other – and hence give rise to indirect network effects. These platforms have the possibility to set different prices for each group of customers; due to indirect network effects, a price increase for one group has, however, repercussions on the other groups involved. If network effects are strong enough, the profit-maximizing price may also be below the marginal cost of supply or even negative.

The author argues that neglecting these characteristics of multi-sided platforms would hinder their proper competitive assessment. At the same time, it is important to keep in mind that multi-sided platforms are a theoretical economic concept. For legal analysis it is the details of the particular case that matter, and it is better to conceive one-sidedness and multi-sidedness as the extreme points of a continuum: ‘it is the task of legal scholarship to identify cases in this continuum and regroup them in order to establish a legal rule how to deal with multi-sided platforms.’

  • Section 3 explores how the economic features of multi-sided platforms make them different from traditional businesses, and how this impacts market definition. First, market power is not inextricably linked to a decrease in welfare in the context of multi-sided platforms: due to indirect network effects, a platform’s value may increase for its users along with the platform’s size. When indirect network effects are strong enough, the resulting efficiencies could be such as to outweigh the welfare effects of a price increase on the other side of the platform – which may lead to the paradoxical effect of price increases being pro-competitive [NOTE: this has become apparent in recent cases concerning payment cards. It was argued by one of the parties in the Ohio v Amex case; and was a topic of discussion in the recent Sainsbury v Visa case in the UK].

Second, it is well established that multi-sidedness affects market definition. Multi-sidedness – even in small doses – means that a conventional one-sided market definition exercise would fail to mirror the market’s competitive conditions. There are at least three implications of this which should be emphasised: (i) each of the groups involved in a platform has to be considered for market delimitation; (ii) economic tools which are traditionally used for market definition have to be modified when applying them to multi-sided platforms; (iii) the fact that no prices are charged for one platform-side is not tantamount to absence of competitive harm, since anticompetitive effects can be felt on the other side of the platform / market.

  • Section 4 reviews a number of EU decisions, and concludes that the EU distinguishes between situations that require the market to be defined by reference to the platform as a whole, and situations that require the market to be defined for each side of a platform. Finding inspiration in the credit card and after-markets case law, the author concludes that: ‘a single market for both platform-sides under investigation may be defined when a sufficient number of participants on one platform-side would switch to another platform if there were a moderate price increase for the products or services on the other platform-side.’

The paper also argues that the EU has broadly followed the guidance provided in the economic literature as regards two-sided markets. Nonetheless, the author identifies three conceptual shortcomings that make the available economic theories difficult to put into legal practice: (i) they rely on a rather vague concept of transaction; (ii) they rely on a binary distinction between transaction markets and non-transaction markets which too simplistic and unable to explain most market scenarios; (iii) they are nor concerned with the development of a legal test. As such: ‘it is still necessary to develop a legal test that integrates the essence of the ideas put forward in the economic theories – but tries to avoid the identified weaknesses’.

  • Section 5 explains that the main failings of existing approaches derive from a failure to  acknowledge that the strength of indirect network effects / interconnectedness is a matter of degree: ‘Even when the focus is on platforms requiring the delineation of various markets, there remains a variety of different constellations ranging from quasi absent network effects to moderate network effects; these differences must be considered in the competitive assessment. Thus, it is necessary to gauge the platform’s competitive conditions on a case-by-case basis.’

Likewise, as regards abuses of dominant positions: ‘in the context of multi-sided platforms, new theories of harm must be developed […] when the markets involved are associated, necessitating all markets of a platform to be subjected to thorough competition law scrutiny.’ For multi-sided platforms, a strong position on one platform side is not tantamount to the capacity to behave independently in the market. Even moderate indirect network effects may effectively limit an undertaking’s ability to restrict the quantity or quality of output. Market dominance may not exist even when the market shares on one platform side are high and competition is weak. On the other hand, indirect network effects may enhance the competitive position of an undertaking when it holds a strong position on other platform sides. Hence, for determining whether a multi-sided platform is dominant, its competitive position on all platform sides must first be assessed, and then the interrelatedness / the strength of network effect between the platform sides has to be gauged.


Comment; This is a competition paper after my own hear – outlining the challenge we face, identifying the contribution that economic analysis can provide for legal practice, but ultimately focusing on the legal implications of multi-sidedness for competition law. Having said that, the analysis stops at a very high level – the article does not really provide much guidance on how to pursue the legal test in practice. As regards this, and for anyone looking for more detail, I would suggest you read the papers presented in our roundtable on ‘Rethinking the use of traditional antitrust enforcement tools in multi-sided markets’ which can be found at

Author Socials A weekly email with competition/antitrust updates. All opinions are mine

What do you think?

Note: Your email address will not be published