The term ‘two-sided market’ originates from the economic literature and does not bear any direct legal consequence in competition law. The term rather offers an intelligible label for the economic phenomenon of indirect network effects that scholars only since the early 2000s have described in concise economic models. The wide dissemination of the term is mainly attributed to Rochet & Tirole (2003). Nowadays, the term ‘multi-sided’ is often used instead of ‘two-sided’, and/or the term ‘platform’ instead of ‘market’. Many, if not most, platforms have more than two distinct customer/user groups. The number and identity of the different platform ‘sides’ is not necessarily a universal characteristic of the business activity at stake but may result from a deliberate choice made by the platform operator. The platform’s competitors may differ across these platform sides. For each side of a platform a separate relevant market may be defined, for example a reader market and an advertising market.
The presence of indirect network effects between user groups affects the price setting mechanism and the competitive interaction in these markets. When cross-group externalities are significant, this may have relevant consequences for the suppliers’ incentives and thus for the competitive analysis. When user group A places higher value on the participation of user group B than vice-versa, it will be rational for the platform operator to offer lower prices to user group B, and this may go as far as offering the service for free to users B or even rewarding them for their participation with payments or other benefits. This means that in some cases one or several user groups will not be charged for the platform service. The presence of significant indirect network effects should be reflected in all parts of the competition law analysis, i.e. market definition, market power analysis and effects analysis. Price-cost tests developed for identifying predatory pricing or excessive pricing in one-sided markets will not yield reliable results when applied without modification for multi-sided platforms. Other things being equal, positive bidirectional network effects between the two or more sides tend to increase both market power and the magnitude of anticompetitive effects.
Although the economics of indirect network effects are nowadays common-sense among economists, practitioners and courts still struggle with their application in competition law. So far, the three most relevant issues concerned zero-price markets, defining separate markets for each ‘side’ or not, and implications of the increased analytical complexity for the burden of proof.
It can be frequently observed that the platform operator charges only one customer group while the service is offered for free to another customer group (‘zero-price markets’). There has been some debate as to whether free-of-charge offers should be considered as (part of) an antitrust market, and some courts even decided that they should not. However, the economic answer, and therefore also the sensible competition law answer to this question seems rather clear. In multi-sided markets, where the provision of a free service is part of a business activity pursuing commercial purposes, zero-pricing should not hinder a relevant market to be defined. With positive indirect network effects, setting a price of zero for one customer group may make perfect sense for the platform provider if it finds it preferable to charge another customer group. There may be competition for non-charged users, and this competition can be restricted through mergers, abusive practices or a cartel. The seminal ‘zero-price market’ case is Facebook where the German Federal Court of Justice decided that inadequate privacy conditions by a dominant provider can be exploitative vis-à-vis private users even though the service is offered to them for free.
As multi-sided platforms involve distinct groups of customers, there are in principle two alternative market definition options: defining separate markets for different customer groups or defining a single market encompassing all customer groups. As a first step, the competitive landscape on each ‘side’ of the platform should be identified one after the other. In comparing the competitive forces on each of these sides, it is easy to assess whether the set of relevant product substitutes/competitors, the single- or multi-homing of the customer groups or the geographic scope differ across platform sides. In case of significant differences, typically separate markets should be defined. Conversely, if these factors do not differ across market ‘sides’, a single market encompassing all customer groups can be defined. In particular, this may be the case where the different groups are inseparably linked by a platform interaction, for example a transaction platform or a matching platform.
Effects analysis is more complex in multi-sided markets compared to single-sided markets. While there is broad agreement that, in principle, multi-sided markets may require more scrutiny from antitrust authorities than one-sided markets, there is a question on whether and how established antitrust tests and their implications for the allocation of the burden of proof remain valid in platform settings. In a controversial ruling, the US Supreme Court decided in Ohio v. American Express that in case of a transaction platform it was not sufficient for the plaintiff to show harm on one transaction platform side. Legislation to reverse this decision is currently being discussed in the US.
While the term two-sided market can be viewed as synonymous to the terms ‘platform’ and ‘multi-sided market’, it should be distinguished from the more recent concept of ‘digital ecosystems’. A digital ecosystem can be understood as a broader conglomerate of several distinct services/platforms whereas the term ‘multi-sided market’ refers to one service/platform only.