Multi-sided Markets and Network Effects
Multi-sided markets are not a new phenomenon. Newspapers financed by advertising have existed since the nineteenth century.7 Payment cards were introduced in 1950 and the explosion of videogames took place during the 1980s. However, it was only when the internet became widespread, in the mid-1990s, that multi-sided markets proliferated, grew multibillion memberships, and attracted the interest of investors and economists.
Early this century, Nobel Laureate Jean Tirole and his peer Jean-Charles Rochet published their seminal work Platform Competition in Two-Sided Markets.’1 In that paper, seven “mini cases” were analyzed, including media, credit cards, and video games. The new paradigm built by Rochet and Tirole has spawned a substantial amount of industrial organization literature.9
Over the years, a consensus has developed that multi-sided markets typically involve two or more distinct types of users, interacting through a third party, the platform.10 Key elements are the relevance of positive indirect network effects, which have to be strong enough to affect business conduct," as well as the leading role that the platform plays in the distribution of such benefits across the ecosystem of interacting parties through pricing and other design decisions. However, let us briefly explain, step by step, how this theory has been built.
It is not by chance that Tirole had originally worked on, among other things, the regulation of network industries, particularly telecommunications. In fact, in Platform Competition in Two-Sided Markets, Rochet and Tirole explicitly stated that their theory builds on network economics.12 We have already explained how, in network industries such as telecommunications, transport, and electricity, the larger the number of users of a specific infrastructure or system, the higher the benefit for all other users. This is the so-called “network effect.” It is considered an externality as it is a benefit that does not derive from the mere single decision of the party that contracts a service.
The benefits derived from network effects can easily be underestimated. They were early identified by the most successful leaders in network industries, such as Theodore Veil, the president of AT&T at the turn of the twentieth century (see his story in Chapter 6) and David Sarnoff, leader of the radio and television network NBC. Sarnoff observed that the value of his network seemed to increase proportionately to the number of users, but he still underestimated the value of the network effects. Robert Metcalfe, working on the development of Ethernet in the 1980s, later concluded that the value of a communications network would be not be the number of users, but the square of the number of users. A network of ten telephones would not have a value of ten, but a value of 100 - the number of potential connections between users. This has been known as Metcalfe’s Law. Furthermore, it was proposed in the early 2000s that as different sub-groups with more than two users could be formed, the value of a network of ten users could be as high as 501, the total number of potential connections including the variable numbers of users (Reed’s Law).13
The key novelty in the analysis of multi-sided markets is that network effects are not limited to systems with a single category of users, as is the case with telephone networks. These are called direct network effects. The interaction of different types of users can also generate network effects. This is the case with advertisers and newspaper readers, merchants and purchasers using payment cards, and videogame developers and players. These are the so-called “indirect network effects.”
Indirect network effects build up in a different way than the more traditional direct network effects. The key difference is that a balance must be reached and then sustained that takes into consideration the interests of all the parties involved in the multi-sided markets. It is not sufficient to sell an innovative good or service at an attractive price, as is the case in traditional industries. It is not sufficient to attract a single category of users to generate the direct network effects. An attractive offer must be provided to all the parties in the multi-sided-market, including to the intermediary itself.
On one hand, negative same-side effects can emerge. If one side of the multi-sided market grows too large, positive network effects may become jeopardized, as is the case with nightclubs, for example. Nightclubs can be understood as multi-sided markets that facilitate the interaction of opposite-sex individuals. A growing number of men might produce negative same-side effects. This is why it is an established practice to allow women free entry to clubs (another example of zero-pricing), while men are charged a fee to enter.
On the other hand, negative cross-side effects can also be identified. In some multi-sided markets, a growing number of users can make matching more complex, more expensive, or slower, or it can just degrade the service provided by the intermediary. This is the case with advertising in the media industry. Advertisers can be attracted by large audiences, but the introduction of too much advertising, such as in radio or TV shows, can discourage audiences from participating in the market; that is, from following the shows.
For a multi-sided market to succeed, positive network effects must be created inside each group (positive same-side effects) and across the participating groups (positive crossside effects). Furthermore, some benefits must be captured by the platform that created the market and that curates it to keep the right balance. Thus, the active role of the platform is key in multi-sided markets.