Network saturation

When evaluating the anti-competitive effect of network scale, the first factor to consider is "network saturation"-that is, the scale of the network is sufficient to promote complementarity and generate positive spillover effects. Once this threshold is exceeded, it is almost impossible to obtain additional user benefits. A lower network saturation threshold means that there will be ample opportunities to customize network services through differentiated functions to meet the specific needs of sub-group users. This means that, in principle, this differentiation can promote market access and platform competition.

The process of early operators in the market competing with their successors for the dominance of the next-generation platform technology (such as in the US home video game market) shows that early operators will quickly reach network saturation, which makes complementary product suppliers turn to smaller but developing platform networks. companies. Similarly, complementary suppliers will be more likely to launch their best and most innovative products on smaller platforms. Compared with large dominant platforms, these platforms provide a more relaxed, clearer, and possibly more conducive environment for their products. At the same time, a similar effect was shown in terms of the innovation of user-created content.

Diminishing returns and local preferences

The research also shows that network effects may be diminishing at a certain node, that is, under certain conditions, it will show a trend of diminishing returns to scale.

Panico and Cennamo model indirect network effects as a function of end users' preference for innovation capabilities (quality and novelty of platform complements) and ecosystem scale (network scale/number of complementary products). They show that the intensity of indirect network effects will vary according to the composition of platform user types at a specific time. There may be many users but low intensity, or there may be few users but high intensity. For example, when users value innovation capabilities more, indirect network effects will be strong in the early stages of the platform market, and then will gradually weaken as the scale of the platform expands.

Similarly, Casadesus and Halaburda said that platforms can create more value for users by limiting the number of available applications. In this case, even if there is an indirect network effect, the application will show a direct network effect- that is, users will be more willing to use the same applications as other users (for example, users entertaining in the same massively multiplayer online game or using the same video conferencing application) due to the increase in consumer income from interacting and communicating with other users.

In this case, users will face a coordination problem, that is, users may tend to use more non-essential applications instead of just a few applications that can meet their needs. In this way, the platform needs to coordinate the demand side and supply side of the platform market by limiting the network scale, rather than simply pursuing the expansion of its network scale.

Taking online dating platforms as an example, due to the asymmetry of user preferences and "attraction", a smaller but more homogeneous user set may be easier to successfully match than a large user set. This explains why eHarmony, which has a smaller network, can successfully compete with Match.com, the leading online dating platform, because it uses a differentiated approach, with more "boutique" members and selective matching services as its winning weapon.

In fact, many platforms may be affected by the effect of diminishing returns, especially when users tend to interact with a specific, small-scale user group. This tendency is called "local preference", which widely exists in the social networks of some user groups.

Intraperiod and intertemporal network effects

Another factor that needs to be considered when evaluating the strengthening or weakening of network effects is the difference between intraperiod (or, one-time) and intertemporal network effects. Even if there is a strong network effect, the huge user network itself does not necessarily create barriers to market access. The reason is that, in most cases, strong network effects are short-lived. Within a certain period of time, there is a strong attraction between user groups. But the network effect of this intensity is not lasting.

In other words, from the perspective of time, the network effect is diminished to weak or even ineffective. In other words, the network effect has a strong attenuation effect- although the number of active users in one market creates a strong attraction for (active) users in the other market, the value of the total number of (registered) users on the platform network to attract users in the other market is very limited. Therefore, there are important differences between active and passive users, new users and old users, and traffic and existing users.

Uber is a typical example. The strength of its network effect is not constant over time, and may rapidly weaken with the passage of waiting time. The analysis found that when the waiting time is more than 5 minutes, consumers will have an indifferent attitude towards the choice of taxi platform. Similar effects will appear on social networks and media platforms. Therefore, the platform must continue to provide users with new content and encourage interaction between users, otherwise, users may lose interest in their service platform.

Switching costs, migration, and multihoming

Multihoming can reduce the possibility of a platform dominating the entire market. For software platforms, there is evidence that when the multihoming costs of software platforms are lower than the platform-specific innovation costs, the network scale of the platform can create positive spillover effects for competitive technologies. In contrast, exclusive transactions can help small platforms achieve differentiation on the basis of content, thereby competing with large platforms. For smartphones, multihoming has a limited impact on platform competition and market share between Android and iOS, because popular applications will be put into operation on both platforms.

When considering the impact of multihoming on competition (and its impact on consumer welfare), we should also evaluate the hidden costs associated with the potential loss of innovation and quality that are complementary to multihoming. Precisely because multihoming complementarity can reduce the differentiation between platforms, platform owners may invest a lot of money in technical design to gain a competitive advantage. Differences in platform architecture, such as differences in operating systems and interfaces, may therefore become obvious.

These differences will in turn prompt suppliers of complementary products to make design trade-offs, because compared with single-homed platforms where it is easier to design core technologies, multi-homed platform architecture is more complex, which directly leads to an increase in the cost of multi-homed design. Moreover, the quality performance of multihoming complementary products on multihoming platforms may be lower than the singlehoming platforms to which they were originally designed. The complexity of the platform is indeed an important strategic dimension, because it affects the extent to which new market entrants disrupt the existing platform ecosystem and may cause users to switch to other platforms.