Competition between global and local online social networks

TL;DR

This paper models the competition between international and local online social networks as a complex ecosystem, revealing conditions under which local networks can survive or be overtaken by global giants like Facebook.

Contribution

It introduces a scale model of digital world competition, highlighting the role of network fitness heterogeneity and social ties in network dominance dynamics.

Findings

International networks can outcompete local ones under certain conditions.

Local networks may persist or dominate depending on model parameters.

Empirical data suggests Facebook's overtaking was unlikely without specific factors.

Abstract

The overwhelming success of online social networks, the key actors in the Web 2.0 cosmos, has reshaped human interactions globally. To help understand the fundamental mechanisms which determine the fate of online social networks at the system level, we describe the digital world as a complex ecosystem of interacting networks. In this paper, we study the impact of heterogeneity in network fitnesses on the competition between an international network, such as Facebook, and local services. The higher fitness of international networks is induced by their ability to attract users from all over the world, which can then establish social interactions without the limitations of local networks. In other words, inter-country social ties lead to increased fitness of the international network. To study the competition between an international network and local ones, we construct a 1:1000 scale model of the digital world, consisting of the 80 countries with the most Internet users. Under certain conditions, this leads to the extinction of local networks; whereas under different conditions, local networks can persist and even dominate completely. In particular, our model suggests that, with the parameters that best reproduce the empirical overtake of Facebook, this overtake could have not taken place with a significant probability.