Abstract:

Viral marketing is widely used as a social marketing strategy in the era of social commerce. Previous research applied agent-based model to simulate the social influence and explored the viral marketing performance under social network. But transactional community, which developed in the era of social commerce, is different from traditional social network in its network structure (dual network system) and members’ psychologies (informational influence). This research, based on the platform of largest transactional website in China (Taobao.com), applied web crawling technique to extract data of consumers’ social relation and activities from two transactional communities (Taobao Bangpai) and used ABMS to explore the effects of social influence, homophily and network topology on viral marketing performance. The results shows: (1) regardless of heterogeneous customer value, the increase in seed size (from 0.5% to 5%) would not increase the performance of viral marketing strategy (due to the overlap effect of each seed)； (2) when the density is controlled in random network, the performance of viral marketing based on social influence is significantly different from homophily mechanism； (3) when the network topology is adjusted into real world relationships, the difference is still significant. Especially, the homophily mechanism would only speed up the viral marketing process in an early phase, it cannot spread as wide as social influence do in the long term. This is because the constraints of community activity in homophily mechanism； (4) the experiment of exploring the effect of types of seeds on viral marketing performance shows that, compared with randomly chosen seeds, the seeds who possess central position (degree centrality) both in two layers of networks can significantly speed up the viral marketing process until the point of saturation effect.

Key words: transactional community, viral marketing, social influence, homophily, ABMS