Publication details
Viral Video Diffusion in a Fixed Social Network: An Agent-based Model
| Authors | |
|---|---|
| Year of publication | 2014 |
| Type | Article in Proceedings |
| Conference | Procedia Economics and Finance 12 ( 2014) |
| MU Faculty or unit | |
| Citation | |
| Web | Odkaz na článek v časopise na ScienceDirect |
| Doi | http://dx.doi.org/10.1016/S2212-5671(14)00353-0 |
| Field | Economy |
| Keywords | viral video; viral marketing; social network; agent-based model |
| Description | Agent-based computational papers on viral marketing have been so far focused on the study of the word-of-mouth knowledge diffusion, and hence merged the decisions to adopt a product and to share information about it. This approach does not seem to capture well the properties of viral videos which are shared with no regard whether the sender has adopted the product. This paper presents the first model of such knowledge diffusion. The model consists of an artificial social network (a mix of small world and power network) that mimics the properties of empirical social networks and a model of node activation where every node that viewed the viral video shares it with a random subset of her neighbors just once. The results of the simulation show that there is a phase transition: in one phase, almost no agents view the viral video, in the other one, a great part of the whole population does. When the second phase occurs, the diffusion of the knowledge in time resembles that of Bass model. What phase occurs and how many agents view the content depend above all on how “catchy” the video is. Other marketing practices as selecting the seed of the first addressed agents are of secondary importance. The marketer can choose between addressing fewer more connected agents or more agents with fewer connections. If the video is “catchy”, then a small number of the first addressed agents is sufficient even when the agents are selected randomly. |
| Related projects: |