Viral Marketing: New Face of Social Network Marketing

Many companies are using social media websites and data for marketing purposes. For instance, Facebook uses the data collected on users to run an advertising based business model. However, research has indicated that click rates on targeted website advertisements is quite low and hence investments in such campaigns may not provide adequate ROI for marketers. This seems intuitive since most people do not visit social network websites to gather information on products and services.

The following link (http://hbswk.hbs.edu/item/6187.html) talks about recent research in the field of marketing applications of social networks. The research quantifies the impact social networks might have on individual purchase decisions. The above article goes on to argue that the best way to use the power of social networks in marketing could be viral marketing.

Viral marketing is the technique of using existing social networks to increase brand awareness and adoptions. It’s so called because word of mouth or network effect which disperses information in a way similar to a self replicating viral process. As long as the reproductive rate (the average number of susceptible users with whom the idea is shared) is greater than 1, the number of people to whom the marketing effort reaches grows exponentially (atleast in the inital phase). The goal of marketers in designing such a campaign is to identify individuals with high Social Networking Potential and create messages which appeal to this segment.

With the advent of internet and the wealth of social information and preferences available about people, it has become technically possible to technically isolate the focal point members of any viral campaign, the "hubs" who are most influential.

This paper by Pedro Domingos from University of Washington (http://www.cs.washington.edu/homes/pedrod/papers/iis04.pdf) talks about the feasibility and development of social network models to quantify the network value of customers. The paper argues that it is computationally feasible to build and analyse social network models with millions of nodes to come up with predictions and aid in marketing decision making. Given the model, the question boils down to a linear optimization problem i.e. choosing a set of nodes to target for marketing in order to maximise net profits. Although with larger networks precise solutions could be prohibitively costly, it is still possible to use heuristics to derive an approximate solution which could provide a large boost in profits.