Our Research

Social Learning and Word of Mouth Networks

Percolation Model Diagram

Social Learning and Word of Mouth Networks

Arthur Campbell

A fundamental issue for economists is understanding the ways social learning, via word-of-mouth (WOM) between friends, affects demand; and hence, a firm’s pricing and advertising strategies. I have explored a couple of different questions in this area.

Percolation Model of Demand

Individuals tend to be more likely to talk about products they buy than those they do not purchase. Conversations often convey information which is essential for someone else to consider buying a product, such as its existence.  In this setting demand for a product is determined, in part, by how far this type of information diffuses across a population. We describe demand in this type of environment using a percolation process on a random graph where the probability a person talks about a product is a function of how much they like the product and the price.

We consider three sets of questions: (1) What properties of the social network and consumers’ preferences for the product make demand more elastic than in a population of fully informed individuals; (2) How does the price change when the social network is: more connected? has more clustering? more homophily?; (3) Are the people with the most friends the optimal targets of advertising?

A Model of Buzz

Many marketing practitioners recommend strategies which limit access to information in order to spur word of mouth. An interesting aspect of these strategies is that they seem to contradict the intuition that wider exposure to product information will lead to more word of mouth and a larger fraction of the population eventually holding the information. We explore why a firm that is seeking to maximize the number of people who possess information about its product may undertake strategies which actively restrict early access to the information.

Central to our explanation is the motivation of individuals to engage in word of mouth communication and to acquire product information. We propose a model where individuals care about how others perceive them (X is technologically savvy; X knows a lot about music/restaurants/wines/Ö). In a signaling equilibrium of our model individuals convey their knowledge of product information through word of mouth communication as a means of signaling their type to others. We find that a firm will optimally restrict (increase the cost of) information to certain types of individuals. We also find that untargeted advertising may crowd out an individual’s incentive to engage in word of mouth, hence a commitment not to advertise is valuable for the firm, and the ability of a firm to target advertising at well connected individuals may reduce the incentives of individuals to engage in word of mouth.