Modeling influence diffusion in social networks is an important challenge. We investigate influence-diffusion modeling and maximization in the setting of viral marketing, in which a node’s influence is measured by the number of nodes it can activate to adopt a new technology or purchase a new product. One of the fundamental problems in viral marketing is to find a small set of initial adopters who can trigger the most further adoptions through word-of-mouth-based influence propagation in the network. We propose a novel multiple-path asynchronous threshold (MAT) model, in which we quantify influence and track its diffusion and aggregation. Our MAT model captures not only direct influence from neighboring influencers but also indirect influence passed along by messengers. Moreover, our MAT framework models influence attenuation along diffusion paths, temporal influence decay, and individual diffusion dynamics. Our work is an important step toward a more realistic diffusion model. Further, we develop an effective and efficient heuristic to tackle the influence-maximization problem. Our experiments on four real-life networks demonstrate its excellent performance in terms of both influence spread and time efficiency. Our work provides preliminary but significant insights and implications for diffusion research and marketing practice.

Introduction

People live in various social networks, and share information and ideas with friends in the form of word-of-mouth (WOM) communication. New technologies and various social media rapidly penetrate into every aspect of our daily life, and provide us new channels and great convenience to exchange information and express opinions. They disseminate massive volumes of information over different social media, and spread influence to each other. As social media becomes prevalent, its influence on business, politics and society becomes evident and significant. How new innovations, behaviors, and diseases spread through social networks has a long history of study in social sciences. Research in this area has exploded and drawn considerable attention from many disciplines over the last decade. Many models of information and influence diffusion have been proposed for a wide variety of applications, such as viral marketing (Kempe et al. 2003; Bhagat et al. 2012), cascading behavior and prediction (Leskovec et al. 2007; Cheng et al. 2014), information spreading (Morales et al. 2014), outbreak detection (Leskovec et al. 2007), etc. In this paper, we look into modeling influence diffusion in the setting of viral marketing in business. In essence, viral marketing is a process of influence diffusion over social networks. An effective viral marketing campaign requires that marketers identify individuals with high social networking potential. A general setting can be depicted as follows: a company would like to market a new product, in the hopes that it would be adopted by as many people as possible in a target social network. The company chooses a small number of influential individuals as the initial adopters/seeds by giving them free/discounted samples of the product and encourages them to recommend the product to their friends, hoping that their friends would be influenced to purchase the product, then influence their friends, and so on. As a result, the influence would propagate through the network and trigger widespread adoption in the end of the diffusion process. Viral marketing is termed to describe such a marketing technique that induces the users in a social network to pass on a marketing message (viral ad) to others so as to achieve the largest influence spread in terms of product sales or brand awareness. Viral marketing is driven by WOM communication and enhanced by network effects. Survey-based statistical research has shown very strong support for the hypothesis that network linkage can directly affect product/service adoption (Hill et al. 2006; Iyengar et al. 2011). The crucial factor to the success of viral marketing is for the marketers to identify the most influential set of initial adopters. Researchers investigated different approaches to seeding campaigns for marketing practice (Hill et al. 2006; Iyengar et al. 2011; Libai et al. 2013; Rand and Rust 2011; Peres 2014). These studies significantly enhance our understanding of WOM behavior and effects on promoting viral marketing. However, the approach these researchers commonly use or rely on is explanatory modeling.