Role of network structure and network effects in diffusion of innovations

Abstract

Why does diffusion of innovation sometimes propagate throughout the whole population, and why at other times does it halt in its interim process? The current paper provides a potential answer to this question by developing a simple computational model of social networks. The proposed computational approach incorporating small-world graphs enables the authors to find that diffusion of innovation is more likely to fail in a random network than in a highly clustered network of consumers. A marketing implication is that the choice of initial target groups and their network structures matter in influencing whether an innovation makes full or partial penetration, in markets where network effects plays a role.

Introduction

Innovations happen everywhere. However, some innovative products take off instantly, and others take a long time to penetrate the market—for example, facsimile (fax) machines needed a century to accomplish substantial diffusion. In addition, there are many new products that succeed in the early market but ultimately fail to diffuse throughout the whole customer base.³ The cases in point are Momenta Corporation and Palm Computing. Both companies sought to exploit untapped opportunities in the emerging pen-based computing market in the early 1990s. From the outset, Momenta tried to target the mass market with glitzy advertising (cf. Harvard Business School Case 9-392-013), whereas Palm initially focused on a specific target group. Palm successfully penetrated into the mainstream market, but Momenta didn't.

Why does diffusion of innovation sometimes propagate throughout the whole population, and why at other times does it stop in its interim process? Although practitioners can provide a wide array of anecdotes, systematic research on this issue has been rather sparse. The purpose of the present paper is to provide a potential answer to the question, focusing on the role of social networks, an under-explored variable in marketing.

To address this question, we develop a simple numerical model of adoption dynamics with two key components: network effects and network structure. First, we consider a market with network effects, where the benefits of adopting the innovation grow as the number of adopters increases (e.g., Katz & Shapiro, 1985). Adoption dynamics of such network products or services are quite distinct from those of conventional ones. Network products and services are quite difficult to get started and often end up being under-adopted (Rohlfs, 1978, Rohlfs, 2001). For example, consider an adopter of a communication service (e.g. email or instant messaging). The first adopter sees no benefit in adopting this service because there is no one to communicate with. Customer benefits will be realized only when other people begin to adopt an interoperable service. Because of such a lack of customer benefits at the early stage, under-adoption is natural for network products and services.

The second essential component of our model is the direct specification of the structure of consumer networks. As cited in Rogers, 1983, Katz, 1961 remarked: “[i]t is as unthinkable to study diffusion without some knowledge of the social structures in which potential adopters are located as it is to study blood circulation without adequate knowledge of the veins and arteries.” But, much of prior work on network effects (with only a few exceptions, which are Morrison et al., 2000, Abrahamson and Rosenkopf, 1997) has so far largely sidestepped the role of social networks, while focusing on the installed base or a total number of adopters. The paucity of research on social networks stems from their bewildering complexity, which defies analytical tractability (Strogatz, 2001, Lee et al., 2006). Researchers have had difficulties even describing or representing social networks mathematically.

Recently, however, advances in complexity theory have offered tools to represent social networks systematically. In particular, Watts and Strogatz (1998) developed the small-world graph model, which is a formal representation of Granovetter's (1973) conceptualization of the architecture of social networks. In his view, a social network consists of two essential elements: (1) cliquish sub-networks and (2) bridges. A cliquish sub-network consists of individuals who are interacting extensively with one another. Consider, for example, a friends and family network. Granovetter, 1973, Granovetter, 1974 found that this sort of sub-network is not very helpful when people look for jobs. The main reason is that information traversing through cliquish sub-networks is more likely to be limited to a few cliques, which tend to share redundant ties. Instead, people get more useful job information from random contacts, or people who are not in extensive relationships. Such connections are called bridges, which serve to connect diverse members from different, or often socially distant, sub-networks. In social worlds, people often create bridge-building mechanisms, such as conferences, parties, or Internet chat rooms, to facilitate interactions among random contacts or strangers. Onnela et al. (2007) empirically confirmed that the above architecture indeed represents the structure of real-world communication networks for mobile phone users.

Key questions are: should marketers focus on a few cliquish sub-networks for launching network products? Or should they exploit random bridges from the outset? Which network strategy is more likely to lead to full diffusion? To explore this question, we use Watts and Strogatz's (1998) small-world graph to represent consumer networks. The main benefit of using this model is that we can address the above questions by generating diverse kinds of networks that lie between a network of purely cliquish sub-networks and that of purely random bridges.

Our numerical analysis shows that the network structure does play a moderator role for the link between network effects and innovation diffusion. More specifically, we found that a new product is less likely to reach full diffusion in random networks than in cliquish networks. Unlike information diffusion, the spread of network effects reveals that randomness in connection topology makes it harder for an innovation to build up network effects or network benefits at the initial stage, and that insufficient network effects in turn result in a lack of momentum for the diffusion to reach the whole customer base. However, once the diffusion process reaches a critical mass, abundant bridges in a random network accelerate the process.

A marketing implication is that the network structure of the target consumer groups at the early market stage may be a critical factor affecting whether a new product makes full or partial penetration. Indeed, Rosen's (2000) argument for Palm's success is consistent with our numerical result. He argued that Palm was successful because it effectively focused on cliquish social networks in Silicon Valley, whereas Momenta's mass-market approach did not exploit such favorable properties of social networks. In addition, many shrewd marketers have been already exploiting these properties. Tupperware's party plan, Amway's network marketing, and MCI's calling circle are examples of utilizing cliquish sub-networks.

The present paper proceeds as follows. Section 2 reviews the related literature, and justifies our research method that utilizes the small-world graph. In the third section, we develop a model and a simulation procedure. Section 4 discusses simulation results, and we conclude by discussing implications and future directions of the current study.