SIeCommunities: Discussion leader/summarizer: ErikWJ

Katz, M.L., & Shapiro, C. (1994) Systems competition and network effects. Journal of Economic Perspectives, 8(2), pp. 93-115.

Discussion leader/summarizer: ErikWJ

Research Questions:

This paper is not trying to address one specific question. Rather it frames the issues that arise when network effects are considered. Katz organized the questions and the paper into the following three categories with discussion of the forth question left for the discussion.

(from page 4)

Technology adoption decisions: how many consumers purchase a given system, and what institutions or market mechanisms arise to internalize the network externalities associated with adoption?

Product selection decisions: what forces determine consumers’ choice among rival incompatible systems?

Compatibility decisions: Which firms will seek compatibility, and which will not? What institutions arise to set product standards and achieve compatibility?

What are the policy implications of the answers to these questions and when should government intervene?

Key Terms/Points/Claims/Methods

The unit of discussion is a system: “collections of two or more components together with an interface that allows the components to work together.” P.1.

Positive Network externality: “The value of membership [to a network] for one user is positively affected when another user joins and enlarges the network.” P.3.

This paper is already a collection of key points from other works, so instead of summarizing each point, I will outline the themes that connect the claims.

Much is known about how individual units (people, companies, projects) behave in isolation. However, as these units start to interact and build networks, the behavior sometimes can take on unique and unexpected effects.

External Adoption Effects: Consumers normally face a complex challenge when making decisions for purchasing goods that have network effects. A fax machine in isolation has little value, but as more and more people buys fax machines, the value of belonging to the network of fax machine owners increases. The purchase decision for the fax machine now contains a person’s value of being able to fax in addition to the perceived strength of the underlying network of people that own fax machines now as well as the value of that network in the future. Will the network increase in size and value (e-mail users), or will it diminish and decrease in value (VCR users). The paper also explores many other issues that arise from network effects including: How to build up a network, the cost of changing networks, and how reputation or large investments can single to consumers the probably strength of the network in the future.

Note: The job talk candidate from Friday the 18th gave a good example of how a network can lead to adoption of a new technology. She demonstrated that a computer software package for watching television and talking online had cascading implementation effects because of the network of people involved. There was a first wave of implementations in the U.K. to watch the World Cup of soccer. Then there was a second wave of instillations in America of the software by people who had contacts with people in the UK. This was explained that as each additional person in ones personal network started using the software, the software became more valuable to implement. Thus, the more contacts one had in the UK, the more beneficial it was to install the software. There was also a tertiary implementation of people in the US who did not have contacts to the UK, but instead had contacts to the people who had contacts to the UK.

Compatibility: Compatibility between two different systems can be achieved with two main mechanisms: standardization and adaptation. Standardization is the mutually agreed upon method of how two or more items work together. Adaptation is a tool that allows one system to operate with another system that would otherwise be incompatible. Companies that are in a position of power normally prefer incompatibility as a way to take advantage of network effects that can promote dominance. See Microsoft. Companies might also want to increase compatibility to take advantage of the networks of others. There are nice examples of how Nintendo build up a network of software developers for its hardware.

The conclusion of the paper is a discussion of when government should step in with policy or management. Anti-trust laws, regulation of industries (Energy, FCC, transportation) all have cost and benefits when operated by the government vs. privatization. Each of these issues can be expanded upon significantly, and some people have spent careers talking about the role of government in marketplaces.

Critique

This paper is a nice introduction to the scope of network issues. The paper sometimes has trouble maintaining its focus while navigating the large number of concepts. Relatively little presented in this paper is an independent contribution, but as mentioned previously, it is a nice framework for a large number of issues that surround networks. However, this paper should not be viewed as a comprehensive exploration. Network effects are so prevalent they can be seen in many disciplines: Information, economics, public policy, complex systems, cognitive science, sociology, communication, and on and on.

I think one weakness of the paper is that there is not enough focused on negative network externalities. There is the discussion that networks effects might slow innovation and that networks might lead to monopolistic behavior. However, these are more effects of the network as a whole and not examples of negative externalities. One example of a negative network externality is the noise of too much participation. In the Slasdot paper that we read last week, it identified that in large online conversation spaces, the ability for central moderation is severely limited as the amount of contributions increases. As each additional person makes a comment, it is more difficult to identify how to allocate one’s limited attention. The distributed moderation system used on Slashdot is an example of a mechanism that has been designed specifically to address the consequence of a negative network externality. Lampe2004, Cliff and Paul Resnick. Slash(dot) and Burn: Distributed Moderation in a Large Online Conversation Space. In Proceedings of ACM CHI 2004 Conference on Human Factors in Computing Systems, Vienna Austria. 2004.

DerekHansen: I agree with Erik that this paper (and many others) only focus on positive network externalities and forget about the congestion costs that often occur in situations like the one we are interested in. This is why I appreciate the Butler2000 article, which looks at both (although doesn't use the economic lingo).

I think that the reason that we reading this paper for e-communities is that membership in eCommunities face many, if not all, of the challenges of network effects. For instance, to sustain a healthy eCommunity there needs to be a critical mass of people. There are sections in the paper that discuss network building. With little imagination and a brief squinting of the eyes, the sections could be seen as prescriptions for community building.

DerekHansen: The additiona reading by Liebowitz and Margolis is a nice critique of this work. In situations where there are significant positive network effects the most stable equilibrium is where only one of the competing goods is purchased. The connonical examples put forth are the QWERTY keyboard and VHS vs. Beta. Liebowitz and Margolis go to some lengths to convince their readers that QWERTY and Beta were not necessarily inferior to their competing goods and that in general there are very few cases of the market selecting the "wrong" product, even when there are positive network externalities.

Connections with other readings, ideas, etc.

Increasing returns and the lock-in effect (p.7) were initially credited to Brian Arthur. In economics the concept of diminishing returns was that one gets less value with the more items that a person has. For example, the value of the first apple or first car is greater than the value of the 100th. Arthur thought there might be the ability for increasing returns meaning that the value of the next item is greater then the original. This concept faced significant resistance for a long period of time but is now generally accepted. The process of trying to prove increasing returns is warmly described in the book Complexity.

DerekHansen: One of the additional readings (Katz1985) is Katz and Shapiro's original work on Network Externalities. It is quite dense mathematically, but lays out the basic model that has been used over the past few decades. Arthur's work is certainly related, although his approach is a bit different.

Waldrop, Mitch(1993) Complexity: The Emerging Science at the Edge of Order and Chaos.

Additional papers and books that focus on network effects and/or design issues:

Rogers, E. M. (2003). Diffusion of Innovations. New York, London, Toronto, Sydney, Singapore, Free Press.

Norman, Donald A. (1990) The Design of Everyday Things. New York: Norman, 1990. New York: Doubleday.

Discussion