by Janet Fitch
CHAMPAIGN, IL – May 2005 saw the debut of YouTube.com, a web site for people to share and download videos. It attracted 9 million US visitors in February 2006. In April, 35,000 new videos were posted and 35 million were watched daily. Other popular user-contributed sharing sites include Flickr.com, for pictures, and KaZaa, for music sharing.
One important feature separates these sites from traditional online forums where users exchange messages. There is little verbal communication among users of sharing networks. It is also notable that these communities are completely built online, and users have no social ties among themselves. What communication exists is mainly from observing the activities of other users. This seems to make the success and survivability of the sites hard to predict. Indeed, many such community sites do not long survive after their launch. Even when online communities have such explosive growth as YouTube, it is not clear how they work.
Illinois Assistant Professor of Business Administration, Mu Xia and colleagues Wenjing Duan from George Washington University, and Yun Huang and Andrew B. Whinston from the University of Texas at Austin decided to tackle this issue with the aim of increasing the knowledge about such communities and to help entrepreneurs and investors evaluate such online sharing communities from a business point of view.
For their research the four researchers focused on one network for music sharing, the most popular kind of Internet sharing community, with millions of users exchanging files every week. They examined Internet Relay Chat (IRC) protocol-based sharing networks, a popular music-sharing application, analyzing data from March 2001 to May 2006 that encompasses over 300 million user activities.
Since there is little or no direct interaction among users of these sharing networks, the only possible impact a community as a whole has on individual members is through member observations of the system characteristics and the combined behavior of all other users. In these non-verbal sharing communities there are two types of users, sharers, who contribute and also download content, and free riders, users who do not contribute content, but simply download it.
Xia and his colleagues discovered that the two types of users play different roles in a community. They believe that sharers are the core element of the community, and their behavior has the dominant impact upon community growth. On the one hand, the value of a community lies in its resources, its content, provided solely by the sharers. On the other hand, sharers are also responsible for a disproportional amount of the congestion on the network. Furthermore, the more sharers who are involved in a community (both sharing and downloading), the more likely sharers are to continue to contribute content.
What really surprised the researchers was the role played by free riders. They found that the impact of free riders was different depending on whether one considers their numbers or their downloading. The more free riders, the more likely a sharer will continue to share content. This unexpected result reveals the true payoff for a typical sharer, who is happy to see more people enjoy his contribution just as long as his own downloading is not being affected. The more free riders download, however, the more likely a sharer will stop sharing. More free rider traffic means increasing congestion in downloading for the sharers, who are, themselves, heavy users of the content.
Also interesting is the role that the number of sharers plays in retaining sharers. The number of sharers contributes to both gains and losses in sharer numbers. The researchers note that the dual effect of sharers may be due to the fact that the number of sharers contributes to three conflicting effects on the community: the total resources available (the more the better), the number of downloads (a benefit) ; and the amount of congestion generated by sharer downloads (a detriment).
What about the practical business applications of this research? Entrepreneurs who are designing non-verbal sharing communities should encourage all kinds of users to join, even if some do not contribute any content-their mere existence will be a reason for potential contributors to join. At the same time, community operators should closely monitor free rider downloads, perhaps even limiting them by volume or bandwidth to ensure that sharers’ downloads are smooth and not disrupted by too much free rider traffic. To retain sharers as members, it is also helpful to provide more features that allow sharers to interact with other users. For example, recognition of some sort has a positive effect on sharer retention, even if it is as simple as adding a special symbol in front of a sharer’s user ID. However, the one thing that the community does not want to do is to make sharer downloads too easy. Based on their analysis, Xia and his colleagues have found that, all things being equal, sharers are more likely to become free riders if it is easier to download content.
For investors who need to evaluate such non-verbal content-sharing communities, they should not only look at the total number of registered users in the community, but also the different user types. While sharers should be valued highly for community sustainability and growth, free riders should also be appreciated for their part in the community, even if their participation does not include contributing content.
Download a copy of Professor Xia’s full research paper (PDF file, 311KB).
View Professor Mu Xia’s faculty profile.