|
|
| |
We live in a society of social networks. New information sources such as social network websites (i.e., Myspace.com, Facebook.com), Wikipedia, and group e-mails record not only a variety of layers of social connections (i.e., friendship, collaborations, coalitions) but also second-to-second individual and group activities. The newly available information, along with the computational power for data analysis, provides organizations unprecedented opportunities to understand customers, harness collective expertise, and respond to the constant change of society. My current projects apply both the empirical and system design/development methods to the research of social network information.
Understanding virtual teams through their communication network
This focuses on the communication network in virtual teams, aiming to identify the types of group dynamics that could be applied to predict the escalation of group conflicts. One key issue in managing virtual teams is to avoid conflict escalations. Moderate levels of conflict could improve the efficiency and productivity of a team. At the same time, conflict can result in negative emotions, decreased satisfaction, and enhanced intention to leave. When the conflict worsens and the intensity increases, conflict escalation could occur. A conflict escalation is characterized by disputants absorbed by the fight, proliferation of team issues, and the entire team being distracted from its task. It is apparent that conflict escalation is extremely detrimental to a team’s cohesion and productivity. Therefore, being able to monitor the status of conflicts and avoid conflict escalation plays a crucial role in the success of a team. The first article (B. Zhu and M. Gaynor, Monitoring Conflicts in Virtual Teams: A Social Network Approach) proposes a theoretical model that provides a systematic way to evaluate the evolution of team dynamics that could lead to conflict escalation. Managers thus can utilize the model to decide when to intervene before the team conflict turns negative. The model was developed based on the fact that the communication networks in effective teams are very different from those in teams with conflict escalations. For example, in an effective team, team members communicate with each other to share information and expertise. They form cliques to accomplish tasks and those cliques usually dissolve when the tasks are finished. On the other hand, when conflict escalation occurs, disputants can be too emotionally involved in a persistent clique that lasts for a long period of time. At the same time, other team members may choose to hold their communication to avoid getting involved. Therefore, several network measures such as number of cliques, the life span of cliques, network size, communication frequency of team members, and the centralization score, when put together, can be good indicators of conflict escalation. The preliminary version of this article with its hypothesis has been published in the Proceedings of 7th Workshop on e-Business (WeB'08), Paris, France. We are currently analyzing e-mail archives of four hundred virtual teams over a ten-year time period to test our hypothesis. The data is from the IETF (Internet Engineer Task Force), a community of network designers, practitioners, and researchers concerning themselves with the evolution of the Internet. Over the past ten years, IETF has organized over 400 working groups for various tasks related to the development of Internet protocols. A full version of the paper is expected by the end of this coming August, targeted at Information Systems Research or Management Science.
The second work-in-progress paper in this project focuses on the design of an information system based on the results from the first paper in order to facilitate monitoring the status of team conflicts. Knowing the network patterns that indicate the beginning of conflict escalation alone does not help because it is impossible for decision-makers to go though e-mail archives and comprehend the overall structure of a communication network. The patterns identified in the first paper can be helpful only when there is an information system designed to automatically extract and make apparent these patterns. The design of the system has been presented in the Fifth Winter Conference on Business Intelligence, Salt Lake City, Utah, 2009. We are in the process of developing the system to validate the design. A lab experiment and a field study with IETF will be conducted to evaluate the effectiveness of the design proposed. A full paper is expected by the end of this year, targeting ACM Transactions on Information Systems or Decision Support Systems.
The future work in this project includes investigating other factors that may be related to the escalation of team conflicts, such as leadership in teams and the sentimental content in communications.
The dispersion of word-of-mouth in social media
This project concerns the dispersion of word-of-mouth (WOM) of a product on social media websites. WOM marketing is an effective way to spread product information and facilitate the adoption of a product. One important issue in WOM research is to identify factors that encourage the cross-boundary dispersion of WOM. Previous studies have shown that the dispersion of WOM across various communities has more significant positive impact on product adoption than when WOM stays within a community. The homophily theory in sociology, however, indicates that people intend to communicate only with those who are similar to them. The psychological discomfort due to communicating with dissimilar people may cause information to “stuck” within clusters of people in a social network. As a result, it is usually more difficult for information to traverse across the boundary of online communities than to repeat within a community. However, only the product information that crosses the boundary can make WOM more effective.
At the same time, online network communities provide an effective channel for marketers to reach potential consumers and to study WOM. The awareness or adoption of products/services spreads from network member to member when people talk about a product to their online friends. The explicit linkage among online community members is well recorded, which in turn provides great opportunity for understanding the interactions among them. It is much easier to measure the WOM in an online community than in a real-world community. In addition, studies have shown that the WOM measured in online communities is related to the adoption of product and thus can be regarded as a good proxy of WOM in the real world. Therefore, using the archives of Myspace.com online discussion groups as the data source, this project seeks to identify the social network features that are related to the cross-group dispersion of WOM. We used the new release movies as the domain of this project. There are two reasons for our choice of research domain. The first is that the detailed information about a movie can be easily found at the website called internet movie database (http://www.imdb.com). The second reason is that the selection allows us to control the start/end time for our study based on the release date of a movie. The article (B. Zhu and C. Yang (2009), “Understanding How Product Information Traverses Across Online Communities”) describing the theoretical development of the project and the preliminary plan for data analysis has been accepted by the Third China Workshop on Information Management 2009, Guangzhou, China. We are also starting to work with colleagues in the Marketing Department at Boston University. A full version of the paper is expected early next year, targeting the Journal of Marketing Research or Information Systems Research.
We are also planning a system design paper based on the results from the empirical data analysis, aiming to provide marketers with an effective tool to monitor the dispersion of WOM about their products in online communities. Because the weekly box office of a movie is also available at imdb.com, we also plan to study the relationship between the temporal change in sales and that in WOM in online communities.
| |
|
Dr.
Bin Zhu