Do Online Friends Matter? Consumption and Creation of User-generated Content

Knowledge-sharing online social networks, such as YouTube, Flickr, and Digg, are becoming increasingly popular. These sites allow users to create and consume virtual objects of text, photo, music and video, commonly referred as user-generated content (UGC). Similar to typical online social networks such as Facebook, these websites also allow users to make friends in the cyber world. An interesting question is whether or not, after they become friends, people influence one another’s creation and consumption of online content.First, we explore friend influence on content consumption behavior. For example, users in Flickr can view one another’s photos and mark a photo into their favorite list, an action similar to product adoption. We examine whether or not a user is more likely to adopt the same object when her friends do so. While similar question has been investigated in other contexts, such as TV viewing behavior and adoption of a new drug, the previous studies are either based on small samples, leading to lower validity, or suffer from certain drawbacks in the research design. We collect a large sample of data from typical online social networks and perform a rigorous test to address the above concerns. This research aims at understanding the success of songs, photos and videos in virtual culture market. It also resolves the skepticisms regarding, for example, the viability to use social networks for promoting products and ideas. If friends indeed influence one another’s adoption, companies may use social media marketing by targeting “influential” users.Second, we examine friend influence on content creation behavior. Most of the literature on online social network focuses on the diffusion of user-generated content and very little attention is paid to the determinants of these content objects. We examine whether or not friends influence a user’s participation in online contribution, the nature of the content that she contributes and its quality. Understanding this question is critical for the sites such as YouTube and Flickr because they largely rely on the contribution of their users. We provide suggestions to the websites about how to encourage high-quality contributions. In addition, we examine how social interaction shapes the collective outcome. For example, does closely connected social network lead to more uniform contents or more diverse contents? Understanding these questions, we believe, is vital to the healthy and sustainable growth of virtual culture market.We propose to conduct empirical studies using real data from typical online social networks. We first propose to collect data from Flickr.com, a photo sharing website. We record the users’ photo uploading and favorite marketing behavior, and the friend links between them. Compared to most studies where the collected network is static, we track the network evolution over time and therefore observe people’s behavior before and after, which facilitates us to identify the friend influence. To examine friend influence on content consumption, we compare photo adoption behavior of user pairs before and after they become friends. We control for other effects due to common shocks and simultaneity. We also conduct robustness checks when varying, for example, the importance and activeness of friends. To examine friend influence on content creation, we first develop methods to quantify the nature and quality of photos. Then, we compare the uploaded photos of user pairs before and after they become friends. We evaluate the collective outcomes such as the variety of content in a network as a function of the network structure. Finally, we perform the analysis across different cultures, e.g., among Asia users and North America users, and examine if the results change.