Online community success - A study of success criteria and user behaviour in online communities
MetadataShow full item record
This item's downloads: 870 (view details)
In a world where online users form communities for various purposes and around many different topics of interest, it has become of social and economic importance for owners, providers and managers of online communities to assess and possibly improve community success. However, the first problem arises when trying to measure success. Success means different things for different people, and moreover, the different purposes of communities determine the criteria under which success must be defined. For example, for assessing the success of a questions and answers (Q&A) community like Stack Overflow, one has to take into account how effectively questions are being solved. On the other hand, social connections among the participants are far less important in a Q&A site than for a social media site like Facebook. The first core contribution of this thesis is to define concrete success criteria for a number of communities. In particular, we focus on Q&A, Life & Health, and Knowledge Creation communities, which are distinct types of communities, each with their own purposes and goals. In order to achieve community success, it is vital to understand which user behaviour is indicative of, or contributes to, success. In the literature, many community features have been proposed as indicators for success, including aspects of user activity, participation, loyalty, and interconnectedness among the community members. However, these indicators have not been evaluated against tangible success criteria. Our second core contribution is to put together a collection of proposed community features, and to evaluate them against the success criteria we defined for the three types of communities in our study. We find that there are no universal success indicators, as the goals and purposes of the communities vary widely, and that it is therefore important to identify the appropriate user behaviour that maximizes success for different types of communities. Lastly, the third core contribution of this thesis is to identify successful user behaviour by studying the relation between user behaviour and community success, utilising prediction and simulation approaches. The prediction allows us to determine the best combinations of user behaviour for each community type, and the simulation enables us to represent user behaviour in a computational model in order to study community success in a controllable environment. This way, we are in control of the various conditions that a community can be subjected to, and observe how these change the community’s ability to fulfil its purpose beyond what is captured in recorded data, and without interfering with live communities. The practical application of our approach allows owners and managers of online communities to gauge the success of their communities, and identify the weak spots and potential remedies to increase success.