Field Studies on User Experience of Automation in Context-Aware Social Media

Abstract

Today’s social media services increasingly contain highly automated context-aware features. However, there is scant research on understanding how automation affects user experience and what are the related issues, which should be taken into account when designing social media services. This thesis starts to fill this gap through empirical user studies and by forming design implications for automation in social media services from a user experience point of view.

Many popular social media services have introduced automatic context-aware means for social interaction. For example, users are able to automatically assign location information to their status updates on Facebook and Twitter, and information about a movie one is watching on the online movie service Netflix, can be automatically and instantly sent to the user’s Facebook profile for friends to see. Human factors research has shown that automation has benefits and costs. On the one hand, it can free humans from performing nuisance tasks; on the other hand, it changes the role of the human user in the task performance loop by making the human a passive supervisor, which can decrease for example manual skills and situation awareness.

This thesis belongs to the field of human-computer interaction. The research takes a cross-case approach on the results of several field studies in which users used context-aware social media technologies in real-life contexts. The studies cover topics such as automatic location sharing, mobile notifications, and automatic video remixing, and majority of the studies were in mobile context. Altogether, 132 users participated in the studies. Individual interviews, data logging, and questionnaires were used as the primary data gathering techniques. In the cross-case analysis, the results were categorized under three themes: automation in input of information, automation in output of information, and automation in collaborative media generation. The benefits and costs of automation in each category were discussed. In addition, broader design implications are proposed to guide automation design of social media services into a direction where automation can support user experience.

The results convey that automation is capable of supporting users in pragmatic tasks. In addition, if the level of automation in the execution of the pragmatic tasks is also in agreement with user’s hedonic (emotional) goals, automation can support a positive user experience. However, at the same time automation assists the user, it also lowers user control. This thesis shows that it can be challenging for the user to, for example, correctly interpret automated social information, sustain satisfactory privacy, and control the impression other people form of them. To conclude, this thesis proposes five implications for design that emphasize the application of alternative levels of automation to achieve a pleasant user experience with social media services.