Towards Intelligent Routing For Sharing Content Online

In today's world, people share content online using an ecosystem of tools such as social network sites, blogs, forums, email listservs, chatrooms, etc. As this ecosystem continues to grow, it becomes increasingly challenging for people to make decisions about how and where to share content. I refer to this emergent problem as the content routing problem, and my goal is to test if we can build intelligent routing tools that can solve this problem.

To this end, I focused on a subset of the online ecosystem—social network sites (SNSs)—and explored the possibility of building an intelligent routing tool for this subset (problem demonstrated in the top half of the figure above). My collaborators and I conducted a scenario-based survey with ~700 people, asking them how they make decisions about sharing posts in the SNS space, and used this data to build prediction models for content routing in this space.

One intuitive approach to predict where to route/share any post would be to leverage the contents of the post (i.e. the actual text, image, video, or other media being shared). However, this has some privacy concerns because people would have to share their actual post with a third party application/system/vendor. In our study, we explored a privacy-preserving approach wherein we asked people a fixed, tractable set of questions about the post-metadata (whether it was personal or impersonal, its topic, etc.) and the affordances (perceived capabilities of SNSs) desired to share the post, and tried to predict the ideal SNSs for them to use to share it. The results of this work are currently in submission to a 2017 conference.

Our approach (presented in the bottom half of the figure above) can be generalized beyond SNSs; what is needed for that to happen is an understanding of the affordances of this online tool ecosystem and examining whether this affordances-based approach is the mental model people use in making these routing decisions. We are continuing work towards building these intelligent routing tools, starting with the SNSs subset, but also increasing the scope to all online tools in the future.