Thursday, March 27, 2014

Exploring Curation and Discoverability from a Man + Machine perspective

At the Annenberg Innovation Lab, some of my research is helping to lead The Edison Project, a multi-year research and executive education project launched during this academic year. The goal is to better understand where the Media & Entertainment industry is heading in the next 5-10 years and, working together with the lab’s sponsors and partners, to discover, embrace and leverage new opportunities in The New Creators + Makers, The New Screens, The New Funding + Business Models, and The New Metrics + Measurement.

To launch each of these research areas, we start with a Think & Do workshop to bring a community together and brainstorm / ignite ideas (you can find more about my Think & Do process here).  In February, we had our first Think & Do for the New Funding + Business Models series asking the provocative question, "How can we create new business models to take advantage of an emerging all-mobile environment?"  One of the ideas (which was referred to at the Think & Do as Janus) had a goal of exploring ways we can take audiences deep into a studio’s library through a powerful recommendation system using both algorithms and celebrity content curation. The system operates at two levels, drawing viewers into both the back catalog and new releases. The audience builds a community around content through social media, fan pages, and geolocations.

I've been thinking for some time about curation and discoverability which stems out of the 5 years of research I've done in designing the participatory learning platform, PLAY!, a prototype that offers  users multimedia tools to build, share and enhance their ideas in collaboration with others. 

Building off of this and bringing it back to the Janus idea from the Think & Do, I think there are four vectors to consider when thinking about the data to collect when designing a user experience that offers a powerful recommendation system with the goal of combining both man and machine.

Think about it from these four vectors...
  • KNOWLEDGE (what you know) ...Here you might ask, "When I create an account and set my profile, what information does the application gather or request to better know the type of content I'm interested in?  How does the recommendation take what I'm interested in and offer discoverability, similar content that I wouldn't choose in a list?"
  • SOCIAL (who you know) ...Here you might ask, "Can I connect with my friends through this application?  Is it as easy as signing up using my FB, G+ or Twitter account?  The friends that I'm connected with via this application, are they interested in similar or different things than me?  How are these similarities and differences shown in the UI / UX?"
  • INTEREST (what you want to know, even if you don't know you want it) ...Here you might ask, "Beyond what I know, is there a way for this application to create phrase net patterns and word relationships between the topics I'm searching and participating in and associate it with improving my profile, who I socialize with and places I go throughout the day (physical graph below) to recommend new interests (whether that is new music, new friends, new advertising / brand opportunities) beyond the knowledge graph?"
  • PHYSICAL (where you are and at what time when you're looking for it) ...Here you might ask, "Are there iBeacons [such as the use of Estimote] within the physical space that offers more of a push notification (rather than a pull) of information, discounts, etc with this application in order to create a fuller experience?"
This is just some initial thoughts, and I'm currently working to refine this ...so feel free to give me feedback on it.  My goal is to identify characteristics of each vector and see how it offers added value across different types of recommendation engines / archives such as Spotify for music, Amazon for retail, Facebook for social network and Flixter for movies.