In the battle for the Interest Graph (essentially a graph of what you like) between Facebook, Google+ and Twitter, Facebook has one huge advantage. It has third party content deeply embedded in its platform. Google+ is actually better in many ways for finding and tracking media content, but it has no advanced API and therefore third parties cannot integrate into Google+ like they can on Facebook. And Twitter is too ephemeral. Media sharing is of the moment in Twitter and it’s not properly archived - like on a Timeline.
There’s an emerging class of media brands that are smart, scrappy and unmatched in their digital DNA. Call them hybrids. They’re digitally native and entrepreneurial. They use social and search to their fullest, yet many of them have a traditional ad-sales network that resembles their legacy-laced brethren. Hence the descriptor.
In a world with too much content and not enough time, hybrids are the future of media. They’re a planet of quick, agile ants that’s challenging the apes for dominance.
Media Curation is the emerging trend toward integrating and pondering media content using a mix of machine and human resources. The practice includes Aggregation (gathering) and Curation (sorting, categorizing, art directing, and presenting) such that material from multiple sources creates a unique editorial experience for readers/visitors. (Wikipedia)
Curation can be performed manually, assisted or fully automated. While in the first case a human brain looks at the content and decides about its relevance for the latter two criteria and algorithms are required by which content can be filtered and ranked.
There are various dimensions by which filtering and ranking can be driven:
Content out of a specified period will be taken into consideration. The span can reach from Last 5 Minutes for alerts to Last 12 Month for scientific publications.
The location the content was published at or the location it refers to. When the US government issues a terrorist warning Washington, DC is the publishing location, international airports might be the locations referred to.
Categories such as International Affairs, Science, Lifestyle can be used as filter criteria.
Entities can be individuals (Bill Gates), organizations (The European Union), corporations (Google) and things (iPhone).
Intention can reach from unbiased to biased, the latter including Marketing and PR material, advertisements, promotional videos, editorials, to name a view. Depending on the purpose biased content might be ruled out by default.
Based on a reader’s relationship with other individuals (Followers on Twitter, Friends on Facebook, Circles on Google+) their sharing and liking behavior can be used to curate content. Prominent examples include Flipboard, Zite and paper.li.
The correlation between contents or metadata can be a ranking criteria. For the first one the semantic profile similarity of contents needs to be measured resulting in “similar content” recommendations. For the latter one the identity of metadata elements (author, source, topic, tags etc.) needs to be analyzed.
A reader’s profile can be defined in specific terms (age, gender, profession ..) and/or derived from tracking his behavior. While the first one is noncritical the latter has drawn the attention of regulators most recently.
The sheer number of criteria already poses a complexity problem to curation. To make things even more complicated these criteria also influence each other creating a chaotic rather than a deterministic system not easy to manage. But that’s going to be another story.