Facebook is now showing an Aggregated Mentions news feed story when multiple Pages or friends you are connected to include the name of the same Page in an update. The new story type shows you people or entities who are trending in your network, such as a public figure who has done something newsworthy and is being discussed by multiple friends or Pages. What’s especially new and important is that Facebook recognizes names written simple text, not only those @ tagged, implying that Facebook is machine-reading the updates.
Facebook used to have a publicly accessible feature called Lexicon which anonymously visualized conversation trends on walls using machine reading, and Aggregated Mention stories could be a new application of this technology.
Aggregated news feed stories are also shown for Places check-ins, birthday wall posts, and shared links. They show how the same content can be shared or the same action can be taken by different friends and Pages, which gives users perspective. In the case of Aggregated Mentions, they show how different friends and Pages feel about a something represented by a Page, such as a controversial political figure. The story type doesn’t appear to be rolled out to all users, though, as we couldn’t recreate it.
In the news feed, users see a story stating “[Friend 1 / Page 1] and [Friend 2 / Page 2] mentioned [Page 3]” with each name linked to it’s corresponding Page or profile. Users then see the original update by each friend or Page with the threaded feedback folded up. Users can click the feedback icons to see who has Liked or commented on that story. It can be interesting to see how a different context or opinion will net a similar mention different amounts of Likes.
Facebook’s ability to extract textual content out of updates holds a lot of potential. Users might one day see stories aggregated because they express the same sentiment, such as one update that includes the word “happy” and another with the word “happiness”. The text of updates could be also be used to recommend Pages, such suggesting the Nikon or Canon Page to a user who frequently mentions “cameras” or “photography”. If done completely algorithmically, users shouldn’t worry about “Facebook reading your private updates”, and instead allow this machine reading to help improve their experience.
[Images via Forbes.]