This article – British Airways sees booking spike after social media posts by celebs like Rihanna and Katy Perry – caused me to reflect on the nature of interest graphs and how influence and search results are impacted by social signals.
One of the strongest drivers of Internet/Web search always has been the ability to get to the long tail of content. From a social information graph perspective, the beauty of networks like Twitter & Pinterest has been to allow diffusion of information through weak-tie connections. Both the weak-tie connections and long tail searches allow for relatively hard-to-get information to become available and flow rapidly across a large graph or network at low cost.
One of the strongest assets for Facebook has been the ability to cement strong-tie connections. Even the basis of “friendship” is based on establishment of a mutual tie. This allows for rituals and tools that help build a stronger connection – but the very basis of the strong tie may impede “fringe” information to flow into the network. In addition, the notion of celebrities and influencers plays a significant role in interest based searches and information discovery. Even LinkedIn has introduced a weak-tie feature (so that you can follow celebs – and not just connect) that allows for interest graphs to develop as so much of trends and curation begins with influencers who may NOT be in your network. This ability to form weak-ties at large scale enables the sharing of disparate information rapidly and easily, and removes the handicap that your strong-tie network needs to have the right expertise in all the different fields that interest you.
So given that Facebook is building a “Graph Search”, one needs to analyze the structure of strong-tie graphs that Facebook typically has, and understand whether they are most efficient for sharing new (and especially fringe or long tail information) quickly or with contextual richness – especially since most network will not have the hierarchy of influencers present. Will this structural issue become a fundamental flaw in “Graph Search”? Perhaps…at the minimum Facebook will have to carefully think about features (liked LinkedIn did) to help balance the social graphs out so that influencers can play at scale – and thus allow for a richer information to mine for searches. Fan Pages were a start – but that concept needs to be developed further. However Facebook does have a tremendous asset with Facebook Connect/OAuth (Read more about how Pinterest & Tumblr and other sites benefit Facebook) – and perhaps mining the information from external apps is the mechanism of opening up the graph? What do you think?
