Influence12 started with a keynote address by Gilad Lotan, VP Research and Development at Social Flow. As a PhD student studying social media data, Gilad’s talk was of particular interest to me.
Gilad started his talk by defining influence as “the ability to disproportionately affect the spread of information.” He then provided few real life examples from online social spaces that gained attention from the audience (such as Long cat, #Kony2012, etc.) and how the analysis of such data can define emerging patterns and help us understand how information spreads online. This information can allow us to make better recommendations and determine an appropriate time to disseminate information to audiences, which Social Flow provides to their clients. Through a software tool, Social Flow helps companies/organizations with what to post and when to post on Facebook and Twitter.
He stated that people resort to simplified metrics (such as number of followers, likes, tweets per minute etc.), yet there are many other metrics at play. He also eluded that we spend an enormous time trying to reach certain users deemed “Influencers” and that the networked dynamics are much more complicated. The key is to understand the system as a whole, understand the network and the norms of the network (attributes of the network, what events influence the network, etc.). This is obviously possible once we have the data and can see what deviates from the norm. And knowing who is influential on a network is nice but what is interesting is the path taken to get to these influencers. Another aspect of understanding the data is to look at the audience to determine how interconnected the audience is along with regional trends (geospan, persistence, volatility – to measure the effect on the network on a specific event).
He concluded his talk by giving us food for thought – that we need to look beyond just the “influencers” and look at deviation from the norm, audience receptivity, network attributes (shapes, centrality, clustering), bridges and context setting.