SocialPortrait is licensable software, based on InferLink Corporation’s research on analyzing how people’s public social media posts reflect “who they are.” For instance, if someone posts numerous times about setting up different computers, and how his daughter uses the computers and asks for help, that implies that he is a Techie and a Dad. It is currently licensed as a marketing tool for the entertainment industry, and has been licensed in the past for persona marketing for brands and as an advanced media planning tool.
SocialPortrait can perform both group level and individual level analysis. For individuals, SocialPortrait produces a “dossier,” which describes a person’s interests and demographics, inferred from what he or she says about him or herself.
For example, based on the public Twitter account below, the system creates the following dossier:
At the high level, this account shows an interest in the news, education, books and science. Digging into deeper, the more detailed topics reflect interests in American culture, politics and the United States. Demographically, the account also reflects someone who is a mom, owns a pet, and lives in a house.
Given this level of information in a dossier then allows us to aggregate sets of dossiers to understand a group or audience, in aggregate. For instance, we can look at a group of Twitter followers, users who use a hashtag, an email list, those who share links, etc. We call these groupings “audiences.”
The example below presents an audience overview for a set of users who Tweeted out links to foxnews.com. As the figure shows, this audience is composed of three major personas. One group are parents with young children who share an interest in the news, sports and the world in general. The second group are parents with interests in science and religion. The third group are grandparents that are dads (hence grandfathers). The bubble next to each shows how much more likely you are to find that persona in this audience, versus the world at large (e.g., it is highly targetable) For instance, in this audience you are 9.6 times a likely to find a member of the first persona than any random audience in general!
Finally, we can leverage these audience summaries by comparing them to other audiences. For instance, the system can find look-alike audiences on TV or Websites (to improve targeting and acquisition strategies), see how followers of two brands compare (e.g., for competitive intelligence), compare brands to media properties (for advertising opportunities), or just allow one audience to be compared to another (say for social science research).
In the figure below, we compared people who have Tweeted links from foxnews.com against people who have Tweeted links from a number of different websites (across news, entertainment, sports, retailers, brands, etc.). Unsurprisingly, the most similar audiences are those who shared links from other news sites. As the figure shows, the sites are ranked by how similar the demographics and interests of those audiences are (and again, although we compared to websites, we could easily compare audiences for TV shows, networks, movies, video games, etc.)
SocialPortrait even allows us to analyze these audience comparisons at a deeper level. The figure below shows the overlap between the BBC News and Fox News audiences, showing shared interests, but also a few divergences (such as the BBC audience also cares a bit more about Food and Technology):
SocialPortrait has been licensed for the entertainment industry, brand analysis and as a media planning tool. If you are interested in licensing please contact us.