The first notable case or example from SXSW that I would like to share with you in more detail is from the session “The Link Between Social Media and the Box Office“, presented by David Herrin, Head of Research at the United Talent Agency. Herrin and his team have developed a tool called “Preact” which allows them to monitor social media conversations surrounding an upcoming movie, up to 365 days in advance.
Preact doesn’t measure views but what UTA calls “engagement”, i.e. the volume of posts regarding the movie. The resulting hits are classified into positive and negative comments and can be used by UTA and the producing studio as a status check to see what the public opinion regarding the movie in question is like at and up to that point. Based on the volume of posts and the sentiments expressed in them, Preact assigns the movie a number between 1 and 100 to predict the movie’s box office performance (1 predicting a bad performance and 100 predicting an extremely good one).
As a result, studios and producers can see 12 to 6 months ahead of a movie’s opening weekend how the movie will perform, and adjust their communication tactics if necessary to improve performance.
Ideal Conversation Balance
In the ideal case, there is a balance between three factors: The volume of posts, a positive sentiment, and organic conversation. Organic conversation refers to posts that were created by users themselves, and not “push”/advertising messages from marketing departments. If there is an imbalance between push and organic conversations, it leads to bad results. For example, if there is too much push communication in relation to organic conversation around an upcoming movie, it means that the users don’t care about the movie enough to post about it themselves and to join online conversations about it. If there are too many organic posts, a studio’s communication department might have to step up their game a little bit to make sure the users’ interest in the movie is met with relevant information about it.
Consequently, analysing both the Preact score of a movie and the nature/content of the online conversation surrounding it allows producers to spot potential messaging and overall communication issues early on. There are a variety of drivers of online conversations about a movie: Online trailers, early & smart TV buys, film festivals showing the movie, conventions, posters, and theater displays, for example. At the same time, however, the message from the producing studio must be clear, otherwise, the online conversation dies out or turns bad.
Preact Example: 21 Jump Street
In the case of the recent adaptation of 21 Jump Street (2012) for example, evaluating the online conversation via Preact showed that the consumers didn’t understand what the movie was supposed to be about, as it was not only a new adaptation, but had also changed the tone of the movie significantly. Sony reacted to these news by repeating trailers again and again and by showing different clips and trailers that gave consumers a better understanding of the new adaptation. 21 Jump Street went on to make over $200 million at the box office.
At the moment, Preact only evaluates the conversations taking place on Facebook and Twitter, but Herrin’s team is already working on a tumblr integration. They are also planning on expanding the tool from the US American and Canadian market to other English-speaking regions soon (e.g. UK), and also to other key, non-English-speaking regions, such as China.
The case of Preact struck a particular chord with me due to its universal applicability. It is not very surprising that such a tool was developed by the movie industry; after all, Hollywood has desperately tried for years to find a formula or predictor of success for their multi-million dollar investments. Now, thanks to Preact, Hollywood seems to have taken a big step into that direction.
However, the use of Preact extends far beyond the entertainment industry. Consumer brands could also make very good use of it and predict the performance of a new product launch, while news services could use the tool to detect the stories most relevant to their audiences. Needless to say, there are also countless applications for detecting trends early on; this knowledge could be useful for companies, investors, or anyone else running a business based on the prediction of trends and user behavior.