Combining Traditional Research with Social Intelligence
Predicting Box Office Results Using Twitter
Every day, more and more people use social media to share opinions and recommendations about movies. This is particularly true on Twitter, where users are often quite passionate about expressing their thoughts in the moments before, during, and after a screening. Movie fans are so active on Twitter, in fact, that recent research has demonstrated that Tweets can predict the performance of upcoming movie releases better than any marketbased predictors to date.
Penn Schoen Berland, a market research and consulting firm whose clients include major movie studios, is keenly aware of this trend. Their business uses cutting-edge box office forecasting techniques to help clients maximize the reach and effectiveness of their marketing investments, and they were looking for a way to help movie studios make better decisions about allocating marketing dollars during the critical final week leading up to a movie’s release. They were also looking for a way to increase their own competitive edge by significantly improving the accuracy and relevance of their predictions.
The Problem
The team at Penn Schoen Berland knew that Twitter, with the rich context and detail provided in millions of Tweets, had the potential to help them achieve these goals. But their existing market research model was already very sophisticated; how could they capture Twitter conversations in a manner that significantly improved their predictions, while also being scalable and easy to use?
The Solution
In Visible Intelligence® from Visible Technologies®, they found the ideal combination of value-added data gathering, ease of use, and forecast model integration capability for these needs. Working with Visible®’s experts, the team at Penn Shoen Berland uses Visible Intelligence to track both volume and sentiment in discussion of movies slated to release each weekend. Visible Intelligence’s ability to track millions of conversations in real time ensures that actionable information on many topics is easily gathered, while its sophisticated approach to scoring sentiment provides a detailed understanding of how the audience is feeling as the release date approaches. Because Visible Intelligence presents all this information in one place, the team is always able to quickly and easily access the data they need to provide up-to-the-minute predictions for their clients. In addition, Penn Schoen Berland uses Visible Intelligence‘s Twitter data to incorporate valuable additional insights into existing forecast models. For example, by identifying audience subgroups for each film and scoring sentiment for those groups, the team is now able to track the effectiveness of a film’s marketing campaigns, showing the client how well the campaigns are (or are not) working. And because Visible Intelligence’s data represents actual individuals, clients can be assured that this picture of the information is real, and can make millions of dollars’ worth of marketing decisions with greater confidence.
The Results
By adding so much additional detail to the existing information, Visible Intelligence’s Twitter data have made a measurable impact in the accuracy of Penn Schoen Berland’s prediction model. For example: Two horror films were being released at the same time. Other market research firms were predicting that film A would beat film B at the box office, but Visible Intelligence’s Twitter data showed poor volume and negative word of mouth for film A, so Penn Schoen Berland predicted film B would win—and they were right. In addition to improving prediction accuracy, Visible Intelligence has also increased the context available for analysis, giving clients a better picture of why the audience holds certain opinions. This kind of well-rounded picture helps clients make smarter business decisions, and has helped Penn Schoen Berland build credibility in the marketplace. By enabling Penn Schoen Berland to be innovative and providing services that other research firms do not, Visible Intelligence has also helped them build a strong competitive edge in a fast-moving marketplace.
We’ve found that Visible Intelligence’s Twitter data makes our prediction
model just that much more accurate, providing context that gives us the ‘why’
behind the numbers. With Visible Intelligence data, we can tap into the mood
of the moment around each film and identify and quantify the impact of Late
Deciders that traditional box office tracking measures can miss out on.Chris Rethore Director, Entertainment Research, Penn Schoen Berland