Time Warner Inc.’s new study of second-screen viewing behaviorputs even greater emphasis on the growing urgency and sophistication of finding out when and how viewers respond to programming. The TW study, being conducted by the Ball State University’s media research group using Time Warner’s year old Media Lab in Manhattan will focus on the interaction between big-screen and handset (or laptop) supplements to TV shows .
Coming on the heels of extensive commentary about Netflix’s use of “big data” research in developing its House of Cards series – including viewing pattern analyses of other shows that affected production design for the original series –the new study adds to the expectation of creating content based on user preferences.
Content research is hardly new; Hollywood’s classic “sneak previews” often resulted in tweaking feature films before theatrical release.
But today’s tools fine-tune the options and augur even more customized program development. Among the new tools is “emotion tracking,” which is being used by several broadcast and cable networks. “Affdex,” the system developed by Affectiva Inc., officially debuted last month just before the Super Bowl, after being quietly tested for nearly two years by Coca Cola, Unilever and global marketin research agency Millward Brown. Affectiva declines to identify the cable networks that have been part of its low-key two-year ramp-up, but we believe that sports and drama networks have been involved.
Affdex uses facial coding software to identify how you react to a TV show or commercial by calculating your facial responses as you watch the video. The technology, for now being used in market research labs, uses a small web-cam aimed at your face to gauge emotions based on the 10,000 different facial expressions in the human repertoire.
“We can identify [reactions] to any kind of media, including the segments in a video where emotional engagement peaks,” explained Affectiva Senior Marketing Manager Elina Kanan. Using related research and demographic information, Affectiva offers content producers (including advertisers) a way to gauge, for example, whether customers would really want to see a show, buy an advertised product or pursue another media-induced activity.
The emotion-measurement system is “unobtrusive,” Kanan added, heading off potential privacy concerns. Aside from the opt-in factor, the facial responses can be anonymous. The purpose is to identify and aggregate how viewers respond to specific visual scenes, she said.
Unlike research methodologies that use technologies such as galvanic skin response or eye-tracking (a core tool in the Time Warner Media Lab), Affdex measures emotional states such as surprise, dislike, confusion and attention based on facial expressions. At its core are advanced computer-vision and machine-learning techniques that recognize and automate the analysis of tacit expressions and then interpret viewers' emotional responses.
Collectively, the newly unveiled Time Warner tests (mostly likely for HBO, which has been a frequent user of its parent company’s Media Lab in the past year), the Netflix deep dive into big data, Affectiva’s “emotion” research and other new predictive viewership tools will play an ever-larger role in what gets onto the screen.
Gary Arlen is president of Arlen Communications LLC in Bethesda, Md., and a long-time interactive TV enthusiast. Reach him at GArlen@ArlenCom.com.