WE CELEBRATE AND ENCOURAGE INNOVATION.
Innovators push the boundaries of the known world. They're change agents who are relentless in making things happen and bringing ideas to execution.
Facial recognition tech puts marketing to the test
Rana el Kaliouby can read your face like a book--and her insights could rewrite the rules of consumer engagement.
El Kaliouby and her MIT Media Lab colleague Rosalind Picard are the minds behind Waltham, Mass.-based startup Affectiva, whose flagship Affdex Facial Coding integrates automated facial-expression-recognition technology to measure and interpret viewers' emotional responses to brands, advertising campaigns and other digital video content. Forget focus groups, telemarketer surveys and other traditional audience research tools hamstrung by self-reporting variables: Cloud-based Affdex probes even the most subtle nonverbal cues to reveal what consumers really think about a program or product.
"You can understand so much about how consumers perceive a brand by analyzing their spontaneous, subconscious responses," says el Kaliouby, Affectiva's chief science officer. "If you're a content creator looking to elicit a certain emotion, we can validate that. In cases where an ad is trying to elicit humor, we can tell you if people get the jokes or not by the number of people who smile, the intensity of the smile and the timing of the smile. There are so many cases where self-reported responses get it wrong. Facial response is more accurate."
El Kaliouby has dedicated her career to leveraging technology to improve interpersonal communication and expression. Affectiva spun out of her research into how facial-coding technologies can ease autism spectrum disorders. "Many people with autism struggle with reading nonverbal cues and acting on them," she explains. "When you lose that ability to understand and process nonverbal cues, you're at a huge disadvantage socially."
She and Picard--director of MIT Media Lab's Affective Computing Research Group--launched Affectiva as an independent company in 2009, encouraged by interest from MIT sponsors including Procter & Gamble, Google and Microsoft. "We realized there are a lot of ways we can use [Affdex]--in cars, in mobile phones and in market research," el Kaliouby says. "I remember thinking, I don't know if I want to be in business! But what sold me was the vision of our technology becoming ubiquitous. It works in all sorts of contexts."
Affectiva has since raised more than $21 million from investors including the National Science Foundation, WPP and Kleiner Perkins, and inked partnerships with brand agencies Millward Brown and InsightExpress, both of which offer Affdex to their clients. The software integrates with any standard webcam across connected laptops, tablets and smartphones, capturing consumers' expressions as they view video content. All footage is streamed for processing to the Affdex cloud server, where computer-vision and machine-learning algorithms or techniques evaluate emotional states like surprise, disgust and concentration. From there, analytics tools slice and dice the aggregated data to generate overall emotion scores, complete with near-real-time, scene-by-scene facial-data playback.
Most Affdex consumer insights are derived from paid panel participants who are given the option to share their responses or disable the webcam. About 50 percent leave the webcam running, el Kaliouby says. "The incentive is that you're getting value back," she explains. "We're optimizing your content. Maybe we're making your favorite game more engaging by tracking your emotions over time. That's a powerful value proposition."
Affectiva is taking steps to get its facial recognition tech into even more hands. This fall the firm introduced a software development kit for Apple's iOS mobile operating system, enabling third-party developers to bake Affdex features into their iPhone and iPad apps.
"People check their phones an average of 150 times a day. To us, that's an exciting opportunity to capture your emotions at each of those moments," el Kaliouby says. "Emotion-enabling apps also opens up a ton of ideas around gaming, usability and messaging. Education is another big, big place to make a difference. Our hope is the SDK will allow us to bridge the gap into a number of new markets. Emotions are core to everything we do." --Jason Ankeny