If Rana el Kaliouby has her way, you will never have to channel surf again. Affectiva, the company that el Kaliouby founded with MIT colleague Rosalind Picard, develops technology that reads minute facial expressions to measure emotion. A television equipped with a webcam and Affectiva’s technology could determine which shows you like to watch, given your past emotional reactions to them, and program your television accordingly.
Affectiva’s facial-reading software, Affdex, is already being used by major advertisers, including Unilever and Coca-Cola, that previously depended on focus groups and surveys to test ads. The problem with the old approach? It requires people to self-report their reactions. Affdex is more scientific. It records viewers as they watch ads on their computers and uses an algorithm to analyze subtle facial cues, drawing from a database of more than 283 million facial frames. Then it adds viewers’ moment-to-moment reactions to a timeline for the ad, so companies can see precisely which segments might need tweaking. Affdex also determines if each viewer’s overall reaction was positive or negative.
Since its launch in 2011, Affdex has helped measure audience sentiments for several high-profile events, including the 2012 U.S. presidential debates and this year’s Super Bowl. The technology has spread globally, too: Affectiva’s algorithm now includes data from viewers in 35 countries. Its ability to detect cultural differences in audience reactions has proved especially valuable to advertisers, says el Kaliouby, Affectiva’s chief technology officer. For instance, when one company tested an ad in Brazil, it found an enthusiastic reception in one region of the country and a lackluster response in another. “If they find these things out earlier, advertisers can save millions of dollars,” el Kaliouby says.
“Affectiva is doing something radically different with data. The future of marketing is about using quantitative information.” –Bob Lord
Originally, Affectiva’s technology was developed for a far different purpose. El Kaliouby and Picard, an MIT professor and Affectiva’s chief scientist, were developing a device that could respond to users’ emotions, with the aim of helping autistic children communicate better. They realized that companies and researchers could benefit from the technology. In 2009, MIT’s Media Lab spun off the project into a separate company. That year, David Berman, formerly the president of WebEx, joined as CEO.
Since then, Affectiva has grown in large part by signing partnerships with market research companies, including Millward Brown and InsightExpress, both of which offer Affdex to customers. It has raised $21 million from investors, including Kleiner Perkins and WPP.
Eventually, Berman says, the technology could be used to test websites for ease of use or become a more scientific version of Facebook’s Like button. In the next two years, Berman envisions Affdex becoming a complement to “smart” televisions that can understand people’s preferences. “If my wife and I both like to watch the same show, it will fine-tune the algorithm,” he says. “It puts the emotion back into viewing.”