Are We Humans or Are We Dancers?

Several major retailers and CPG houses are using software developed by MA-based Affinova to evaluate packaging and other aspects of products usually handled by brand teams and focus groups. Affinnova's premise is that consumer markets evolve like the natural world: the strong products survive while weaker ones die out. Using a "genetic algorithm" called IDDEA, they recreate this evolution in a series of generations covering possible products or packaging until the strongest design emerges. While not really machine-designed (human managers oversee the testing and choose options from every consecutive generation), the software speeds up and streamlines the process over what is able to process and analyze these preferences in a way no focus group can accomplish Because Affinnova's surveys include consumer profiles that contain basic demographic information, customer beliefs, consumer habits, with such data splicing especially useful in brand redesign.

Affinova is behind the launch of Dannon’s Activia, a probiotic yogurt that generated more than $1bn in sales in Europe. Parent company Groupe Danone of France had two prior unsuccessful attempts at importing a successful European products into the US market. Affinnova's research identified imagery and messaging that appealed to Americans, including expressions such as "body in balance" or "running smoothly" over descriptions of bacteria or cultures. The icon chosen by the software was a sexy photograph of a woman’s abdomen instead of an illustration. Activia sold more than $130MM in 2006 (its 1st year on the market), and $300MM the following year.

In another potentially game-changing use of technology, fragrance-dispensing kiosks in France are relying on face-recognition technology to decide whether the spritzes they deliver should be for men or women. Facial recognition specialist Quividi, which developed the kiosks in partnership with fragrance diffusion device maker Presensia, developed software that first determines the nearby life form is human, then estimates the corresponding gender of the prospective customer, choosing an "appropriate fragrance" to be puffed into the air. The kiosks then "can diffuse up to four different scents up to a 5-meter distance." Similar technology, recently displayed in Japan delivers ads based on the ages and genders of people passing by. NEC’s advertising kiosks “see” people as they approach, check out their facial characteristics and deliver ads based on what they believe to be their relative age and gender. For retailers, the goal is targeting advertising to those consumers most likely to be engaged. But the effort is running headlong into some sticky ethical questions involving personal privacy.

What if face-recognition-powered digital signs and kiosks can identify shoppers by name merely by scanning their faces and matching those images against known images? Such techniques are already being used in airports, with federal agencies comparing file images of suspected terrorists against those that video cameras pick up at the airport. Many applications today—such as Xobni—already scour the Web trying to match images of people with those who are apparently sending E-mail to the application’s user. But with the E-mail effort, the trick is looking for an E-mail address match and then trying to find a photo. Starting with a photo is a much more challenging task. Of course, there’s the question of how reliable such a system would be, especially in the early stages. Face recognition age-verification systems have had more than their share of embarrassingly inaccurate moments, including Japanese teens fooling a cigarette dispensing machine with photos ripped from magazines.

For information about licensing, contact Broad Street Licensing Group (tel. 973-655-0598)