Facial-Recognition Technologies Reach Madison Avenue
Gesture- and facial-recognition
technologies, so far developed for security purposes and more recently for
video gaming, are finding traction in the advertising world.
According to an
article in The Wall Street Journal, the technologies
may enable marketers to alter images and messages on a billboard, for instance,
based on whether the viewer is paying attention. It can also tweak that message
based on whether the person is male or female. Other potential applications
include using sensors to customize TV spots based on the age and gender, or
enabling interactive TV commercials based on the viewer.
Many consumers are
expected to first experience facial-recognition technologies through Kinect,
Microsoft’s motion-sensor gaming device that lets people play videogames without
holding a traditional controller, according to the article. It includes a camera,
a microphone and an infrared depth sensor that analyzes a player’s body movements.
Basic
applications of the technology have also been tested in stores. At 77kids,
the children’s chain of American Eagle Outfitter, children can stand in front
of a digital screen to try on a virtual outfit.
According to Journal, “The
ultimate application from the industry’s point of view: mirrors in a department
store or a device connected to a TV in a consumer’s home that could calculate
a woman’s dress measurements or a man’s trouser size and instantly place an
order.”
Separately, Forbes.com last week debuted a web-based facial
expression reader app. Participants turning on their web cams can view video
ads from Doritos, Volkswagen and Google. Employing technology developed by
Affectiva and the MIT Media Lab, the ads change based on human facial expressions.
A confused look, for example, may trigger more information about a product,
while a look of joy may send viewers straight to the purchase page. Dave Berman,
CEO of Affectiva, told Media Post that his marketing clients want to
know more about how online consumers are responding to ads.
“They’re very interested to know how their customers feel,” Mr.
Berman said of his many clients, “Click-throughs don’t tell you that.”
But
the Journal article noted that privacy concerns are rampant. The
technology can read a person’s face as soon as they enter the store, then link
that information to a profile of that individual.
“If you want to do something evil with it, I’m sure you could. It is
the same thing with anything else technology-wise,” said Benjamin Palmer,
CEO of Barbarian Group, an interactive-ad agency working on such interactive
billboards.
Discussion Questions
How receptive will consumers be to gesture or facial-recognition devices in marketing?
Different cultures of the world have varying responses to this kind of technology. Here in the US, we are hypersensitive to privacy concerns. However, if a service that is perceived high-value is offered to a shopper, the guard often goes down. Even basic loyalty programs that only require email addresses or telephone numbers breach basic privacy barriers.
The challenge is to create awareness of the benefits of the technologies and the services they provide. Articulate promotion of the offers can result in the overcoming of privacy threats.
There are some great technologies available and the implementation of them can generate a compelling shopping experience.
Facial, racial and gender recognition are just further steps towards the promised land of marketing relevance. Why show a 40 year old man a L’Oreal ad if, in an instant, the ad can be switched over to Old Spice? Then if you can go a step further and measure the viewers actual response to the message, you’re moving into an unprecedented level of consumer understanding.
Facial recognition in advertising will become a dominant force in the future, but that future is likely to be a long way off. Advertisements will eventually be able to target consumers individually through this technology, but a lot of factors will need to come together for this to work properly, such as knowledge of the individual consumer’s buying habits, the data infrastructure to store and rapidly retrieve that information, and the accuracy of the facial recognition software. That last one is the largest initial hurdle to this technology; most existing facial recognition technology checks a face against a relatively small set of faces, whereas advertisers will need to be able to identify a single unique face from the tens of thousands of faces stored in their records.
Early adopters of facial recognition software in sales will likely be plagued with complaints about inaccuracies of the software and false positives. Genders will be mistaken, people will get mixed up, and sooner or later some facial recognition ad will try to sell ham to someone keeping kosher or announce a special on an embarrassing personal hygiene product. Look for this to do more harm than good, at least initially.
However, once the systems become more robust, I expect to see widespread adoption of this advertising system. Just don’t look for it too early.
I didn’t read the (original WSJ) article, but I’m having a hard time picturing how one would “customize” a billboard: aren’t those, by definition, seen by multitudes of people?
Anyway, this all seems reminiscent of those Popular Science articles from the ’50s where everyone would be flying mini-helicopters in 10 years…the future is usually just a fraction of what it could be (and that’s often a good thing, too).
I’m a fan of these electronic personal recognition systems, I really am, but I wonder how effective they are in non-interpersonal practice. When you’re reading ads on your computer, do you exhibit facial expressions or are you poker-faced like me? In a crowd or on the street without a companion, are you one of those straight-ahead, no-nonsense, minimum-eye-contact folks like most are? How can a machine read those expressions as anything but lack of interest? But down in your expressive little heart, you might be interested, or repulsed, or amused. It usually just doesn’t show on your face–unless you’re with someone.
And racial/gender recognition? What about instances in which several people are looking at an ad–billboards, air terminal signs, etc.? For which person does the racial/gender recognition system adjust the ad? As for the other applications, electronically displaying virtual outfits on shoppers is really cool. And physically active games, too.
Sometime ago I was advised that I might be happier living on a desert island than the real world. I do not disagree. The only advertising related story that would bring “a look of joy” to my face is that I would never have to be subjected to another one. I cannot remember ever purchasing anything based on seeing an ad and cannot imagine ever wanting to have an ad tailor-made for me based on facial recognition. I think you can imagine, from that, what gesture I would be making.
That’s a great idea, as long as apps give a benefit to the customer and not just for the purpose of gathering data for customer research.
A discussion about facial or gesture recognition is really about agent vs. principal. Do we seek out proxies for purchase or purchase intent? Or do we instead look for the actual behaviors?
Audience recognition is powerful technology in search of the right application. Digital screens that change content using count or gender admittedly have some potential to advertisers. Ads that adjust to a web visitor’s emotions are a red herring. People do not noticeably emit elation or agitation while web surfing.
A better evolutionary path than recognizing faces is capturing response the closer to the purchase the better. Response may include SMS opt-ins, QR code scans, website visits, GPS-defined shopper traffic, and, in a retail environment, actual purchases.
Current generation analytics and technologies don’t require us to rely on substitutes to evaluate advertising effectiveness. Spend engineering cycles on truth, not proxies.