Whatever happened to the facial recognition trend in digital out-of-home?
Advertising gets its share of bad PR, but some incidents stand out as particularly extra-bad. Back in March 2008, an article in The New York Times drew readers’ attention to new biometric technology that enabled cameras in digital out-of-home displays to scan dozens of facial features to determine demographic information about passers-by including their age, race and gender, which in turn allowed computers to target advertising to them based on these factors. For example, a digital out-of-home display employing the technology might “show one advertisement to a middle-aged white woman … and a different one to a teenage Asian boy,” according to the article, provocatively (and not inaccurately) titled “Billboards That Look Back.”
Public reaction to the article was entirely predictable, which is to say almost entirely bad. The issue which immediately leapt to mind for most people was privacy: Should advertisers be able to collect such personal information without your permission or even knowledge? Yes, pundits conceded, individuals moving in public spaces are probably fair game, but still — it’s just creepy. And that kind of thing matters (or at least, should matter) to advertisers concerned about brand perceptions. No brand in its right mind wants to be associated with unwanted, intrusive electronic surveillance. What was the right way to handle this?
These issues seemed especially urgent for two reasons. First of all, the applications described above are actually just the tip of the iceberg when it comes to the full power of facial recognition software, which can actually go beyond collecting demographic data to identify specific individuals by their unique facial patterns. Indeed, the technology was originally developed by the Israeli military for counterterrorism and is capable of matching a face in a crowd against a database of known individuals at the rate of at least 100,000 faces per second. (It’s probably even faster, in its latest intelligence applications.) It didn’t take much imagination to turn this into a “Big Brother” nightmare of total surveillance and control. Second, it was widely agreed that facial recognition was the wave of the future for digital out-of-home advertising. The technology was so powerful, and the benefits in terms of ad targeting (and therefore cost savings) so great, that within a few years every DOOH ad platform would be equipped for facial recognition. This seemed especially likely as the cost of facial recognition software plunged, along with DOOH hardware, enabling any “mom ’n’ pop” DOOH network to use it if they saw fit.
But like many bold predictions, it didn’t come to pass — certainly not to the extent originally promised (or feared). While there aren’t many statistics on facial recognition adoption in the U.S., the generally quiet news scene suggests DOOH networks aren’t exactly rushing to implement the technology. Thus most of the growth reported by major players in the facial recognition arena, like Quividi, has come overseas, in Europe and Asia. So what happened to the great American facial recognition wave that never was?
Many DOOH networks have probably been scared off by “potential legal and privacy problems,” says Bill Gerba, cofounder and CEO of WireSpring Technologies, who has followed facial recognition technology closely. Gerba adds: “There have been lots of small-scale pilots and projects, but there hasn’t been any serious, widespread adoption — a lot of interest, but not a lot of traction yet.”
One contributing factor, Gerba noted, is the advent of new technologies that to some degree obviate the need for facial recognition technology — for example, near-field communication, which allows consumers to voluntarily identify themselves in interactions with digital marketing platforms in order to get targeted offers. Although this technology is also in its infancy, manufacturers and tech companies are hurrying to add NFC technology to mobile devices and apps to enable various kinds of interaction, including mobile payments.
Google and its partners are pushing NFC through a number of platforms, including Google Wallet, PayPal’s Android-based mobile payment system, and Foursquare’s latest Android app. (In the Android universe, NFC is called “Android Beam.”) Visa is incorporating NFC features into its new mobile account functions, including the V.me digital wallet, courtesy of Monitise, and MasterCard is quickly moving in the same direction. Meanwhile, major advertisers are also expressing interest in NFC in a variety of media: Most recently, Lexus inserted NFC chips into copies of the April edition of Wired magazine to promote its 2013 Lexus GS sport sedan.
Some of the most compelling applications for NFC are in DOOH advertising, where NFC chips can enable ecommerce transactions as well as identify the user. In April 2011, Reach Media Group and Blue Bite partnered to introduce NFC interactivity to RMG’s entire network of 75,000 DOOH displays, enabling easier downloads of mobile apps, coupons and offers, in addition to facilitating purchases and participation in rewards and loyalty programs. In January, NBC announced a partnership with VeriFone to enable mobile purchases in taxicabs and at gas stations using NFC technology; passengers who want to buy a product advertised on a video screen in a cab or gas pump can do so with an NFC-enabled mobile device.
Gerba points to other technologies, which allow users to voluntarily identify themselves, their interests and their locations to advertisers — for example, location-based networks like Foursquare and Gowalla. While these networks don’t necessarily give precise spatial coordinates for users, their demographic information can be quite detailed, and “people are obviously less squeamish about sharing personal information when you’re announcing to the world what you’re doing anyway,” Gerba observes. DOOH displays can easily integrate location-based networks, enabling them to deliver targeted advertising, promotions and so on through interactions with mobile devices with NFC or Bluetooth.
Facial Recognition Is Still in the Fight
It would be a mistake to conclude that advertisers and DOOH networks are simply choosing to skip facial recognition technology, however. Because while NFC and place-based networks offer many opportunities for identifying and locating individuals, the fact remains that only facial recognition technology is capable of gleaning identifying aspects of an individual. Likewise when it comes to precise tracking of individual movements — a key source of data for retail marketing — there are few alternatives besides technology that “look back.”
Gerba notes that location-based social networks “don’t give you the idea of how shoppers are moving around in the store, or traffic patterns, so there’s still plenty of space for complementary technology.” Similarly, location-based networks and mobile devices equipped with NFC don’t necessarily convey key demographic information including age, race or gender. Only facial recognition technology can provide these for every individual whether or not their mobile device is equipped with certain hardware and, indeed, regardless of whether they have a mobile device at all. While new, gesture-based DOOH systems, which invite participation, can make some data-gathering voluntary, there are other situations where passive data collection is the only or most effective option.
In fact, recent months have seen renewed interest in facial recognition technology, though it’s not always identified as such. Of particular note is the Audience Impression Metrics Suite created by Intel for DOOH networks following its acquisition of Cognovision in November 2010. The AIM Suite pairs biometric scanning with powerful embedded computers to ascertain “actual impressions, length of impressions, potential audience size, and gender and age range demographics.”
A number of DOOH networks have tapped Intel’s AIM Suite for audience measurement. In January of this year, InWindow Outdoor announced a pilot program to create fixed “experience stations” in malls and hotel locations nationwide in partnership with Intel. The freestanding, seven-foot-tall interactive kiosks will enable InWindow to develop a variety of “immersive” brand experiences, including gesture interaction, multi-touch screens and near-field-communication capabilities. Provision Interactive Technologies also announced that it was integrating Intel’s AIM Suite for audience measurement into its 3-D holographic kiosk displays. Provision will be deploying the kiosks in big-box retailers nationwide through a partnership with Premier Retail Networks (which has already worked with facial recognition technology for merchandising displays). And in February DS-IQ, which works with PRN affiliate Walmart, announced that it is integrating AIM into its Retail Media Platform, using AIM to incorporate real-time audience characteristics and engagement as inputs for store analytics around foot-traffic patterns, product engagement and product abandonment by store.
In short, some of the nation’s largest retailers will be deploying facial recognition technology in the near future — indicating that the technology is hardly in danger of being passed by. Of course, this simply raises the same privacy issues which bedeviled facial recognition technology when it first debuted several years ago — issues which haven’t been resolved since then. Chief among them: How should DOOH networks alert consumers that they are being watched by sophisticated technology in order to deliver targeted advertising? And will this admission generate any kind of backlash?
As it turns out, a set of privacy guidelines for using facial recognition technology in DOOH networks has already been drawn up by the Point of Purchase Advertising Institute. The voluntary POPAI guidelines include recommendations about the way information is gathered and used, covering key issues like whether data is retained. For example, “In no event should image, video or biometric data … be stored without an explicit consumer opt-in to do so.” Most currently available facial recognition systems for DOOH measurement conform with this rule, only using biometric information once before erasing it. On disclosure, POPAI advises “marketers must provide a disclosure notice to consumers who may be monitored … [which] should describe the … collection methods in effect and whether data collected … will be combined with other data including, but not limited to, register receipt information, credit card information, any NPPI information or data collected by third party and/or affiliate marketers.”
However, to date no major retailers or DOOH networks have adopted the POPAI guidelines, leaving it somewhat ambiguous what rules — if any — will actually govern the use of facial recognition technology. In other words, the industry is still at square one.