Facial detection and recognition are burgeoning technologies that are beginning to be applied in the marketing industry in a multitude of ways. The technology is already having a profound effect on how we analyze audiences in the real world as it provides hard data on consumer behavior that we were traditional only privy to online. But that is just the start. Facial detection could reshape everything from security to loyalty programs and payment…and privacy concerns abound.
Facial Detection vs Recognition
There is a key difference between detection and recognition.
Facial detection is the anonymized analysis of facial features to determine characteristics like age, gender, or ethnicity. The image of a person in view is compared to a vast database of faces to make a determination. This is being used for audience measurement, particularly enhancing OOH and retail as well as emotional tracking for research.
Facial recognition is the positive identification of a unique individual based on their facial characteristics. It’s the difference between identifying a young adult male vs. Jack Pollock. With recognition, the image of an individual is directly matched against their archived profile. This is primarily being used for security purposes within residential complexes, prisons, and airports but in the future it could hypothetically power loyalty programs. Imagine walking into a Starbucks in which a camera correctly IDs you and relays your favorite order to the cashier with your payment details already on file. You can simply pickup your coffee and walk out the door.
Detection and recognition are both done through software solutions that traditionally run on webcams and surveillance cameras. Facial detection, however, is the primary technology for marketing use cases.
There are a number of companies that are offering audience measurement solutions to determine the demographics of people exposed to OOH advertising, entering retail stores or other venues. These solutions can detect age, gender, size of the audience, distance from a display, dwell times and more. Lighting and range are certainly a factor and depending on the configuration, these solutions can range from 90% accuracy to far less. The result is a granular look at foot traffic, attention and audience makeup that can be used for ROI tracking and more. Simply put, it’s Google Analytics for the real world.
Beyond measurement, this data can be used for creative optimization, displaying media according to the audience in real-time. Just take a look at Plan UK’s Because I am a Girl campaign which only displayed OOH video to females in order to raise awareness around gender discrimination.
Because this software can detect facial features and movements, we can also ascertain emotions or mental states. If my mouth curls upwards, for instance, I am generally happy, provided I am not Jack Nicholson in The Shining…which is something to consider. We are not robots. We respond very differently to stimuli so coming up with a standardized emotion score is impossible.
That said, emotional recognition can be a powerful tool for market research and focus groups, providing quantitative analysis of a measure previously untrackable. And there are a lot of vendors out there doing just this.
Additionally, clever marketers have used emotional tracking to trigger different events and media. Just take Douwe Egberts Yawn Activated Coffee Machine, which uses facial detection to identify tired passersby before dispensing coffee.
Currently, facial detection has been deployed across the globe. There are often some careful limitations to how the data is collected. For instance, camera images are often not stored and at some times—as was the case with Plan UK—the metrics aren’t even being collected. Even so, consumers are wary about being on camera and being tracked. If marketers do decide to leverage this data for communication, they should use it to the consumer’s benefit, whether it’s offering enhanced service or simply being clever. If they do not, they will pay the price. Just check out some recent backlash from this tech being used to identify the makeup of bars in the US for reference.
Interested in learning more? Shoot us an email at [email protected]