Two years ago a Japanese company started to attach a little camera onto its billboards to be able to measure who much attention they attracted, or to be more precise, to measure how many people looked at their billboards. There is also a second camera below the billboard which counts the total number of people who are near the billboard, allowing to tell how many of the people passing by actively looked at the billboard (eg 50 out of 200 people walking by).
This technology got revamped recently, adding the ability to tell if the person looking at the billboard is male or female and also estimating the age. The billboards however do not instantly react to the number/gender/age of people passing by but assists the owner of the billboard to maximise the use of the billboard by displaying that kind of advertising that proves most efficient for the individual audience.
A similar system, lacking the age recognition is allegedly already under operation in s.Oliver stores where they are used in the waiting areas of elevators or around the changing rooms.
For both systems it is claimed that they do not store information about the individuals, just whether they have looked at the billboard or not.
In my opinion these two techniques are in fact pretty similar to web-analysing software, as they also measure, how many people in total were exposed to an ad during a certain period of time and how well the ad actually performed.
While in the online-word the “click” (or in some cases even a “conversion“, that is a certain action carried out by the user, such as an online purchase or an subscription) has become the leading marketing currency, “long time fixations” might become a currency in the offline world. (My usual caveats about the phenomenon of Inattentional Blindness of course also apply 😉
So while usually off-line concepts are being applied to on-line matters it looks to me that this time it was exactly the other way round.
And, in case you really want to avoid being identified by a camera you will need to put on some cool make-up to deceive the face-recognition algorithm: