/* ---- Google Analytics Code Below */

Sunday, February 01, 2009

Proactive Merchandising

It is an old idea to connect sensors to advertisements and make them respond to people. As simple as motion sensors; my local 'Jungle Jims' has a huge Campbell's soup can hung from the ceiling that talks to you when you approach.

But what if you added advanced sensors that could detect the age, gender and other demographics of people as they approached? Even track the attentiveness of those interested. Tracked the interaction of their eyes? Then adjusted the message based on what it detected? Simplest thing would be to play alternative videos or images. We experimented with some of this in the innovation lab. Also because each interaction sets up an experiment and outcome, this could be linked to neuromarketing sensor data to classify specific behavior.

The sensor classifications are not perfect, usually 80% or so, but a miss is usually not a big problem, depending on the category. Or how about adding it to a system like Prism (abandoned by Wal-Mart and then by Nielsen), to get some additional measures for your in-store metrics?

News out now about Tru-Media, which describes their approach as 'Proactive merchandising', actually better described as adaptive merchandising. Their site has some good videos of the concept. Worth continuing to experiment with. As sensors improve we will see more of this. Most of these systems do not try to identify anyone, so there would seem to be little concern from the privacy direction. If you trust the people analyzing the images.

No comments: