What Are Your Customers Thinking? Next-gen technologies predict consumer behavior.

Technology researchers are working on something big: a device that combines a digital screen and camera to analyze your physical characteristics and play back personalized advertisements. It isn't ready for market yet; but, when it finally is, it could revolutionize the way businesses reach out to their consumers.

Here's how it would work. When shoppers pause in front of a monitor, a computer reads their sex, age, race and expression to gauge interest levels, and then decides the best commercial to play. If the person turns or looks away, the device attempts to draw attention back to the screen--perhaps the music will suddenly crescendo--and the customer will see a different product the computer thinks may be of interest based on the information gleaned earlier.

Spearheading the project, called "Targeted Advertising Based on Audience Natural Response," or TABANAR (thankfully) for short, is NICTA, an information and communication technology center established by the Australian government. According to Glenn Downey, NICTA's commercialization manager, progress depends on how quickly researchers can develop sophisticated ways to identify emotional cues from facial expressions. She says a breakthrough would represent a transition from dynamic to responsive technologies. "Next generation technology will react to interest levels and shift content accordingly."

What's Available Now?
TABANAR is a work-in-progress, but there are existing technologies that businesses use to get a read on what customers are thinking.

For example, mining video for customer data is gaining popularity. Rajeev Sharma, founder of Pennsylvania-based VideoMining Corporation, created the company in 2000 to meet demand for retail intelligence revealed through image recognition software. In VideoMining's client stores, feeds from security cameras are sent to a main computer, which extrapolates information on everything from what products people are looking at to how long they stand in front of a particular display.

The business applications are clear. "The store design and merchandising would be fairly critical for all business owners, and there's a growing interest in video mining because everyone is trying to compete, small or big, and trying to differentiate their stores," Sharma says. With this kind of feedback, store owners can assess the effectiveness of a display and alter the design of the store to maximize sales.

In one grocery store, Sharma recalls, they discovered there were too many product types in the juice section. "Ten percent of shoppers spent 90 seconds in front of the display before leaving without buying anything. We pinpointed that people must be overwhelmed by products. They responded by reducing the number and organizing it better. It worked."

Right now, VideoMining's clients are mainly large chains and consumer brand companies, but, Sharma says, any business with a storefront can benefit from a better understanding of consumer shopping behavior. And services are more affordable than you might think, given that the biggest cost is usually outfitting the hardware.

The company is also beginning to offer clients real-time measurements on an analytics website. "Just like an online business gets immediate feedback on the number of hits, we can do the same with different displays in a bricks-and-mortar store," Sharma says. "This way, retailers can change displays quickly and experiment to find out what works best for their individual needs."

Merchandising firm YCD Multimedia is rolling out a digital media platform featuring video screens that play advertisements related to what people put in their shopping carts at points of sale. For instance, at the cash register at Aroma Espresso Bar (one of YCD's clients), a woman buying a coffee might see a commercial for a scone on the monitor.

CEO Barry Salzman believes the retail environment of the future will integrate the ability to measure which promotions are working and which ones aren't with the ability to act upon this information immediately. YCD's platform allows retailers to cheaply measure, in real time, the performance of marketing campaigns and concept testing. This, he says, unlocks a realm of possibilities for retailers to try things depending on time of day and location. "Our clients can log onto our analytics website and see that promotion X got put up in a particular aisle at noon, and at 12:15, what's happening at the cash register."

Salzman adds that the initial capital investment that scares people away saves them money--and increases sales--in the end. "Once you get the system in place, you push a button and get everything at a minimal cost," he says. "It will replace the waste of shipping and having to change posters and other static materials that may be outdated by the time you get relevant sales data."

A Look Inside
While Sharma and Salzman are decoding consumer behavior by observation, neuromarketing expert Martin Lindstrom is doing so by reading shoppers' minds--literally.

In his book Buyology, Lindstrom describes the findings of a three-year, $7 million study that examined subconscious shopping behavior using brain-mapping techniques. "We know that 85 percent of every purchasing decision . . . is made in the unconscious part of the brain," he says. "Now we can access this using fMRI and EEG scans [of the brain]." One aspect of his research measured how different regions in the brain reacted (or didn't) to certain advertising-related stimuli, including sound, smells and visuals.

Turns out, the sense of sound makes the biggest emotional impact, followed by smell and then sight. "If you expose people to sound, all five sensory regions are activated, which means that sound has much more influence on our mood, on our choice of brands and our emotional engagement than visuals."

These findings have several practical applications. First, Lindstrom says, companies should leverage the internet's sound capabilities. Less than one percent of business home pages use sound. He suggests that even something simple--like a short tune when a credit card is processing, or an introductory theme at log-on--works to get people into a certain mindset when thinking about a brand.

It's also important to make sure advertisements are placed in the right context. "If you're watching or listening to a dramatic, fast-paced program, a commercial break featuring Dove beauty soap will be forgotten," Lindstrom says. "If the brain can't figure out how it fits with the storyline, it will literally delete it." So even if you're using a great advertising medium, an ad appearing at the wrong moment negates all the benefits.

Another interesting discovery was that people subconsciously purchase more expensive brands when others are around. "If they're totally alone, people are more likely to buy generics," Lindstrom says, adding that designing a more open space with lower aisles can also deter theft.

Now we know that store design can influence what people buy, he says. And as brain-scanning technologies become increasingly portable, with studies done in real shopping environments, the results will yield further insights.

Reading Ahead
Given the trajectory, Guy Hagen, president of technology consulting firm Innovation Insight, reflects that as these technologies are refined, the pool from which to extract market intelligence will grow enormously. "The demographic data we can get right now is important, but if we could access data on expressions and attitudes, that takes things to the next level," he says, which could open up all consumer-generated video and photographic footage to piece together a larger picture. "We would have richer information, and more of it."

And although privacy issues will certainly arise, Hagen thinks that is also part of the natural process. "Research has shown that in general, people will put up with privacy invasions if they get enough benefits from it."

One thing, however, is clear: With next-generation advertising technology, you may not be able to read your customers' minds, but you can get pretty close.