The Science of Shopping
Next, he says, marketers realized they had to consider not just their product, but the entire assortment of available products, the similarity and complementarity of the items, and the allocation of shelf space. This evolved into category management. At the same time, the next wave of technology hit, the UPC scanner, offering detailed point-of-sale data. Manufacturers and retailers could track what was being sold, when, with what and at what price. Volumes of data were being collected, and so grew marketing research techniques to mine and warehouse the data.
Up to this point, as Burke explains it, manufacturers and retailers had a very product-centric view of the world. With the advent of consumer loyalty programs about 10 years ago, there was a switch to collecting more consumer-centric data-who's buying what, when, where and the characteristics of the household-and tracking consumer behavior over time. From these data, loyalty measures were calculated along with lifetime customer values.
Today, "we know what consumers are buying, but not how they're buying," says Burke. The tracking and observation technology Burke and his graduate students have developed allows them to monitor consumer behavior and study shopping patterns. A panoramic video camera with six lenses takes in a view of the entire store, capturing 30 gigabytes of data an hour.
By observing customers' paths through the store, Burke can determine the obstacles to purchases, and he can also identify what will cause consumers to stop in their tracks. "Consumers have a limited amount of time to shop. For retailers to remain competitive, they need to make the shopping experience as convenient and enjoyable as possible." This type of merchandizing optimization information helps retailers increase sales while improving customer satisfaction and loyalty.
The computer tracking information Burke collects is anonymous. The location and path of the customer is tracked and whether or not a purchase is made, but not the identity of the customer. The data that Burke and his colleagues collect is not shared with others. It's only made available to retailers with whom they work, and only in the form of aggregate statistics.
Although huge amounts of data are being collected, the tools to analyze the data are still being developed. "The whole area is becoming more analytical," says Burke. In the future, he expects that the market will see the same kind of gains from merchandizing optimization and product presentation that are now being achieved through price optimization techniques. In the meantime, Burke's students are getting a hands-on look at this evolving technology as they get opportunities to work in the lab with live data from collaborating retailers.