What the Shelves Can Teach Us

Published in CSP Daily News

Mondelēz develops high-tech displays to promote products, learn consumers’ habits

DEERFIELD, Ill. -- Mondelēz International Inc. will bring new “smart shelf” technology to stores, pairing sensors with analytics, in an effort to break the code of consumers’ impulse shopping habits, according to a report in the Wall Street Journal.

The maker of Cadbury chocolates, Trident gum and other items found next to the cashier or in the cookie aisle will use sensor technology in the high-tech shelves to identify the age and sex of the would-be snacker, determine what type of guilty pleasure might most appeal to that customer and present a video to deliver custom advertisements.

"Knowing that a consumer is showing interest in the product gives us the opportunity to engage with them in real-time," Mondelēz CIO Mark Dajani told the newspaper.

This approach to personalizing with technology the impulse buy comes as the company continues to be challenged by the macroeconomic slowdown where potential snackers may cut down on non-necessary food items.

The smart shelves represent the latest example of the so-called Internet of Things, where everyday objects, outfitted with sensors, can tell businesses more about their customers and how their products are being used. Gartner Inc. predicts that net revenue generated by the Internet of Things will top $1.9 trillion dollars in 2020. Today sensors are found in everything from cars, to clothes and dumpsters.

When the smart shelf, now in prototype, goes public in 2015, the hope is that they will deepen the company's understanding of the customer. "When people walk by, it's a missed opportunity," Dajani said. "We must know how the consumer behaves in the store."

Here's an example of how it could work:

Sensors determine that a male teenager is standing in front of the shelf by analyzing facial structure and other characteristics used to determine age and gender. The sensors use this data to alert the display to feature something that a teenage boy is more likely to buy, such as gum or a chocolate bar. The shelves also use sensors based on Microsoft Corp.'s gesture-based Kinect for Windows technology, and if the boy looks at the shelf long enough, the shelf's display may play a video targeted for his demographic. Product weight sensors will know when and what product has been picked off of the shelf. If the boy has picked up Cadbury bar, the display could show a coupon to boost incentive.

The shelf will relay data in bulk to Mondelēz, whose engineers will analyze it with software. The system will compose vast anonymous profiles of consumer interactions with products, using the aggregate data on age, gender and product choice. For example, the analytics may conclude whether teenage girls prefer certain brands of gum, or whether middle-aged men cotton to Triscuits. Dajani said that collecting this data can be a powerful indicator of brands that are well placed--or not--in stores. Ideally, this will help the retailer change product placement, remove products and target in-store promotions. "Knowing that 500 people picked up the product helps determine if the smart shelf placement in-store is optimal to drive traffic," he said.