Carrefour to Use Bluetooth Beacons to Track Carts, Baskets

By Claire Swedberg

The company will be the second retailer to pilot a solution from Proximus to identify the locations of carts and baskets moving around stores, providing analytics regarding shopper behavior.


Multinational retail chain operator Carrefour is planning to test Bluetooth beacons to track the movements of shopping carts at three of its Madrid stores this fall. The technology, provided by Proximus, is intended to allow the anonymous tracking of shoppers as they move through the stores, by identifying not an individual shopper himself but rather the shopping cart he is pushing. While in the long run, Proximus intends to offer the solution as part of a retailer’s loyalty program that would track individuals via their mobile phones (as long as they opted in), the first step is merely to determine the carts’ locations, rather than identify the shoppers using them, according to Jorge Garcia Bueno, a cofounder and the CEO of Proximus.

Bueno and the company’s other two cofounders launched Proximus in September 2013, to develop and sell a technology-based product that would allow businesses to better connect with their customers in the way that online stores currently can. For instance, he says, online retailers can identify which products have been viewed and then tailor advertising to an individual based on that information, whereas brick-and-mortar stores do not have that option. Previously, Proximus’ founders had started up a company known as beMee, which leveraged Wi-Fi technology for tracking carts’ locations. In 2013, with the advent of iBeacons and Apple‘s release of the iOS 5 software development kit (SDK), they launched Proximus to focus on solutions leveraging battery-powered tags that transmit 2.4 GHz signals compliant with the Bluetooth Low Energy (BLE) standard.

Proximus’ Jorge Garcia Bueno

The company, which is headquartered in Spain but also maintains an office in London, is now trialing and releasing its technology in phases. The first phase involves the system now being installed at three Carrefour stores; the solution consists of battery-powered Bluetooth beacons, each of which transmits a unique ID number via the BLE protocol, as well as sensors that receive those transmissions and forward the data to a cloud-based server via a Wi-Fi or wired connection. While the next step will include the use of mobile phones (customers would load an application enabling the phone to transmit BLE data), there are initially no phones being used in the system.

Proximus had the beacons and sensors manufactured by a third-party vendor, and created the cloud-based software to manage the collected data. Bueno says Proximus spent three months piloting the technology this spring at two small Nisa neighborhood food stores in London. The pilot was completed this month, and the company is now installing the system at three Carrefour supermarkets, with a go-live date set for fall 2014.

With the Nisa trial, some beacons were attached to shopping baskets, while approximately seven sensors were installed in each store’s ceiling. The two sites have about 50 baskets apiece, Bueno says, though not all of them were equipped with the beacons. The beacons that were in use there, however, transmitted a signal at regular intervals that were received by the ceiling sensor devices, and the software employed the sensor-based data to determine each basket’s movement as it passed through the store. The system deployed at the Nisa stores could pinpoint location to approximately 1.5 meters (4.9 feet), Bueno says, which provided sufficient granularity to know in which aisle an individual’s basket was located. In this way, the software knew where the shopper went, where she paused, how long she was at the checkout area, and when the basket was returned to the storage area for another shopper’s use, thereby indicating that the individual had left the store.

Such data, Bueno explains, provides the retailer with considerable information. For example, a store’s manager could determine not only which products may be of most interest to shoppers, but also which aisles may be over-crowded and are causing bottlenecks. In addition, the software could identify when shopper traffic is highest and lowest, and which aisles experience less traffic than others.

Proximus is still evaluating the results of the trial at the two Nisa stores, Bueno reports. In the meantime, the BLE company is launching a more ambitious pilot with three Carrefour hypermarkets in Madrid, where thousands of shopping carts—also known as trolleys—will be tagged. Bueno says he cannot yet indicate how many ceiling sensors will be used, but notes that they will be wireless, sending data back to the cloud-based server via a Wi-Fi Internet connection.

The technology will be offered as software as a service (SAAS), with retailers paying a fee for access to the data and weekly analytic reports from Proximus; the price of the hardware could be included in that cost. The system is designed to be inexpensive, Bueno says. A retailer could deploy the solution temporarily at one store in order to gather data before moving the system to another site. However, he adds, Proximus’ long-term plans to transition into a second phase of the technology, to include loyalty programs using customer cell phones, would require that the technology be permanently deployed.

During the second phase, retailers would be able to use the technology as part of an app enabling customers to participate in a loyalty program. In this case, they would opt into a store’s loyalty program and could then earn rewards for providing such data as demographic information or a name or e-mail address. They would also download a retailer’s app that could be modified via an SDK (supplied by Proximus) to allow the retrieval of beacon-based location data.

In the case of such a loyalty program, the system would collect data regarding a participant’s movements, and could send promotional material, coupons or rewards. Conversely, it could simply collect demographic information relevant to that individual’s shopping patterns, based not simply on what she buys (with traditional programs tied into the point of sale), but also which products she pauses to look at.

The location data can be utilized for details beyond product placement as well, Bueno says. For example, he explains, if a retail chain intends to build a new store, it can determine the necessary parking area size or aisle width, based on Proximus system data derived at another store indicating how many people enter that store and how long they remain.

Bueno expects the Carrefour pilot to extend about six months, after which the results of that test will be evaluated as well. For now, he says, it is too early for Carrefour and Nisa to comment on those results. Proximus’ goal, he adds, is to offer the technology to Europe’s largest retailers, and to open an office in the United States by the end of 2015 for marketing in this country.