Retail and RFID clearly have a very ambiguous relationship. On the one hand, the technology seems to come from another age when compared to the Internet of Things and the first AIs currently invading the specialist media. On the other, RFID seems never to have reached maturity to fulfil retailer fantasies about geolocation, real-time inventory, and end-to-end tracking.

Yet, today, according to a report from Auburn University (Alabama, USA), 96% of retailers say they are interested in RFID and plan to adopt the technology.


A look at the Gartner hype cycle seems to confirm that the RFID reached its productivity plateau three years ago. This is due to several factors, starting with the drastic drop in the cost of the technology throughout the chain. The price of single tags is now close to 3 cents, while the cost of an RFID reader is around €2,000 - not much more than most radio terminals used in shops today. The number of RFID suppliers on the market today is also helping to significantly reduce costs. Yet very few companies integrate the entire RFID chain of chips, tags, readers, application, and integration. In addition, the EPC (Electronic Product Code) codes are standardised, for example by GS-1, which guarantees their uniqueness.

Moreover, most retailers have in fact already adopted the technology in warehouses and shops, with varying degrees of success. Across the Atlantic, Macy's is now aiming for 100% RFID and enables "Pick to the Last Unit" thanks to almost real-time stock accuracy. Decathlon has been tagging nearly 90% of its articles at source for nearly four years, and Hervé d'Halluin, in charge of the RFID programme at the French sports equipment leader, recently announced the launch of phase 2 of their programme: targeting 100% RFID in June 2018. The result means fast checkout, easier customer information via interactive terminals, and accurate stock control. RFID is also used to prevent counterfeiting at Moncler. And it equips interactive booths at Rebecca Minkoff, John Lewis, and many others.

However, beware of the promise of "100% geolocation – JIT stock - real time". To have a real-time view of the entire stock would be like lining all your ceilings, drawers, cupboards and even storerooms with RFID sensors. All of which is ideal, but only in a metal-free shop. Geolocation to the nearest drawer would require not only the notion of a drawer in the shop back office, a feature that is extremely rare, but also having a sensor in each drawer and isolating each drawer from its neighbour. With immediate consequences for your CAPEX, and a degree of operational efficiency that would need to be demonstrated. Even if the technology continues to progress with the appearance of solutions for perfume bottles, for example, enabling Sephora to embark on its DIY adventure, there is still some way to go.


It is interesting to draw a parallel between the coming of age of RFID, which until then had difficulty finding its place in the retailer's value chain, and the ambition of omnichannel. Could the latter have supported the development of the former? Accurate stock assessment is a prerequisite for aligning an off-line with an on-line offer. RFID makes it possible to meet the customer's promise of item availability without having to concentrate in-store efforts on counting stock. The most recent surveys of retailers, particularly in the US, show that stock accuracy is close to 95%. The remaining 5% is attributed to process errors or technology limitations. Stock accuracy is the first step in any omnichannel ambition.

Évolution des usages de la RFID

Achieving ROI on stock accuracy alone seems difficult. Unlike other technologies which appear to provide the same guarantees (GPS, Bluetooth, NFC, Wi-Fi, etc.), RFID makes it possible to achieve a faster ROI due to the addition of operational gains throughout the retailer's value chain. A Beacon is expensive, NFC requires too much proximity, Wi-Fi is not secure or precise enough, and so on.

Most retailers who have implemented RFID have seen direct effects on reducing shrink (average of 5 points) and breakages (average of 8 points), or increasing the operating margin (average of 5 points). In addition, RFID reduces non-sales and improves additional turnover. By uniquely identifying the article, you can complete your customers' shopping basket by cross-selling. As well as avoiding abandoned shopping baskets, a fast checkout will also allow you to put your employees back in the middle of the sales area, in contact with the customer. The tracking of articles entering, leaving, and passing or not passing through the checkout, will offer a new source of data on possible fit problems for your articles or those of your suppliers. And there will be other gains specific to each business model that each retailer will have to invent and collect.


The question seems legitimate. As we have seen previously, it will not achieve the miracle of "100% - Real Time - Localisation". But what if the technology’s limits were precisely the key to the success of RFID projects? Before choosing the technology, the precise model must be selected as it would be very risky to "use RFID because it’s trendy". But the next question is which technology and RFID. Is Beacon better suited to what you want to achieve? Is robotics with image recognition more suitable for your model? Is your item-naming system adapted to the unique code specific to RFID? Are you looking for operational efficiency, or a new customer experience, or simply innovation for innovation's sake? These are by no means incompatible. But they must be RFID-compatible to be truly feasible.

RFID may be a fickle technology, but it is the simplest in the contactless category. A unique code, a sensor, and a database. One of the biggest risks is therefore to force the technology to fit into your processes and correspond to what you want it to do. In short, if you add complexity to a technology that is not intended to be complex, this is the best way not only to add extra costs but also to rob RFID of what it can do best and as simply as possible. An EPC, because it is captured on the fly, cannot, for example, communicate to several shelves. RFID's contactless and distance concept prevents you from picking retail items. No point in trying - you won't succeed. Retailers who have implemented RFID have had to rethink most of their shop processes to get the most out of it. And an RFID programme is often the perfect opportunity to do just that.

As you can see, an RFID installation is a special project. First of all, because although it will partially transform your business model, it is not a Big-Bang project that will revolutionize your shops overnight. Decathlon has therefore broken down its RFID programme into two parts for the time being. Behind the term "RFID programme", your model will be equipped with a range of functionalities that will be added to each other on top of a technology base of tagging - inventory, quick checkout, anti-theft, connected mirrors, threshold crossing location, etc.

Next, some of your specific features will have an impact on how your project is implemented, and these should be studied to measure their impact. What is your marking strategy for your products? Do you control your own production chain? What types of products do you sell and what is your shop architecture? Which processes are integrated in the shop and which in the warehouse? Which application architecture do you use? In-house or outsourced inventory? Regular or annual inventories?

By using this set of questions, you will be able to outline your RFID strategy and measure the impact of your technology implementation.