Executive Summary
Overview
The pending reconversion in grocery retail
The burden of barcode
Starting in the seventies, the barcode extended to achieve a complete and rapid identification without errors of article families in the cashier lines with the UPC and EAN standards. Over time, the new optical technology developed increasingly less demanding devices for the manual reading procedure: pencils, guns and a multidirectional scanner.
What was an unprecedented advance is currently a drag on the grocery distribution value chain that other sectors have overcome with RFID technology or are in the process of doing so. The disadvantages of optical technology in a context towards industrialization 4.0 are overcome with the advantages offered by radio frequency identification: non diaphanous omni-directional reading independent of a manual procedure, recycling of store human resources towards the online channel (order preparation and last mile logistics), traceability and perishables at the unit item level (SKU), inventory robotization, and packing with automatic reading of products in the box.
The 7 points of the pending conversion
The delicate crossroad in which the distribution of grocery is found, with unavoidable challenges (online cost overruns, manual identification, perishables and traceability without SKU, omnichannel stock, inadequate HR profiles and demographic threat), advises a change in the identification of products without the participation of a manual procedure that enables balance and management integration between face-to-face purchase and online demand, a demand that could end this decade by absorbing a third of grocery distribution:
1. Recycle non-added-value human resources from the stores to the online channel logistics and order preparation services.
2. Reduce and speed up the stay time of customers in their purchases in store, reducing physical contact (contactless, better prophylaxis) and without payment friction (frictionless, less time).
3. Reduce the high cost of managing perishables, and change batch-based traceability to unit code (SKU).
4. Facilitate a single omnichannel inventory (O2O strategy).
5. Reduce the cost of store inventory with robotization.
6. Provide pickers in online distribution warehouses with products with omni-directional identification without manual procedure.
7. Guide the future demand in homes with automatic replacement items (IoT, home automation in the kitchen, refrigerator, pantry …)
Fundamentals and Solution
1 – Safe migration to RFID (technological factor)
The migration of the sector towards RFID is inexorable in this decade with a progressive abandonment of the barcode as the preferred identifier for reading. The fast4shop ecosystem makes transactional security reliable, as a technological requirement.
2 – Cost maturation (viability factor)
In recent years the cost of the RFID Inlay with chip and small antennas for short-range contexts (50 cm, Near Field, fast4shop scope), has matured to the current environment of 2 c$, exceeding the cost limitation with the expected returns.
fast4shop is a solution to the 7 conversion points with a patented identification method that proposes a complete ecosystem for the safe migration of the sector to an RFID identification in an integral way, both in store and in the online channel:
Advantages RFID vs. Vision / AI + sensors fusion
The only two technologies that remain in competition for self-service and self-checkout without product scanning are: Radio Frequency with RFID and Vision/AI with Sensors Fusion.
fast4shop focuses on the RFID perspective with transactional preeminence in the Grocery sector, and proposes a residual and complementary use of Vision / AI technology: Merchandising, delivery of products in machines with facial recognition, payment systems…
Profitability
The surface in a store with Vision/AI maintains an inverse function with its profitability: agglomeration makes the cost of cameras, customer monitoring and the back office grow non-linearly.
Some cost and back office aspects of Vision/AI:
- Previous creation of image recognition patterns for each product in the laboratory with Machine Learning (ML) techniques, which may include printing special codes for remote recognition (hypercoding).
- The design change of a product by the supplier requires its validation in the ML laboratory for its display on shelves (rigidity).
- An essential support of the store with Vision/AI are the sensors and precision load cells (weight) for all the products on all the shelves, with their linkage and corresponding maintenance (extra costs).
Scalability
The size of the physical environment in which the technology works reliably is a limitation to the reconversion of the sector.
The reconversion of stores with Vision/AI technology + sensors is not fast or cheap, it requires a complete remodeling of the store, from the shelves, the sensors and the infrastructure with cameras. The conversion of a traditional store to RFID lies in the simple incorporation of tags in the articles by suppliers with industrial applicators, as well as fast4shop self checkout kiosks in Minimarkets, or the use of fast4shop carts and baskets in convenience stores and supermarkets.
Reliability
The vision systems as well as sensors and load cells on the shelves rely heavily on new image codes for remote recognition, which requires a particular ex novo labeling of part of the packages (blister packs, tubs, thermoformed, etc.)
Some weaknesses of Vision/AI systems with sensors fusion:
-
Bad faith practices for substituting used (consumed) product for a new one.
-
Bad faith practices due to intentional change of attire during purchase (change of recognition pattern).
-
Long wait for the receipt of some purchase tickets, probably due to purchase scenes outside the safety margin of the algorithms.
-
List of items in the purchase ticket open to the refusal of the client and trusting in their honesty.
-
Items are assigned to the customer who takes the item even if it ends up in the hands of another.
fast4shop with its patented 3-sensor heuristic is reliable even for products with liquid or aqueous nature and with metal packaging. Guiding the products through the fast4shop window ensures efficient readings at the distances of use of the RFID magnetic coupling (Near Field).
Traceability, perishables and sale in bulk
Identification by product image (Vision) does not resolve the health directive of traceability of the articles in a unitary way (SKU), nor does it allow a mechanized control of perishables, nor the sale in bulk except for products with certain conditions.
The EPC standard with 96-bit RFID achieves the level of resolution to encode a single item or SKU. The simple reading of the tag of an article resolves its direct traceability without the need for the external lot attribute (barcode), as well as its expiration if it is a perishable. The trend of self-service bulk sales is solved with scales that print labels with RFID tag.
Comparison RFID vs. Vision / AI
