Radio Frequency Identification (RFID) can be used to improve an organization’s efficiency in track & trace of its goods and assets. We explore how an RFID application is structured to achieve the same.

A Supply Chain consists of customers who buy goods or services from a retailer. The distribution channel could be a store, mail order or an e-commerce website among others. The retailer stocks goods from a wholesaler or a distributor. The distributor buys goods in large quantities from a manufacturer who makes the goods in a factory or a production facility. The manufacturer buys raw materials from suppliers.

A typical supply chain has one or more of each of these entities. There could be multiple tiers of suppliers, manufacturers, and distributors. As materials move from the initial supplier in the chain to the customer, value and costs are added at each stage. As you get closer to the retailer, the supply chain becomes more complex, with different products sourced from many different business partners; a large retailer would sell thousands of products sourced from thousands of direct suppliers.

The key to making supply chains more efficient is by improving the information sharing through all stages. At each stage in the supply chain, there is a need to forecast, collect actual sales from the next stage, plan on what and how much to make or order. RFID technology can help with all of the above to improve efficiency and earnings.

Current processes are complex and labor-intensive. As products arrive from manufacturers, a physical scan is required to read the barcode on the pallet and on each case on the pallet for each stage. For entities downstream, this is simply a cost of doing business. Our goal is to automate, to achieve significant labor savings, reduced data-related errors, and improved product availability.

While RFID can impact many different areas of the supply chain, let us look at a few scenarios between a supplier, a manufacturer, and a retailer, and identify various processes that can benefit from enabling RFID.

Let us take the hypothetical case of Acme Industries which is a manufacturer of widgets. A-Mart is a retail chain store which sells these widgets. It has around four distribution centers located around the country. Perfect Parts Ltd is one of the suppliers of widget components (widgelets) to Acme Industries. Mapping these entities to our Supply Chain stages above, we get the following:

Widgelet Supply Scenario
Acme industries receives widget components(widgelets) from Perfect Parts Ltd in response to demand from A-Mart for widgets. Prior to a RFID solution, there was no way to forecast demand or plan for widgelets supply in real time. For e.g. in response to a special one time bulk order from A-Mart for a custom widget, Acme placed an urgent bulk order for a custom widgelets on Perfect Parts. Perfect Parts had to scramble to get the order filled by shipping these widgelets through a high priced courier to Acme. They also had to eat the extra cost of transportation, cutting their margins. A side effect was their demand forecasting system incorrectly generated a report indicating a spike in demand for those widgelets. This resulted in having unused inventory in their warehouses.

Solution : Acme Industries and Perfect Parts went for a RFID solution. Acme would have a three widgelets storage areas fitted with RFID antennae that are connected to an RFID Reader. As RFID-tagged cases of widgelets arrive from Perfect Parts, antennae read the tags and update the widgelets inventory. When widgelets are taken out to be assembled in to widgets, inventory is reduced appropriately. When the inventory goes below the reorder point, a replenishment signal is generated and sent to Perfect Parts.

This solution enables Perfect Parts to know in real time what the demand is. They don’t have to scramble as previously to satisfy demand. They would know much earlier when the reorder point was reached.

Widget Assembly Scenario
Acme Industries has different assembly stations for different types of widgets.
Based on the day’s orders for each assembly station, a requirement list is generated and sent to the storage area. Every morning, the assembly warehouse, picks up the widgelets according to the assembly workstation requirement list and drops them at the appropriate assembly stations. As a forklift leaves the storage area, widgelet tags are read and the inventory is updated in real time. As each lot of widgelets is dropped at each assembly station, tags are read and the system verifies that the right type of widgelets is dropped at the right assembly station. Once a widget is assembled and packed, a Case RFID Tag is created and placed on the widget. The system associates the tag ID of the widgelets to the tag ID of the widget.

As widgets are assembled, a notification to deliver the widgets to the quality control area is sent. The widget cases are loaded onto a fork lift and driven away. The widgelet inventory at the workstation is automatically deducted. If the inventory falls below a certain level, a notification is sent to the assembly warehouse to replenish more widgelets from the storage area.

Quality Control Scenario
Before Acme Industries started using RFID for widgelets and widgets, they had to pick a widget case randomly from a lot, open it, take the widget out and perform QA procedures. With RFID, they simply scan each case that contains the widget, and the system reads the widget tag and the tag on the widgelets automatically. It checks those tag IDs against the bill of material (BOM) assembly information and alerts if there is an error. By using RFID, they are now able to process more cases during the day. As each widget that is assembled in their factory is inspected, their quality has improved from 80% to 99.99%

Widget Packing & Shipping
When the widget cases reach the packaging area. Start Pallet Shipment prints a new RFID tag to be affixed to the pallet. The appropriate cases that need to be shipped from the storage area for the order are placed on the pallet. An antenna in the shipping area picks up the new pallet tag and the case tag for each case being placed. The system alerts, if wrong case is placed by mistake. Once all the cases are placed, the Finish Pallet Shipment process associates all the case tags placed on the pallet to the pallet tag. The system also updates the Finished Goods storage area inventory. The assembled pallets are loaded on a waiting truck. As pallets are loaded, the system verifies that the right products are being loaded for the right customer distribution centers or stores. If there is an error, the system generates a visual alert to respond and take corrective actions. Once all the pallets for a retailer are loaded, the Complete Shipment process generates an Advanced Ship Notice and sends it off to the retailer.

Before Acme started using RFID, they had many shipping errors. Wrong widgets were being placed on wrong pallets and wrong orders were being delivered to stores. Customers simply returned the shipments. Acme not only had to take charges for the late deliveries below promised service level agreements, they had to pay for reverse logistics too. With RFID they could reduce shipping errors from 1 in 100 to 1 in 1000!

A-Mart Distribution Center
A container of widgets arrives at the A-Mart Distribution Center (DC). The driver’s identity badge is scanned at the receiving dock. The system verifies identity and opens the receiving door. The pallets are unloaded into the DC. As the pallets move from the truck into the warehouse, they pass an antenna-portal connected to a RFID reader. This reader picks up the pallet tag and identifies the ASN sent by Acme. As each pallet is unloaded, the system verifies the pallet contents against the ASN. A screen next to the receiving door displays any errors. Once all the pallets are unloaded, the Finish Shipment process is initiated. The system records the shipment, updates the DC inventory.

De-palletization and store orders
Every morning the pallets received from manufacturers like Acme are taken to a de-palletization area. The DC runs two conveyer belts from the de-palletization area to the various store order assembly areas in the DC. The operators at the de-palletization area take out the cases and place them on the conveyor belt. The conveyor belt routes the cases to the appropriate store assembly location. Each location assembles the order for a given store for a given duration of the day. Near each store assembly area there are antennae that sense the tags on the cases moving over the conveyor belt, and open and shut the appropriate gates to route the boxes to the right store assembly area.

Before A-Mart DC started using RFID, they were using barcode scanners on the conveyor belt to route the cases. This was a very error-prone process, as less than 60% of cases were being read correctly because the bar code scanners were ineffective when the bar code on the box was not facing the scanner. As RFID does not require a line of sight, A-Mart is achieving almost 95% accuracy, their store orders are being assembled more quickly, and the average time merchandize stays in the DC has been reduced from three days to one and a half days.

A-Mart Store
A-Mart Stores receive products from their Distribution Centers and use RFID to track their inventory in the back room and on the store floor for selected items. Store Level Inventory Control is done using a perpetual SKU-level inventory process that tracks and traces items through receiving, shelf-stocking, and point-of-sale within the store.

Receiving
As soon as the order is shipped from the DC, an in-store receiving function receives data indicating, to expect a specified shipment. When the shipment is unloaded at the back of the store, the system reads the shipment’s RFID tags, accessing item descriptive data, summarizing it if necessary, and comparing it to the expected shipment. Any discrepancies are noted and reported. When receipts are read through the RFID back room readers, they are logged into a store perpetual inventory database as available for sale. Any plus or minus from the expected orders are logged, but the actual items read are assumed to be correct and put into the store inventory of record.

Shelf Stocking
As pallets of merchandise are received in the back room of the store, this data is made available to a shelf-stocking scheduling function. This application is aware of the current inventory levels in the store and will schedule stocking of merchandise that is either out-of-stock or at a low inventory level first. Merchandise that is hard to stock or bulky will be scheduled during times when the store is closed or light traffic is expected. Merchandise is stocked in a sequence that spreads the available stocking labor throughout the aisles of the store.

Available store labor resources are taken into account when a store stocking schedule is produced. The stocking application can present the store associate with a shelf-stocking list either on a printer or wireless terminal. The shelf-stocking work list indicates the location of the merchandise on the pallet, as well as indicating a shelf location for the merchandise to be put away.

When the stocking associate indicates that the shelf or rack stocking is complete, an RFID shelf checker application can audit the restocking function and the store shelf inventory levels. Some stores have backroom or secondary stocking areas within a store. The stocking function includes not only the putting away of new merchandise, but the moving of merchandise in secondary locations to a primary selling area.

Store Replenishment and Ordering
Store inventory levels are checked on a preset schedule by RFID technology and wireless ordering terminals. Inventory discrepancies between the items “read” on the shelves and other store locations and the perpetual store inventory can be noted if reasonable difference percentages are exceeded. Some discrepancy is to be expected, from some items “in transit” to the front end and a certain level of unreadable tags in the store. After inventorying the shelves and including the custom orders, the system generates the order and checks it against the supply chain for any likely problems.

Proposed Software Architecture
We can now layout the architecture based on the solutions that we have identified above.

The above architecture can be extended to interact with business partners, vendors etc. The final implementation would change based on the specific situation encountered.

Conclusion
In this blog, we first discussed the major challenges that could possibly occur while developing an RFID-enabled supply chain. Then we saw how a RFID enabled solution would alleviate and significantly improve supply chain management. Then, we proposed an architecture, keeping in mind these challenges and how we would overcome them.