Communications of the ACM
Auto-ID: Managing Anything, Anywhere, Anytime in the Supply Chain
Auto-ID is poised to be the next big wave to hit the supply chain. Emerging from the research and development of bar codes, bar code readers, and universal product codes, the tagging of items for sale has been used in the retail industry since the 1950s. A new development is Auto-ID, or radio frequency identification tags (RFIDs) embedded as a chip in each item a manufacturer produces [6].
Incorporating a 96-bit electronic product code, the chip is affixed to each physical object (see Figure 1). The numbering scheme means the tags can be used to uniquely identify more than one hundred million manufacturers, along with one million of their products, yet leave enough numbers for items to be produced and tagged in the future. Antennas enable the chips to communicate wirelessly with radio frequency readers. The readers trace where a particular item is in the supply chain. The manufacturer can place wireless readers strategically along the supply chain, so the unique Auto-ID is transmitted to the Internet (where more detailed information about the product can be stored) and communicated to manufacturers, distributors, retailers, and even third-party logistics providers, on demand.
Auto-ID technology thus creates an "Internet of things" [11]. For example, a manufacturer of soft drinks can identify with the click of a button how many containers of its soda cans are likely to reach their expiration date in the next few days and where they are located at various grocery outlets. Using this information, it might modify its future production and distribution plans, possibly resulting in significant cost savings.
Auto-IDs are likely to affect every aspect of supply chain management (SCM), helping improve demand management, customization, and automatic replenishment of out-of-stock goods while reducing inventory and distribution costs, as well as counterfeit versions of name-brand items.
Flow of Materials
Although its scope may differ from industry to industry, SCM generally refers to all management functions related to the flow of materials (supplemented with information) from an organization's suppliers to its customers. The main SCM objective is maximum customer satisfaction at the lowest possible cost. Achieving it requires organizations maintain an optimal balance between supply and demand throughout their supply chains.
According to the Supply-Chain Operations Reference model [12] developed by the Supply Chain Council, SCM involves five core processes: plan, source, make, deliver, and return. They are conceptualized, planned, executed, and monitored at three distinct yet interrelated levels: strategic, tactical, and operational (see Figure 2).
Strategic SCM typically involves a long planning horizon and top-level executive participation; examples include market evaluation, business partner selection, capacity expansion/contraction, product introduction, and technology adoption. While the extent and frequency of planning are least demanding at the strategic level, their effects are the most significant. Tactical SCM moves forward through business models adopted at the strategic level, with further planning involving yet more details. The planning horizon at this level is shorter (usually weeks or months) than at the other two levels.
Tactical SCM consists of demand planning, inventory planning, and master supply planning. Demand planning helps generate accurate sales forecasts and a sales plan. Inventory planning helps adjust optimal inventory levels (safety stocks), hedging against variability in demand and the lead times of various activities in the supply chain. Master supply planning involves procurement, production, distribution, and transportation. These decisions are further analyzed, executed, and monitored at the operational level.
For operational SCM, the planning horizon is the shortest (usually hours, days, or weeks) of the three levels, and the number of people directly involved is the greatest of the three levels. It includes demand fulfillment, the scheduling of procurement, production, transportation, and monitoring, as well as corrective measures in light of changing conditions.
A typical supply chain (see Figure 3) consists of supplier, manufacturer, distributor, retailer, and customer. Auto-ID promises to promote an open system [3] at all levels of SCM, enabling any company to read all the tags and reap maximum benefit from its accrued information.
A manufacturer of soft drinks can identify with the click of a button how many containers of its soda cans are likely to reach their expiration date in the next few days and where they are located at various grocery outlets.
Auto-ID in Supply Chain Management
Auto-ID is likely to affect all three SCM levels:
Strategic. While Auto-ID may promise many benefits, senior managers must determine whether it is right for their particular organizations' supply chains by conducting feasibility studies covering cost-benefit evaluation, sensitivity analysis using multiple scenarios, technology readiness/ease of implementation, competitor positions, and alliance partners. They must also envision an appropriate adoption plan, including the division(s)/location(s) that will implement it, as well as the timing for and type of adoption process they will use.
Tactical. Suppliers, manufacturers, distributors, and retailers all need sound demand planning for the efficient allocation of resources. One of the main difficulties in demand planning is a lack of reliable data. Adopting Auto-ID would produce accurate information related to the inventory of finished goods, work-in-progress, and in-transit stages with reliable due dates. Without Auto-ID, such data is either not readily available or inaccurate due to potential human error. Auto-ID promises to eliminate these inaccuracies. With real-time data provided by Auto-ID, the forecast will be more accurate, and demand planners will not have to create extra buffers to ward off uncertainty.
Auto-ID will enable process automation in picking, shelving, cross-docking, and implementing consolidation operations. It will reduce mistakes (such as sending an item to a wrong destination and not dispatching the right item at the right time). These measures will reduce the cost of operations for manufacturers and distributors and cut transportation and distribution lead time and lead-time variability. As a result, more limited safety stocks will be needed throughout the supply chain, resulting in better inventory planning.
Safety lead time is an important consideration in master supply planning. During each cycle of master planning, the number of days of supply in buffers in the supply chain are reviewed and adjusted based on this lead time. The time fence, or the portion of the planning horizon in which supply plans are firm, usually spans a period corresponding to the time it takes the organization to modify its resource availability. Organizations today increasingly look to reduce this frozen period so their operations are as adaptable as possible to changing conditions. With enhanced process automation and tracking capabilities enabled by Auto-ID, the velocity and visibility of products in the supply chain will likely improve, thereby reducing the duration of the frozen period. With more stable demand planning, balancing supply and demand is also likely to be easier. The three components of master supply planningprocurement planning, production planning, and distribution and transportation planningare all likely to benefit from these improvements.
Operational. The process of demand fulfillment, whereby corporations make promises to customers and keep them, relies on available-to-promise (ATP) and/or capable-to-promise (CTP) figures. With near-perfect supply chain visibility, ATP and CTP calculations will be more accurate, helping suppliers, manufacturers, and distributors streamline the order-fulfillment process. Using Auto-ID, manufacturers will be able to instantly check the availability of parts and their locations in the supply chain. If they're not available on time, sourcing them from alternative suppliers will be easier.
If the major components of a finished item are tagged with Auto-ID, the organization might choose to adopt more automation in the production line and reduce its manufacturing cycle time. Passing through different machines, Auto-ID-tagged components can help track time spent at different stages, reducing idle time while possibly increasing production throughput.
Since the early adoption of Auto-ID is likely to involve finished goods, transportation and distribution operations in the supply chain will be affected in the near term, and distributors and retailers will be able to improve their operations.
Labor costs are typically 50%80% of a distribution center's total cost of operations. Auto-ID will enable automation or semiautomation of key operations (such as picking, put-away, replenishment, and shipping), thereby reducing labor requirements. A recent study [5] estimates that labor could be reduced up to 40%, depending on the number of handling points and the degree of technology deployment. Moreover, the accuracy of operation, asset utilization, and inventory carrying costs can be improved through the tracking and monitoring functions associated with Auto-ID.
On the retail side, the moment an item is removed from the shelf, the message requesting replenishment is automatically sent out. The increased velocity in replenishment and the reduced lead time variability help reduce inventory levels, prevent stockouts of certain items, and cut costs. The Gillette Company reports its retail sales would be 15% higher if store shelves were always stocked with its products [8]. The potential for theft by employees and customers could also be decreased through Auto-ID. Big department stores could improve their reshelving by tracking items at various locations in their facilities. Gap, Inc. recently tagged one of its suburban Atlanta stores, enabling store personnel to create a computer snapshot of the location of every pair of boot-cut women's indigo jeans in the store [11]. This effort decreased the potential for lost sales due to the sales staff being unable to direct customers to the right products in a particular style, color, or size. Retailers would also benefit from reduced labor cost, improved asset utilization, and better order fill rate from the use of Auto-ID.
The recall and return of defective products is common in supply chain operations. Using Auto-ID, the path followed by all products could be traced back and matched with that of the defective product(s), thus supporting the recall process. Similarly, unsellable items from different retail stores could be tracked and consolidated with less effort for return to the manufacturer.
An essential aspect of any supply chain operation is monitoring and taking corrective measures in response to changing conditions. With enhanced visibility and near real-time information processing provided by Auto-ID, these SCM functions are likely to experience remarkable improvement in the future. The everyday lives of customers are likely to change as well through, say, instant checkout and payment at stores and automatic preparation of grocery lists by smart refrigerators as their contents are depleted [2].
The Supply-Chain Operations Reference Model [12] (Version 5.0) lists a number of level-one SCM performance metrics (see the table). We expect Auto-ID technology will improve all of them, except production flexibility, which is typically influenced by the design of production processes and is a core competency of lean manufacturing.
Auto-ID can succeed if it is supported by a software system capable of consolidating the large amount of data captured by the wireless readers.
Adopting Auto-ID
Using Auto-ID in SCM still requires the resolution of a number of issues:
Cost of tags. Research indicates that the cost of an Auto-ID tag is about 40 times greater than the cost of a bar code and that Auto-ID readers can even cost several thousand dollars each. This means that a significant upfront financial investment is required by retailers and manufacturers. The price of an Auto-ID tag is reported to be about 50 cents, but large-scale adoption of Auto-ID is likely to be feasible only if the price of a tag is less than five cents each [10].
Item-level vs. pallet-level tagging. The benefits of wireless product identification without manual handling at the item level were elaborated in [7]. Pallet-level tagging is an alternative to item-level tagging. How can an organization determine whether item- or pallet-level tagging would be better for a particular commodity? Item-level tagging may be better for high-priced items (such as designer clothing and electronics), whereas pallet-level tagging would be better for inexpensive everyday items (such as groceries and household goods).
Frequency. The frequency of the wireless communications between Auto-ID tags and tag readers must be reserved. Would a ubiquitous frequency be sufficient for reception? Would separate bands of frequencies be better for certain product types?
Data management and data mining. Increasing amounts of data collected from each individual item, as well as the numbers of items, prompt several issues regarding the management of the data. Efficient methods for analyzing and interpreting it must still be devised in order to make the technology deliver its promised benefits.
Privacy. Auto-ID promises to track each product more closely from manufacture to consumption. It also promises to make available various data items related to individual consumer likes and dislikes, as well as consumer buying behavior, to retailers. Consumers may strongly dislike this capability if they feel it violates their privacy while failing to provide them sufficient benefits [9]. If, for instance, tags on individual items are not deactivated after a sale is completed, the manufacturer would be able to track how the consumer handles the goods under consideration. While consumers may tolerate this lack of privacy for certain products, they may be less tolerant for others.
Big bang or phased implementation. The big-bang approach is unlikely to be used to implement this technology. For example, although trial runs have been conducted in Tulsa, OK, [1] by the Auto-ID Center, it is not clear which part of the supply chain should be automated first. At the retail end, it is unknown whether the technology will produce equally low cost-benefit ratios for all retail outlets of, say, a chain store, irrespective of location. However, retail giant Wal-Mart currently requires all its major suppliers to adopt Auto-ID technology [4], likely motivating general adoption of the technology.
Integration with enterprise software. Auto-ID can succeed if it is supported by a software system capable of consolidating the large amount of data captured by the wireless readers. The data must be filtered and smoothed, made centrally manageable, and connected to enterprise Web services or Java environments where it can be used by existing SCM software. Such integration usually involves development of customized middleware using application programming interfaces. Several database management vendors (such as ConnecTerra and Provia Software) are developing software solutions (such as RFTagAware and RFIDware) to handle these operations, making custom programming unnecessary, speeding enterprise implementation of Auto-ID, and easing the linking of Auto-ID to existing Web-based SCM software.
Conclusion
The only way to address the problems challenging implementation of Auto-ID technology is through a more inclusive collaborative partnership among the various players in the supply chain. The Auto-ID Center, formed by a consortium of 88 companies and academic institutions, is a positive step in this direction [8]. Auto-ID is a promising technology with the potential to revolutionize all facets of the supply chain; a well-organized, focused effort can help it quickly fulfill that promise.
References
1. Albano, S. and Engels, D. Auto-ID Center Field Trial: Phase I Summary. Working paper. Auto-ID Labs; www.autoidlabs.org/whitepapers/MIT-AUTOID-TR-006.pdf.
2. Albrecht, K. Supermarket cards: Tip of the retail surveillance iceberg. Denver University Law Review (June 2002).
3. Atock, C. Where's my stuff? Manufacturing Engineer (Apr. 2003), 2427.
4. Boyle, M. Wal-Mart keeps the change. Fortune (Nov. 10, 2003).
5. Chappell, G., Durdan, D., Gilbert, G., Ginsburg, L., Smith, J., and Tobolski, J. Auto-ID on Delivery: The Value of Auto-ID Technology in the Retail Supply Chain. Working paper. Auto-ID Labs; www.autoidlabs.org/whitepapers/ACN-AUTOID-BC-004.pdf.
6. Finkenzeller, K. RFID Handbook Fundamentals and Applications in Contactless Smart Cards and Identification, 2nd Edition. John Wiley & Sons, Chichester, U.K., 1999.
7. Karkkainen, M. and Holmstrom, J. Wireless product identification: Enabler for handling efficiency, customization, and information sharing. Supply Chain Management: An International Journal 7, 4 (2002), 242252.
8. Keenan, F. If supermarket shelves could talk. BusinessWeek (Mar. 31, 2003).
9. Maselli, J. Privacy group focuses on RFID. RFID Journal (Aug. 26, 2003); rfidjournal.com/article/articleview/547/1/1/.
10. Mayfield, K. Radio ID tags: Beyond bar codes. Wired News (May 20, 2002); www.wired.com/news/print/0,1294,52343,00.html.
11. Schoenberger, C. RFID: The Internet of things. Forbes (Mar. 18, 2002); www.mindfully.org/Technology/RFID-Things-Forbes18mar02.htm.
12. Supply-Chain Operations Reference Model: Overview of SCOR, Version 5.0; www.supply-chain.org/slides/SCOR5.0OverviewBooklet.pdf.
Figures
Figure 1. Layout of a grocery store equipped with Auto-ID readers at point-of-sale, storage, and receiving locations.
©2005 ACM 0001-0782/05/0800 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2005 ACM, Inc.