Wal-Mart suppliers shoulder burden of RFID; ONS and PML go begging?

11 November, 2003

category: RFID

A ComputerWorld article from November 10, notes that a supplier to Wal-Mart will have difficulty justifying the cost of RFID implementation, while Wal-Mart reaps the benefit. The article then fluffs the issue of how the Wal-Mart 100, Wal-Mart’s top suppliers, will have consider their internal EPC implementation architecture, irrespective of Wal-Mart’s decisions.

According to the article, the analysts and consultants seem to agree regarding who benefits from RFID versus who pays for it. Despite the article’s headline (“Wal-Mart suppliers shoulder burden of daunting RFID effort”), the analysts and consultants seem to be split on the issue of how to make the internal justification for RFID for the Wal-Mart 100:

“Right now, the benefits are primarily for Wal-Mart, and the costs are the responsibility of the suppliers,” said Kara Romanow, an analyst at AMR Research Inc. in Boston. Romanow estimated start-up costs at $13 million to $23 million for a supplier that ships 50 million tagged containers per year. Costs include RFID tags, readers, system integration and changes to supply chain applications, she said.

And then there’s …

But some analysts and consultants said the only way suppliers will be able to justify the expense is to do the necessary supply chain and business process re-engineering that ultimately will help them to take advantage of the more granular and accurate information.

That will take time. Jonathan Loretto, a Toronto-based consultant specializing in RFID at Cap Gemini Ernst & Young, said the “Wal-Mart 100” face 18 months of hard work to comply with the retailer’s requirements. He estimated first-year costs at $15 million to $18 million per supplier for the hardware, software, other systems and labor.

After a quick discussion of the classes of tags that Wal-Mart will support (Class 0 and Class 1 as specified by the Auto-ID Center), the article points to the issue brought up in an a post on TWF – the DoD’s support of ISO standards rather than EPCglobal’s specification.

And then we come to the final two paragraphs of the article, each of which gets to me for a different reason:

Analysts, however, said they don’t expect the ISO vs. EPCglobal issue to drag on long term. “They will eventually merge, and people shouldn’t worry about that now,” said AMR’s Romanow.

Sorry, but if you’re a product line manager, with a responsibility to satisfy Wal-Mart, your other commercial customers and maybe the DoD, you’re also about to go to your CFO with a plan that’s going to cost $15 to $30 Million over the course of 18 to 24 months. The CFO asks you what portion of that expenditure is actually going to meet the normal 5-year depreciation timeline, and how much of the implementation required to meet the January 1, 2005 deadline will have to be ripped out early and replaced. You really don’t want to tell him that, in the spirit of standards harmonization, $5 – $10 M worth of equipment and software may have to be thrown away in order to meet the new merged standard on January 1, 2006. Now, do you?

The final paragraph reveals one of the more troublesome sides of the Wal-Mart RFID initiative.

Another potential burden that has been lifted from suppliers’ shoulders, at least for now, is the need to use the Physical Markup Language to tag product information and an Object Name Service database server, according to analysts and consultants familiar with Wal-Mart’s plans.

This assessment bespeaks a misunderstanding, or just not enough of the “thinking through”. It’s become clear that Wal-Mart itself will NOT be implementing RFID internally based on the reference architecture defined by the Auto-ID center. Rather, they’ll be integrating their RFID reader networks with the Wal-Mart proprietary ERP and internal management systems, which already handle data from other forms of automatic identification technologies (e.g., bar code scanners and human input). Wal-Mart won’t rely on the Object Name Service (ONS), nor will they store RFID record data in the standard form, Physical Markup Language (PML) for purposes of generalizable access. Once the pallet or carton reaches a Wal-Mart distribution center, Wal-Mart’s legendary IT infrastructure will handle it the way Wal-Mart’s IT organization wants to handle it.

But, that leaves the Wal-Mart 100 suppliers with some big questions regarding architecture. If I (the Wal-Mart supplier) am to use RFID/EPC technology internally, justifying my expenditure on the basis of re-engineering and optimizing my supply chain, to what ERP systems do I integrate? Will my company need to expose RFID record data or go outside the corporate network to request RFID data from my suppliers? Do I architect my RFID infrastructure on the basis of the standard architecture, assuming that I’ll need to provide support not only to Wal-Mart but to other commercial customers?

As long as the company asking these questions is not the endpoint in the supply chain, this architectural consideration of proprietary vs “standard” ONS and PML will lead to major turmoil over the course of the next few months. The Wal-Mart 100 will be deciding whether to “bolt on” RFID to their existing ERP and fulfillment systems, re-architect their proprietary systems, or implement “standard” systems even though the standards are still in flux. Wal-Mart’s decision not to use ONS and PML internally does little to address the question for the Wal-Mart 100.

Related articles: