Going for (Not So) Broke: The True Cost of RFID
The cost of RFID implementation has been over-hyped and exaggerated. Early cost predictions had vendors seeing green, but in reality, an RFID implementation need not bust your whole technology budget. Find out the true costs of taking the RFID plunge.
Early predictions on the cost of implementing RFID horrified CFOs. Each of Wal-Mart's top suppliers would spend between $1 million and $500 million on their first RFID pilots in 2004. These figures became keystones for the rest of industry, which recoiled at the idea of sky-high costs.
These early predictions proved false. Instead, companies implementing RFID to comply with Wal-Mart's mandate actually spent only a fraction of those amounts—approximately $500,000 each. Outside of those Wal-Mart vendors, the costs are even lower.
So, just what does an implementation cost, and where does the money go? We queried a panel of RFID implementers and integrators, who combined have conducted more than 100 implementations—more than 40 of which were with Wal-Mart vendors participating in the RFID pilot.
The new oft-quoted estimate for RFID implementation is $500,000—far less than the predicted millions. According to our panel of experts, however, that figure is still high.
"I've never seen a vendor give a half-million-dollar quote," says Steve Halliday, president of High Tech Aid, a Pittsburgh, Pa.-based consultancy specializing in automated identification and data collection. "Quotes of $200,000 to $300,000 are more common."
Implementation costs can be as low as $15,000 for small vendors, and up to $400,000 for larger vendors, depending on the size and number of facilities, says Joseph Leone, who implemented RFID at Wal-Mart, and is now a consultant with RFID Global Solution Inc., Mount Airy, Md.
As a general rule, most companies taking a conservative approach spend less than $250,000 per facility—including those serving Wal-Mart stores and distribution centers, says Leone.
Because those adopters are taking pilot approaches to implementation, they experience lower costs, while the multi-million-dollar figures are based on full implementation at large vendors. Ultimately, total cost depends on the level of implementation, and on whether the implementation is a single- or multi-site deployment.
The Three C's
Of course, not all implementations are created equal when it comes to depth of integration, either. Three common levels of implementation are:
- Compliance: slap-and-ship only.
- Conservative: light RFID capabilities within a limited geography.
- Committed: implementation at all distribution centers.
Of these "three C's," defined by Yankee Research senior analyst Michael Dominy, early adopters tend to fall somewhere between the compliance and conservative approach, into what Leone calls a "bottom up" approach. Long-term cost advantages make the bottom-up approach popular.
"Any estimate of cost or ROI, however, will only be somewhat accurate because every supply chain is unique. Uncertainties make those estimates impossible to pinpoint," says Leone. He recommends a "thread approach"—selecting several products and/or processes to implement against.
"Companies selecting a 'top-down' implementation may find they could have gotten the same value with 500 RFID readers versus 5,000, or that they've spent $10 million only to save $2 million," Leone says. "When you work from the bottom up, and start to see diminishing returns, you know you are nearing your goal and do not need to go much further."
Large vendors committing to RFID are also practicing limited implementations. Companies that implemented RFID in 2004 will tag between 6 percent and 10 percent of their merchandise with RFID this year, according to a survey of 300 top suppliers by the NRF Foundation (a nonprofit education and research arm of the National Retail Federation), and BearingPoint Consulting. Our panel estimates they tagged less than 4 percent during the pilot.
But does limited implementation also limit the return on investment? Or limit competitive advantage?
Unlikely in the short term, as most adopters have taken the cautious approach.
"The adage, 'The more you invest the more you get back,' applies later on in the process," says Jeff Richards, president and CEO of R4 Global Services, an RFID solutions provider based in San Francisco. "Most businesses are pacing implementation to match the needs of their customers at this point. If they're implementing RFID on their own without a mandate, it's because they've identified a significant benefit to their business."
While the approach is cautious, these companies are not limiting themselves to slap-and-ship compliance implementation—rather, they are starting small, and staying focused on particular product lines and specific business benefits.
Where the Costs Come From
Implementation costs can be broken down into four broad areas:
- Hardware: tags, readers, and transponders, as well as any relevant capital equipment.
- Intangibles: consulting, analysis, and services. These costs include business process analysis, site surveys and reliability studies, training and implementation, research and development, and troubleshooting.
- Changes to existing supply chain applications: purchasing middleware, upgrading enterprise systems, or purchasing RFID software.
- Storage and analytics of large volumes of data.
Analysts predict these costs will break down over the long term as follows: hardware, 44 percent of total implementation cost; integration, 22 percent; changes to supply chain applications, 22 percent; and data storage and analysis, 13 percent.
But early returns tell a different story. In the short term, the cost breakdown is quite different, experts say.
The intangibles of consulting, analysis, and services are the largest expense over the long term—typically more than half the project cost, and up to 80 percent in some cases.
The good news? This high cost is front-loaded, not ongoing. The need for intense planning before an actual implementation takes place is what piles on the consulting fees. One food manufacturer in Wal-Mart's top 100 vendors, for example, spent seven months in business analysis before spending a dime on technology, notes Richards, who was involved in the implementation.
"After a while, the majority of expenses come from tags and labels, but early on, most companies spend a lot on services to develop the IT architecture and think through the logic," he explains. "Before implementing with the food manufacturer, we went through detailed requirements and a very specific implementation plan, as well as the system design and technology selection."
"The aim is to spend 90 to 120 days on the implementation itself, then reach a point at which the solution is just being replicated," explains Leone.
Hardware, specifically tags and labels, is predicted as the highest long-term cost, because it will be so prevalent once lowered tag prices permit item-level tagging. But in these early stages, hardware accounts for less than 10 percent of total cost. Early adopters are tagging at the case and pallet level, and only on fractions of their product throughput.
The much-anticipated drop in equipment costs is well underway. Passive tags are now available as inexpensively as 20 cents per tag in bulk. High-end readers, now priced at roughly $1,000, are expected to drop to $300 during the next two years. Tag printers, however, are still in the several-thousand-dollar range. But even at today's costs, equipment is a secondary budget item.
No Small Change
Changes to existing supply chain applications are the second-largest budget item overall.
"Middleware accounts for between 25 percent and 30 percent of the total cost," says Steve Brown, executive vice president of Acsis Inc., Marlton, N.J., which performed 22 RFID implementations in 2004, 20 of which were Wal-Mart or DoD sites.
"Integration with enterprise applications, and data filtering—selecting and mapping specific data among the gigabytes of data that RFID produces—are included in middleware costs," he says.
Upgrades to enterprise applications are not necessary in all cases, but "those companies that are going beyond compliance will bear upgrade costs to take advantage of back-end data," explains Brown. "RFID can track individual cases; most enterprise systems don't track that way."
Companies will not likely need to budget for new enterprise systems, however, as vendors are beginning to offer RFID functionality in system upgrades and add-ons.
Storage and analytics expenses depend largely on the adopter's business goals and the early system configuration. So far, less than one third of adopters have had to upgrade their existing servers, according to our panel.
This is where the early investment in consulting and middleware pays off, as data filtering and system configuration are key to keeping manageable levels of data, as well as finding the right data.
"The issue comes down to what information you're trying to capture. Companies starting simply, with smart readers and filtering systems, may get away with not changing their current IT system," notes High Tech Aid's Halliday. "Companies that buy inexpensive readers may get multiple readings from one tag, which could overwhelm the solution if there's no filtering system in place."
Starting with minimal data is a good idea, recommends Brown. "The data is not overwhelming if the filtering technology is working well. RFID generates more granular data, but it can be limited to data for a specific business event, such as seeing whether a product has shipped," he says.
Beyond the initial heavy costs of consulting and planning, adopters can expect their largest implementation expense to come from tags and labels, and whatever system and IT upgrades are necessary for a broader enterprise-wide adoption.
Ongoing expenses also include third-party service provider fees, such as EPCglobal's sales-based subscription fee—$75,000 for companies with annual sales of $1 billion to $10 billion—and the Uniform Code Council's e-commerce provider UCCNet—$40,000 for companies with annual sales of $1 billion to $3 billion.
EPCglobal membership is necessary for full access to standards, and both EPCglobal and UCC membership are necessary to register uniform product codes. Companies need to determine which services they need, and what the ongoing costs will be, when calculating their RFID budgets.
The overarching conclusion is that RFID is not the budget-buster originally predicted. Early adopters have paced themselves to specific business needs and the needs of their customers, versus investing heavily to gain competitive edge. Planning and consultation are the highest short-term costs, but are necessary to keep overall project costs down.
"Start small and stay focused," says Leone. "'Overimplementing' brings no value. But there is value in picking a thread of your operation—whether inside your four walls or in your supply chain—and finding out how RFID works for you. That's the best way to figure out where to invest your money.