RFID: cost, benefit or a bit opf both?

RFID, radio frequency identification, is currently the subject of more reports, White Papers, speculation and, yes, hype and half-truths than virtually anything else business-technical, even the web and e-business. With good reason: it will be transformational, and although the technologies and standards are still evolving, they're coming together very fast, costs are falling – and the big retailers and defence procurement agencies in particular, sensing significant efficiencies and cost savings, are issuing mandates. The clouds are gathering and that tends to galvanise serious attention, good and bad. Time then to get a measured appraisal of RFID facts: what it is, what's possible and what's not, the key issues and considerations, and some of the technical and system implications. Just as important, time also to get a handle on the potential impacts and business opportunities, the likely cost/benefits and thus what we should all be doing about it and when. So what is RFID and why is it causing such a stir? Actually, it's anything but new, with its origins – depending on who you talk to – back in the Second World War. Certainly, RFID technology is used in all sorts of applications, from car key remotes to access cards, ski lift passes and the like for authentication. Marathon runners are increasingly issued with RFID tags, and likewise livestock for identification and tracking. Similarly, there are implementations in some of the more advanced libraries, in hospitals and on airport baggage handling systems. And in manufacturing, they have been used for years, for example on flexible production lines and spray booths in the automotive sector and on safety systems in petrochemicals. But things are changing fast, and what have been relatively specialised niche applications are about to be overshadowed as mainstream RFID goes very large and very commercial. You can think of RFID today as advanced barcoding – but remote-sensing, much higher speed, much higher data storage, interactive, potentially intelligent, read/write and very robust. Key differences to note include its ability to uniquely identify any item (not just product type), and provide information about it – what it is, potentially the audit trail of manufacture, storage, transportation, maintenance – and to do so very fast and without 'line of sight' scanning. Fast and convenient Consider cars going through a toll booth: the old way involves time and hassle in stopping, queuing, finding the change and so on; with RFID you drive through at 40mph and you're debited automatically. It's not difficult to see the attraction, or to apply that metaphor to the movement of materials through your supply chain. And it's in this context that RFID is about to explode as the likes of Wal-Mart, Tesco and the US Department of Defense (DoD) drive massive cost savings in inventory and labour as well as boosted sales from reduced out-of-stock conditions, from their considerable investments in the technology. That's not to say RFID will replace barcodes, or indeed other sensing systems with which we're all familiar: it won't for years to come. Neither is it likely to be used over the next decade for individual items on anything but higher value goods where a return can be demonstrated. But at the case, carton and pallet level, the technical attributes of RFID – bulk, box-closed, automatic scanning and recognition – make it compelling right now as a new and far better way of automating and cranking up materials movements monitoring. Likewise, no-one is saying that supply chain improvements represent the only business driver: developments with the kinds of applications already referred to will continue and gather pace. But with the top 14 CPG (consumer packaged goods) and retail companies in the world – the likes of Nestle and Proctor & Gamble – producing around 555 billion items per year, it's by far and away the greatest in terms of potential volumes, and global business impact, spend and therefore also challenge. And manufacturers are going to be in the thick of it. Stepping back for a moment, the technology relies on RFID 'tags' (also referred to as smart labels and radio barcodes) which are attached to whatever needs to be identified and monitored. There's no magic: they comprise a micro-miniature chip, aerial and substrate, and already come in all sorts of shapes, sizes and formats to suit the detail of the multitude of packages, contents and industry sectors. The rest of the equation is then the antenna to sense the tags' presence, the readers (fixed or mobile), and the middleware and host applications and systems that handle and harness the information, and the integration into your back-end systems. Looking at the tag details is where it starts to get interesting, because there are choices. As Simon Holloway, manufacturing and engineering specialist at Microsoft, explains, essentially they break down into two main types: active (self-powered) and passive (without power). "Passive tags have to be pinged by radio waves from the reader/antenna to reveal their presence and/or data, while active tags continuously broadcast." Key distinguishing features are that active tags' ranges are far greater than passive – up to several hundred metres instead of one to five metres – but that they're also a lot more expensive. Ed Cowley, director of the UK's RFID Networking Forum (www.rfidforum.com), makes the point that evolution is carrying on apace, with, for example, even that broad distinction blurring as hybrids conferring different cost/benefit balances are developed. He also observes that both main tag families are available as 'item-attendant' devices, to which large amounts of information – like shelf life, batch number, maintenance history, dwell time at an operator station and so on – can be written. "You can think of those as providing a virtual page of A4," he says, "but at the other end of the spectrum, there are the 'number plate', simple radio barcode devices that carry component ID only. It's these that are favoured by the retail industry." Why? Because they're perfectly adequate for improvements like being able to track and recall individual goods automatically, and in UHF passive tag form, they're also cheap, small, relatively reliable, simple to apply and fast – scanning at up to 800 units per second, depending on power. And with virtually nothing costing more than £3 in the retail sector, all that has been precisely the thrust of the smart development money. They've also very nearly achieved standards recognition ahead of proliferation: and while there are still two main standards initiatives – ISO (broad-based and inclusive) and the big one, EPC (Electronic Product Code) Global (passive number plate), stemming from the US Auto ID Centre with its roots in retail and CPG – there is some convergence and no doubt where the initial dominance will be. EPC Global's only outstanding issues remain frequency and power – the points being to maximise range and read rates, and to get world-wide uniformity where currently Europe and the US remain at odds. Why now? Which answers the question, 'Why now?'. RFID is maturing very fast; prices are falling; and the number of suppliers, and those with at least some implementation experience, is rising – although it's worth being aware that suppliers lag the technology, being still somewhat fragmented and awash with proprietary solutions. Nevertheless, the big boys see the timing as imminent and are calling some pretty demanding shots of their suppliers. Wal-Mart and the DoD, for example, issued their compliance requirements in the summer of last year. The former is now at Phase One trials, with Gillette, HP, Johnson & Johnson, Kimberly-Clark, Kraft, Nestle, Proctor & Gamble and Unilever testing EPC systems at Wal-Mart superstores and its Dallas Fort Worth regional distribution centre, ahead of the January 2005 deadline for the top 100 suppliers to be tagging in-bound pallets and cases. They're not alone: Tesco went public on its RFID requirements last autumn and its trials are underway; Marks & Spencer stepped up to the mark in January this year; likewise the German Metro store group; and the Target group declared its position in February. As Cowley says: "We're not yet at the holy grail of 5 cents a tag that the retail sector wants, but this year will be a critical foundation year for RFID. It's the start of a move from fragmented to consolidated; from expensive and individual to cheap and universal; from doesn't quite work to works… It will definitely take off, and it's going to be massive." Observers are expecting exponential growth starting in the next couple of years, and manufacturers will follow the curve. As Tier One producers for the major groups look to their suppliers to get on board, and they do the same in turn to their suppliers, this is going to ripple through vast swathes of manufacturing right across industry sectors. The RFID requirements may well be different in detail, but come they will. So what of the costs and implications? Passive tags are currently in the 20 cents to $1 range, while active devices are around 10 times that figure, at nearer $5—10. It's not difficult to do the maths against your output to see the impact, but that's just the start. On top of that you'll need readers, which cost in the range $2,000 to $6,000, and probably also RFID-enabled hand-helds that also handle barcodes, costing $200—$1,000 apiece. Beyond that you'll also need tagging systems, perhaps the 'print on demand' type currently emerging that can be embedded in production and packaging lines. And you'll require systems and the infrastructure, possibly including wireless networks to facilitate implementation, for machine-to-machine communications and to record and implant the required information onto the tags. So there will be additional requirements of our IT – including also the host applications and the systems they run on. As Holloway says: "What's important there is what sits on the server. What are the key movements you need to update to and from your ERP system? How many of them are there? How real-time do you need them to be? Almost any system will have to filter the noise from the exceptions – and be configured so that it does something about those events: it will need to push actions and alerts through the ERP system, for example." And the upshot of that? "You might have to buy a larger server, or you may not depending on what you have and what you want to do: you may be consolidating anyway so you could use the spare resource. On the other hand, you may need newer server technology to handle the volumes and nature of data, the transactions, workflow and interfaces." Equally, you might want to consider updating or even changing your main systems if the changes that RFID forces, or indeed enables (see later), make reviving that supply chain project you'd put on ice viable. Certainly, the IT industry is expecting great things to come from RFID as industry conforms, but also, importantly, looks for its own business benefits. Stuart Facey, senior vice president and general manager of TrenStar, the third party asset management company that's been running RFID for its and its clients' businesses for the last couple of years, confirms that all its systems are on Microsoft technologies. "The hand-helds use Windows CE so they can run with different manufacturers' equipment. Our web-based development tools are all Microsoft... We're looking at BizTalk server now for the back-end data capture into ERP systems." And we mustn't underestimate this part of the equation. As Keith Bell, head of supply chain solutions at SAP, says: "Data has to be visible for it to be any use… So apart from read/write rates and distances, the technical challenges are around management and dissemination of the collected information [and] integrating that into the back-end systems." No-one is saying that integration per se is any more difficult than with any other project or system, but it's the volume of data and the fact that it's real-time. We're bound to see increasing roles for automated e-business systems and portal-based information pushing and pulling. And the IT industry is already working on systems to manage information sharing and security. Meanwhile, at the operational level, Holloway's view is that in a sense, nothing changes: this isn't, after all, a million miles from any shop floor data collection (SFDC) exercise. "Your ERP system needs to be connected to the shop floor, warehouse, wherever managing material movements and operations better could make a difference. So you need to look at movements that are relevant to ERP, those that could usefully be automated, with a view to providing the data in real time. Then the scale, frequency and touch points of data increase, and your system, which may not be set up for the real-time data rates, will require attention. The difference is that now there's more motivation to do it." Make it pay Just so. Because there's no escaping the conclusion that getting the benefits that the big boys want is going to cost the rest of manufacturing dear. So all the more reason to play our cards right, not just in terms of negotiating deals, but looking for benefits. And a key point to take away is that there are plenty so that even as we're pushed to provide RFID tagging, we can also harness it for our own ends, rather than simply sticking RFID tags on products as they leave packaging. So what are they? Ken Douglas, technical director at BP, says: "We have a very simple model for RFID. Imagine that you could have an expert with a clipboard alongside every asset, and that you could interrogate that expert at any time – whether the asset was on a production line, in the supply chain or in a marshalling yard – and you could pull information about its location and status… There may be problems where you could gain half a percent – and that could make a very big difference." Holloway agrees: "Organisations need to see that the technology can have positive implications where moving things is concerned, cutting costs, reducing inventories and improving stock accuracy and efficiencies. It can enable improved shop floor manufacturing processes, and internal as well as external supply chain processes, taking Lean thinking and Six Sigma, for example, further." It can provide visibility of the black hole of production so that last minute, late or changing orders can be pulled through, and manufacturers can be more responsive. And he reminds us that RFID enables items not just to be tracked spatially, but in terms of what's happening to them – their status – giving a new lease of life to flexible manufacturing, automation and production analysis. There's also the potential for better customer service – including self-service via web-based tracking extensions – and for improving forecasting through better information right from the point of use. Similarly, RFID can help in warehouses, managing first-in, first-out logistics, mixed consignments, products that age and so on. And there are opportunities in terms of enhancing product and operational safety, tackling counterfeiting and thus brand image issues, and cutting losses through pilferage. Those are on top of the main supply chain efficiency improvements and methods of managing recalls. It's worth noting that to achieve some of these, the tag type and technology may well not be that enshrined in the EPC Global standard. Some estimates are that, longer term, up to 80% of manufacturing's applications will require higher data storage, dynamic devices. But with increasing momentum will come more acceptance and understanding, development and again, falling prices. The leading GCI ITAG Group ranks likely priorities specifically for manufacturers as improving inventory management, warehouse management, out-of-stock avoidance and transportation and logistics management. It then suggests better track and trace, demand planning and order fulfilment, followed by theft avoidance, asset tracking, enabling VMI (vendor managed inventory), providing better customer information, CPFR (collaborative planning, forecasting and replenishment), procurement and returns management. Real value comes from improving the big picture – streamlining the whole supply chain, from manufacturers' suppliers right through to their customers – cutting costs, improving flexibility and responsiveness to customers' requirements all the way. In this sense, RFID should be seen as part of the technology revolution that absolutely includes web technologies, and supply chain planning, management and execution/fulfilment. Microsoft's pilot project with Dutch mid-range snacks and crisps manufacturer KiMs in Holland makes the point eloquently: the success of that trial is in its delivery of what KiMs knew it needed: visibility into its supply chain, with automated monitoring and alerts on the supplier side and finished goods tracking for internal, warehouse and partner and customer use. Cost/benefits We can't close without at least attempting to examine ROI and viability. So far, it's difficult to be precise. Bjarne Schon, business development manager at Microsoft Business Solutions, points out that most projects are at the pilot stage or just starting to deliver hard benefits. "There's not enough data yet. KiMs, for example, is just starting to build the business case now. We can say that the benefits in that kind of implementation are significant in inventory control and visibility." And the cost? "We're not talking millions of dollars of investment," he says. "Tags are the single biggest cost; the hardware and software elements are limited to extensions of the ERP system. It's certainly not hundreds of thousands of dollars." Which is less than helpful. Although he has nothing but good to say of the project, KiMs CEO Jørn Tolstrup Rohde can only agree that as yet it's difficult to assess cost/benefits. He says simply: "The prospect that RFID can help us dramatically increase our ability to read and anticipate our inventory flow is compelling." Few doubt that the cost/benefit equation is going to be a tough one, but as Holloway says: "If you're in CPG or food and drink, you need to be thinking now anyway because, when it comes, your customers will want you to do it quickly. You need to work out what benefits you can get out of it by getting involved – otherwise you won't get any." Consensus advice: get started, but specifically do so by being pragmatic, collaborative and challenging: difficult to square that circle I know, but what you're looking for are applications that are going to bring business benefit. Rope in other parties (internally and externally) because the benefit is likely to be plural, and gain experience in the areas that are problematic. Other than that, "People need to see immediate benefits," observes Holloway, "so receipt into the warehouse, or from the warehouse to the customer distribution centre are good, practical starting points, not least because there are gates so it works for antennae and reader positions. But equally, you might want to explore using RFID hand-helds to see WIP movements on the shop floor." The point is, it's viable now, and you need to think 'By doing this, what else could it let me do'