Building your RFID business and IT

From a business perspective, RFID should be treated like any other technology: if there's a problem it can solve or specific value it can add that matches the agreed business direction and proposed projects, then a business case is made and it goes on the agenda. Then it's considered in the round against the other projects clamouring for attention and resource. But there's one small snag: that approach presupposes that organisations know where RFID can make a difference against existing technologies in existing projects, and that they also have a good picture of the kinds and scale of opportunities it opens up: the art of the possible. And for many, that remains unlikely: most organisations we speak to are aware of RFID's potential in only one or two of its potential applications, so there is room for some education first. So try the following: if you could use RFID to automatically track back into your ERP system exactly what's going on where, you could not only better fulfil customer service, but also potentially improve internal efficiencies, material and logistic costs and robustness of processes, including some of those out towards, or indeed into, the supply chain. You could also find out exactly how much it's really costing you to make and deliver something – and you'd have a great tool and catalyst for process improvement. Such a system doesn't come free, but what's the business case for you, compared, for example, to using barcode systems? And then if you thought about asset management – location, maintenance history, appropriate skills, permissions to work – and were told that RFID could strip out time, admin and waste, how does that start to stack up? Without wishing to teach egg sucking, RFID as an advanced mechanism for automatically identifying and tracking things and people, has many different applications with the capacity to significantly shift some existing limits. The detail will always be different for everyone: for example, in a supply chain context it's not necessary or even useful to track every movement of every item. As Simon Holloway, manufacturing industry solutions architect at Microsoft, says: "You look at the business issues – how much can you save? Identify the highest priority, build a proof of concept, perhaps a trial, then a pilot and finally deploy." Unless, of course, the business issue is compulsion: regulatory requirements, or a supply chain initiative like Wal-Mart's. Whatever the initial impetus, the first project is unlikely to be the last: once RFID has a foothold, and the company has got to grips with tags, antennae, tuning and other hardware issues – which in the end are not the big deal – other applications for the technology are likely to reveal themselves. Some of those will be dependent on emerging global standards and initiatives; others can and are being implemented right now. So let's get that straight first. If, for example, your RFID project involves exchanging data with partners in the supply chain, you'll need meaningful data. And that means sorting out internal and external data synchronisation. Teresa Jones, senior researcher with analyst Butler Group, says: "Sorting out your data is the first step you should be taking. Don't even think about RFID yet if you have invalid cross references, duplication and so on. Synchronise different databases if you have them, and ensure you are tracking the right part numbers internally. It's not very exciting but the reality is that everyone has inaccurate data." Sort out data first? Then externally, useful exchange of data in the supply chain increasingly means adopting GDS (Global Data Synchronisation) – which involves a unique identifier per item or collection of items by all parties in any transaction. Working with this is just as important. Indeed, Nigel Montgomery, director with analyst AMR Research, goes as far as saying: "You should be doing GDS regardless of whether you do RFID or not." All that said though, RFID and GDS are not inseparable. You can get business benefit from RFID without anything to do with GDS, and Montgomery believes that most manufacturing companies can deal with their internal data issues and successfully embark on RFID "without a massive investment." Montgomery gives the simple example of one supermarket chain using RFID tags on delivery trucks. Instead of relying on a security guard at the gate to call the dock – on time, and in the right order – to let staff know a truck has arrived, an RFID tag can signal that information automatically, helping to cut the goods inward bottleneck. There are many similar examples relying on the fact of automatically detected, and potentially very detailed, arrival of materials and goods. No need for GDS, and no need for complex infrastructure and integration: at least, not at first. Applications like these allow manufacturers not only to try the technologies involved, and gain experience without going the whole supply chain GDS route, but to work at achieving and proving business benefit early on. Automatic ID And as Montgomery says, actually, the standards and technology are not interesting from a business perspective: you merely have to be aware of the current physical and standards-based limits and costs – and the former are essentially solved case-by-case using technicians, not SAP consultants. "It is going to take you 18–24 months to roll out your project anyway, so start looking at the business case now to see where RFID can make sense," he advises. For most of us, it makes most sense at this stage to investigate the potential for existing projects where RFID could be a better fit, and to re-examine business operations and service offerings in light of what can now, or will soon, be achieved. Within the factory walls, for example, Stephen Proud, RFID practice lead in this country for consultancy Accenture, cites one company using tags to identify goods once they've been received to the plant and, by fitting antennae to forklifts and tags to the floors, also what's being moved, where and when automatically. "That has led to a 20–30% improvement in forklift utilisation," he declares. And the benefits haven't stopped there. By putting tags on containers, the company also tracks and controls which raw materials go where in the manufacturing processes themselves. If a forklift driver attempts to tip a container into the wrong hopper, the truck locks and is prevented from operating. It's already saved the cost and time of stopping a process, and cleaning down or blending away the problem. Clearly, that kind of project is eminently transferable. Another area ripe for improvement is returnables: whether products or containers – and what about movable assets like production tools? It's not just about tracking and automating processes either. For example, RFID tags embedded into retreaded tyres by one manufacturer in India are putting a stop to tyres being processed by unauthorised dealers and fraudulently returned to the manufacturer for warranty claims. RFID will have a huge role in preventing counterfeiting, fraud and theft – and not just in retail. Returning to the issue of data, there is another point. Ronald van Zanden, senior manager with the Internet business solutions group at Cisco, warns: "If you start tagging things you will generate a vast amount of data, and be adding a lot more devices to the infrastructure. Initially the impact might be fairly limited: with data being gathered at, say, pallet or case level, or reusable asset level, you are not talking about very high volumes. But when things get to item-level it will be very different. Ten years from now the majority of all network traffic could be related to RFID." The picture is unlikely to be as bleak as he paints, given work being done by IBM and its partners around the so-called 'Edge' domain, geared to filtering data to exceptions, but it is a challenge we can't ignore. How intelligently – especially at the edges of the network – do you manage the filtering and controlling of the RFID data? What results in a useful event that you need to publish higher up in the network hierarchy? Some will look to their ERP providers to modify their applications and steer best practice and templates – and indeed the big boys are already doing exactly that. But Stewart Holness, CEO of content management software company Reqio, favours the purpose-designed middleware approach designed specifically to cope with the volume of data generated by RFID tags in a rules-based environment. Just one other point: "Whatever you do, you can't take the traditional batch data collection attitude," warns Holness. "Ultimately, your applications need to be integrated almost in real time." And this applies whether your RFID application is internal, or part of a much wider supply-chain initiative – but especially with the latter.