PDF File: 

asc7_wp_gibbs2.pdf

Intel Corporation

Adoption of RFID across industries has avoided the lowest points of the typical hype/disillusionment curve that most new technologies experience. Now that implementation isunder way, the next step is to perform the thorough analyses required to achieve full value.

Last year I wrote an article for ASCET that described the adoption of RFID in the context of other technologies that have struggled to gain acceptance before they take shape and deliver on their potential. It’s common with all technologies ranging from railroads to electric motors and is particularly common in information technology for hype to enter into the equation before the technology is mature enough for broad adoption. As the philosopher said, history repeats itself, and that was clearly the case with RFID last year as the subject was being discussed nonstop in conferences across multiple industries, but in most cases very few had piloted it, yet alone deployed RFID in a solution. The situation was being fueled by mandates in the retail industry, the U.S. Department of Defense and regulatory agencies that required suppliers to start tagging pallets or items. My concern a year ago was that the combination of too much talk and not enough action often results in a backlash that is as unwarranted and sometimes more pronounced than the hype. I was concerned that some executives who are not as close to technology as I have been would be disappointed if the early experiences were negative, and the backlash would slow down the adoption of this important technology.

Thankfully, the early experiences with the mandates have been remarkably positive, and now we can get on with the business of retooling business process and legacy information architectures to take full advantage of the potential of remote sensing and communication technology, of which RFID is one of the first commercial instances.

The phenomenon of industry hype in advance of a mature technology is so common the consultants at Gartner built a model to describe it. The model is typically referred to as the hype curve, which is shown overlaid with the adoption curve (see Figure 1). On the hype curve a technology trigger is followed by positive hype until it reaches the peak of inflated expectations, which is then followed by negative hype until it lands in the trough of disillusionment, at which point the industry retrenches until positive real results lead to the slope of enlightenment and finally the plateau of productivity. On the adoption curve, which follows an S-shape, discovery is followed by development where the industry then moves to adapt both the technology and business processes until things take off and ramp to maturity. Last year RFID was at the height of the peak of inflated expectations on the hype curve and, outside of a few notable examples such as highway transportation where RFID is used to speed the flow of traffic on toll roads, still slogging along through the development phase on the adoption curve. Today, in my estimation, RFID and the related technologies of autonomic sensors are just entering the knee in the adoption curve just prior to ramp.

RFID apparently dodged the trough of disillusionment due to some very hard work by IT professionals in the retail and consumer packaged goods industry who, with some good fortune, were able to make the first phase of the mandate work with positive results. By the middle of the first quarter Wal-Mart had received over 23,000 tagged pallets and over 600,000 tagged cases, taken over 5 million tag reads and was making RFID data available to suppliers within 30 minutes though its retail link extranet website. At case level, read rates have exceeded 90 percent for cases on carts, 95 percent on conveyors in distribution centers and 98 percent in trash compactors in the back room of stores. There are still issues with reading RFID tags on individual cases on fully loaded pallets, where read rates were just over 60 percent.

As exciting as these early results are, we are only at the end of the beginning. The focus this past year has been on getting the basic reader tag interaction to work, which was clearly a prerequisite to making a complete solution, but now companies have to put the data to work. Capturing value requires a focus on the business process changes and modernization of the IT infrastructure to support them and history again suggests that this takes significant time and effort.

These investments will enable the transformation of RFID data into right-time, actionable information – which in turn can impact four operational cornerstones:

• Visibility – capturing and filtering accurate data throughout the supply chain;
• Reaction – executing business rules in real time based on events captured by the data and alerting the mobile workforce to handle exceptions;
• Collaboration – sharing data in real time across firewalls; and
• Analysis – comprehending and acting on trends and events, from the central core enterprise to all decision-making edges.

These steps may sound straightforward, but they require a detailed understanding of the current business processes to understand where visibility matters and what is required to react and/or collaborate. In my mind analysis will end up being the most nettlesome and interesting area of development. Statistical analysis of real-time data as it is gathered is still a relatively new science in the financial services industry and there are no proven algorithms such as black shoals in financial services to guide industries looking to take advantage of improved visibility to increase forecasting accuracy and operating efficiencies in industries from consumer packaged goods, petroleum, pharmaceutical and healthcare. So while some businesses are seeing value today, there is still a lot of work to do.

Where Is the Value?

Before we dig into the hard work required to take full advantage of RFID I have to answer the question about value. As the industry was riding the curve of negative hype over the last year, questions about value or the lack thereof were rampant. My answer is that there is value everywhere, but not for everybody – yet.

The last year or so a number of companies across multiple industries successfully completed pilots and early deployments with real ROI in transportation and logistics, healthcare, discrete manufacturing and even consumer packaged goods, which was the sector prompting most of the questions. The root cause of most of the questioning was that the mandates were applied to a relatively broad cross section of suppliers, and many of those suppliers did not have a business case to deploy RFID now.

Some suppliers have developed very efficient supply chains using sophisticated processes to overcome the lack of visibility, or the goods they shipped have such razor thin margins that any increase in cost is hard to tolerate. In many of these cases with issues like outof- stock and shrinkage there are still multibillion-dollar problems, so there is room for improvement, and RFID and remote autonomic sensing technologies will help over time.

In the here and now companies like NYK Logistics, St. Vincent’s Hospital and Avis are joining companies around the world to generate real business benefit.

NYK Logistics was recognized with a Computerworld Top Ten award for use of wireless technology for applying RFID and wireless notebooks to their workforce to improve productivity by 39 percent. NYK is a global company with 11,000 employees in 320 offices in an industry where productivity is critical to sustainable success.

St. Vincent’s Hospital in Birmingham, Ala., completed a pilot study last year that projected an ROI of 151 percent based on increased patient turns. Patients were admitted and discharged faster, and that means increased revenues. It also meant that soft indicators of patient satisfaction and staff morale improved.

Avis, the global rental car company, uses RFID in fleet service management. In this application the company equips each vehicle with RFID sensors that are linked to a service and maintenance database. Following each rental, the RFID tag is read automatically and, based on mileage and time of service, the system automatically updates the records and sends a message to the service personnel to perform the appropriate maintenance.

While companies are already benefiting from solutions that use RFID, the technology is still maturing and businesses still have to develop new processes to share data in a collaborative and secure fashion along with new information technology architectures to take full advantage. However, there is value today and in some business cases there is no need to wait to deploy, as early pilots indicate. To help provide some insight, I developed a simple model to indicate where the combination of business case and technology adoption should yield positive ROI.

Essentially for those businesses with high volume and high unit margin or overhead cost there is opportunity to take advantage of RFID now. As the cost of the building blocks comes down over time the ROI potential will open up to more industries and businesses. Accelerating the cost reduction through volume is one of the objectives of the mandates, as well as the real-world application of new business process.

Innovation Matters

If you look into any one of the multiple success stories with RFID as a new technical capability you see companies applying innovation to their value chain process. It should be clear that technology by itself will generally not provide value. Value comes both from the application of new capabilities of which RFID and sensor networks is one important class, and the application of new business processes. Companies that apply both of these will stand to do much better than the competition and realize comparative business advantage.

To support this generalization Deloitte Consulting LLP did a detailed analysis of 300 large global manufacturing companies with revenues ranging from $220 million to $10 billion or higher. Roughly half of the respondents established a baseline of value chain process and capability, and related profit margin. The analysis indicated that 37 percent of the respondents had greater value chain process complexity with roughly the same level of capability such as unique IT, and they were 17 percent more profitable than the baseline. The analysis further showed that 7 percent of the respondents had greater capability with roughly the same level of complexity, and they were 19 percent more profitable.

What is stunning from the analysis of this survey is that only 7 percent of the respondents exhibited innovation on both the process complexity axis and the value chain capability axis, and they were 73 percent more profitable than the baseline. The conclusion is clear – companies that innovate with both process and capabilities fare better, a whole lot better, than their competition.

The question then becomes what to do next, and that is where the going gets tough and is a major cause of why only 7 percent of the manufacturing companies in the survey by Deloitte pursue this level of innovation. It requires considering human factors, process optimization, new technologies and sources of data, new approaches to information architecture where applications and data may need to be treated in a different and flexible way and then considerations on infrastructure where traditional approaches would be prohibitive in cost and complexity.

RFID and sensor networks provide outstanding value due to the data they provide, but this also presents challenges to traditional computing architectures, which are based on a much simpler data and usage model. This is a primary impact of Moore’s Law, which states that as devices get smaller and cheaper they become ubiquitous. Moore’s Law is an empirical observation by Intel founder Dr. Gordon Moore, who postulated in the mid-1960s that the number of devices that could be economically placed on a silicon substrate would double every 18 to 24 months. The law has held true for 40 years and looks to have another decade before human ingenuity runs out of steam. People often ask me, when will we see a 1-cent tag? My answer is, it will be a long time before we can manage all the data.

Moore’s Law has an interesting corollary, which is that the number of users per device is cut in half almost every 18 to 24 months, only the rate isn’t as smooth. For example, in the 1960s when Moore first documented the law, the number of users across a corporation per device – which would have been a mainframe computer – was roughly 1000-to-1 because only large corporations and governments could afford them. As the minicomputer became a significant feature in the enterprise in the mid-1970s, the number of users per device was reduced to 100-to-1. The personal computer and workstation started the trend toward 1-to1 in the early 1980s, but in the beginning only a very few workers had these devices. As the PC ramped into the enterprise and some employees added laptop computers the number approached 1-to-1 in the mid-1990s. With smart phones, RFID and sensor networks the number of devices per user will shift toward one user per 1000 devices and this will have a pronounced impact on information architectures if companies want to turn the devices, processing power and data into value.

The legacy architectures that may have worked with some tweaks and band-aids as the industry moved from a usage per device model of 1000-to-1 toward 1-to-1 will not be able to handle the next move from 1-to-1000 without some significant changes. RFID and sensor networks are adding to the overarching trend of convergence of computing and communications that along with continued penetration of the Internet has been causing another corollary to Moore's Law, where data is doubling every 12 to 18 months, raising the need for a simpler IT environment that is more cost-effective, flexible and secure. To help streamline this evolution we have developed a framework we call the service-oriented enterprise (SOE).

The Service-Oriented Enterprise

As I’ve discussed, the potential for information-age business transformation through RFID and sensor networks is exhilarating. Yet it can also be daunting. How can IT organizations effectively deal with the explosion of data and systems complexity, satisfy the ever-more stringent requirements for compliance, privacy and safety, all while reducing costs?

Enabling the kind of transformation that will drive competitive advantage requires a fundamental shift to innovative business processes. These new processes must be supported by a modular, manageable IT infrastructure based on a service-oriented IT framework. An SOE provides automated assistance to set up and coordinate all of the interactions implied and required by highly integrated business systems. Through this orchestration of these multiple systems, a service-oriented enterprise:

• Delivers software and data as services;
• Organizes hardware as virtualized resources;
• Uses autonomic data sources such as sensor-based and RFID solutions;
• Supports occasionally connected or “mobilized” software applications; and
• Enables the same services to cross firewalls to provide availability to customers, suppliers and distributors as needed.

Using a service-oriented framework, businesses can transform themselves by providing new capabilities, eliminating complexity and simplifying systems integration and management. They also can enable real-time, adaptable and agile IT organizations that can more effectively compete in a challenging business environment.

“A service-oriented enterprise is really connecting business processes in a much more horizontal fashion,” explains Martin Brodbeck, Director, Global Application Architecture for Pfizer Inc. “It’s having a federated infrastructure that provides an architecture and security foundation to be able to run these components consistently across your enterprise.”

How can companies achieve greater service orientation? The first step is to break their large monolithic solutions into flexible components, otherwise known as services. These services can then be reconnected in a loosely coupled manner, which enhances agility of the resources. It’s important to use ubiquitous Web services and standard protocols to connect everything for ultimate flexibility. Finally, services can be reused whenever appropriate for maximum efficiency.

Intel’s vision for a true service-oriented enterprise is based on open industry standards – for both hardware and software building blocks – that are delivered as complementary, interoperable services that interact via standard Web protocols. Intel is helping accelerate the transition to SOE by designing its computing and communications platforms to deliver new capabilities with greater end-user value and benefit. In addition, Intel’s platforms support a range of new features and technologies, ranging from multi-core processors to key silicon technologies known as the “T’s” that enhance manageability, multithreading, security and virtualization.

Intel is actively supporting RFID at multiple levels, from building blocks for readers to Itanium Processor-based servers to solution services practices that focus on adaptive architectures that help enable inevitable changes in business processes. Intel works with leading software, hardware and service providers to bring the most advanced solutions to our business customers.