5.8.1 Current Situation
In Section 5.6, we mentioned several measurable benefits when using RFID in supply chains. Many studies have been conducted to evaluate the return-on- investments (ROI) based on the comparison between direct costs and benefits. Un- fortunately, costs and benefits resulting from processes and production system transformations due to RFID are often difficult to ascertain. In other words, intro- ducing RFID results in changes in production systems that depend on the way they are used, and these factors cannot usually be evaluated precisely.
For example:
• The amount of work required to adapt the information system to RFID is diffi- cult, or even impossible, to evaluate accurately. It may be very costly since, in some circumstances, it requires the design of a completely new information system that is usually much more complex than the previous one: the fact that RFID provides real-time data advocates for automation of new activities, which
186 5 Radio-frequency Identification (RFID): Technology and Applications requires new software developments and new resources. This aspect has been mentioned in Section 5.3.3.
• We are able to evaluate the direct loss due to inventory shrinkage, but it is much more difficult to evaluate the cost due to stock-outs resulting from shrinkage (indirect loss). A study conducted at MIT, see (Kang and Koh, 2002), concluded that indirect loss can be 30 times greater than direct loss.
• The increase of production efficiency is expected, but is difficult to evaluate in terms of benefits.
• Benefits resulting from product availability in retailers’ shelves are usually evaluated in terms of sale improvements observed “a fortiori”, but it is usually impossible to evaluate the consequences of these benefits on the image of the company, and thus on the long-term effect on the company.
• The previous remark also holds for reduction of loss associated with product obsolescence.
It should be noticed that even “first range” benefits are not clearly evaluated.
Studies conducted to evaluate the reduction of inventories when using RFID provide estimates that are neither precise nor consistent. Some estimates show in- ventory reduction of 10 to 30% in supply chains (Booth-Thomas, 2003) while oth- ers display 5% (Arens and Loebbecke, 1981)or8to12% (SAP,2003) or 1 to 2%.
Reduction of labor cost due to RFID is evaluated at 30% in distribution, 17% or 7.5%, depending on the study, for retail stores. Other estimations claim that saving in receipt of products in inventory facilities is 7.5% or 5 to 40%, depending on the study. Other figures are 9% in manufacturing, 90% or 100% in physical inventory counting, 0.9 to 3.4% in stores.
Concerning shrinkage and out-of-stock reduction, results mentioned in white papers and reports are also varied. Some of these results are listed hereafter:
1. The reduction of error when picking a product in inventory is 5%.
2. Thefts at retailers are reduced by 11 to 18%.
3. Shrinkage at retailers decreases and becomes 0.78% instead of 1.69%.
4. Thefts on shelves decrease by 9 to 14%, while the reduction ranges between 40 to 59% in stores.
5. Stock availability increases by 5 to 10%.
To conclude, the least we can say is that the benefits resulting from the use of RFID are fuzzy and often simply qualitative. The situation is even worse when evaluating RFID system cost. We have little and diverse quantitative information on the cost of acquiring and implementing RFID systems. Indeed, the costs of tags and readers, as well as the training costs are known. But, as mentioned earlier, the costs of analyzing and designing an information system that fits with the RFID system is usually unpredictable and often very high.
5.8.2 How to Proceed?
As aforementioned, evaluating RFID technology is still an open issue. Similar situations have occurred in the past when computers, and later robots, were intro- duced in production systems. At that time, nobody was really able to evaluate what were the consequences of introducing computers to help managing invento- ries or robots to automate manufacturing tasks. In-house reports at that time show that they were more tailored to support decisions made “a priori” than technically suitable.
Today, everybody understand the usefulness of these technologies despite the fact that lots of users were initially very disappointed and numerous projects failed when these technologies were introduced. The situation of RFID is similar. It is very risky to be among the first who try RFID, but using this technology is inevi- table in the near future.
Three main approaches are in use to evaluate the introduction of new technolo- gies in a specific environment:
1. Ask experts to make their evaluations, and then organize brainstorming until the experts share the same conclusions. Experts proceed by using their profes- sional backgrounds, looking for similar situations in their past and extrapolat- ing the conclusions related to these situations. Indeed, the greater the gap from their work experience to the technology they are trying to evaluate, the more questionable the results. For the introduction of RFID in supply chains, com- parison to bar codes is usually one of the technologies used by experts to build their evaluation.
2. Build a simulation model for the system under study and derive the evaluation of the technology from the results of the simulations. Indeed, this approach ap- plies to specific problems. Furthermore, since simulation models do not capture all the characteristics of a real situation, some expertise should be put at work to infer the evaluation of the system.
3. Make an accurate and systematic analysis of the effects of the application of the technology (RFID in our case) to the problem at hand.
The first two approaches are quite reasonable for technological advances that are evolutionary in nature. In other words, using a rear-view mirror is acceptable to evaluate situations that are reasonable extensions of past situations.
But a specific analysis is required to evaluate technologies that present a gap with the current ones. The analysis needs the use of analytical models that link op- eration behavior to the decision-making system and ultimately to performance measures. It seems that RFID requires this type of approach.
When applying the third approach, we have to keep in mind the following points:
• Inventory shrinkage is a major problem for retailers and manufacturers. RFID can help addressing the problem in two ways:
188 5 Radio-frequency Identification (RFID): Technology and Applications – RFID makes inventory records closer to actual inventory: this results in a
more accurate replenishment policy that, in turn, leads to fewer stock-outs.
– RFID prevents misplacements, avoids frauds, reduces transaction errors, which leads to more accurate inventory records and, as above, leads to fewer stock-outs.
• Reducing inventory shrinkage leads to stock reduction, inventory savings, out- of-stock reduction and sales increases. The benefits associated with these im- provements are difficult to capture in terms of money. The best way to reach this goal is to build models (analytical and simulation models) designed for the specific problem under study. Indeed, building such models requires making assumptions on the efficiency of RFID, the behavior of customers, the behavior of employees, the quality of products, etc.
• The use of RFID increases the visibility of the supply chain and, as a conse- quence, provides better forecasting and more efficient planning, services and inventory managements. But how to evaluate the benefits associated with these improvements? The answer is the same as the one given in the previous subsec- tion.
• To evaluate labor cost saving, you have to examine the tasks that can be sig- nificantly shortened (in terms of operation times) or removed by RFID, but also tasks resulting from new required competencies.
Finally, the use of RFID is not recommended if the supply chain is not “opti- mized” beforehand taking into account RFID characteristics and functions, other- wise we may reach a situation worse than before and, at least, we will be unable to identify the improvements due specifically to RFID.