Everyone seems to agree that the deployment of radio frequency identification (RFID) technology will result in a massive explosion of data. This, of course, brings up an important question: How are organizations going to solve the analysis challenges presented by the deluge of data flowing from manufacturing, point-of-sale (POS), distribution, and inventory management systems by way of RFID tags?
First of all, RFID is going to create a lot of new electronic data management requirements. Initial RFID applications will include tracking pallets of products for shipment. Such RFID applications will generate a lot of what you might call “thin” data — data pertaining to time and location.
In addition to providing insight into shipment and other supply-chain process efficiencies, such data should also prove valuable for determining product seasonality and other trends.
However, companies are already exploring more advanced uses for RFID. For example, tire manufacturers have embedded RFID chips in tires to determine — with greater accuracy — time to failure. Pharmaceutical companies have embedded RFID chips in drug containers to better track and avert the theft of highly controlled drugs such as OxyContin.
Airlines are considering RFID-enabling key onboard parts and supplies to optimize aircraft maintenance and airport gate preparation turnaround time. And Starbucks is considering using RFID chips and readers to enable its suppliers to make after-hours deliveries to stores to avoid disrupting their staff’s key function of selling coffee.
All of these scenarios entail different systems generating varying data sets. This is because different systems tend to value similar attributes of a product line differently or track some attributes, such as price, and not others, such as color. For example, point-of-sale systems typically emphasize price and quantity, while a warehouse distribution system can focus on weight and size. These differences can create problems when attempting to analyze performance in making, selling, and stocking products.
But consider what a company can expect once it starts using RFID-enabled applications. Let’s take a clothing retailer, which on average might market, say, 10 different lines of clothing. Each one of these lines of clothing also comes in a number of different sizes as well as multiple colors. Because RFID tags are designed to track all of these attributes plus the product’s name, description, ID number, and price, the amount of data that will be generated is tremendous. In fact, the amount of data generated by RFID-enabled applications is going to be so huge that most companies are going to have difficulty just storing and managing it, let alone analyzing it. Obviously, it will not be practical to analyze all the data generated by RFID-enabled applications.
Thus, companies will have to get very innovative in how they decide to analyze this data.
So what’s to be done? First of all, companies are going to have to deploy some form of global product information management system that will provide capabilities for translating the information generated by RFID-enabled applications into meaningful information. Such a system will provide the architecture that will enable global data synchronization and centralization of RFID data (and other supplemental product data) by allowing companies to collaborate with their customers and partners to create, manage, and synchronize product information within the company and across the entire supply chain.
In addition, this product information management system will serve as a centralized hub for feeding various transaction-processing systems — like enterprise resource planning (ERP), logistics, and manufacturing — with the data needed to take advantage of RFID-enabled applications.
It is also against this centralized data store (of combined RFID data and supplemental product data) that organizations will apply BI tools and analytics to obtain more detailed information and insight into trends such as stock visibility, return management, product assortment performance, and promotion campaign effectiveness, among other uses.
The major enterprise players, including SAP AG, PeopleSoft/Oracle, IBM, i2 Technologies, and others have been feverishly enhancing their products to support RFID-enabled applications.
The BI vendors have been busy, too. For example, Business Objects recently announced a deal with Velosel Corporation to integrate Crystal Reports into the latter’s collaborative product information management system to support analysis of RFID data generated by applications used by consumer packaged goods companies.
The bottom line, however, is that RFID-enabled applications are going to open up all kinds of scenarios for applying BI analytics, many of which are only going to come to light once companies have had a chance to gain some experience with RFID systems.
These are my thoughts on analyzing RFID data. I’d like to hear yours. As always, your comments and insights on this announcement and the business intelligence, data warehouse, and CRM markets in general are welcome.
Send your comments to firstname.lastname@example.org or call me at +1 510 848 7417.
— Curt Hall, Senior Consultant, Cutter Consortium Business Intelligence Practice