RFID (radio frequency identification) is impacting the supply chain in a big way, but don't look yet for business intelligence (BI) solutions that exploit the mountains of data that will be generated by this emerging technology - at least not from the pure-play BI vendors. Today, no BI vendor has a legitimate road map for leveraging RFID data for BI, and neither do the enterprise application vendors - although that is changing slowly. Until BI solutions emerge in 2007, product-driven enterprises should capture RFID data and model it for analysis themselves. In other words, don't hold your breath waiting for the ultimate RFID BI solution. Build it yourself.
Today, BI Vendors Offer Almost Nothing for RFID Data
RFID represents a new technology for collecting data about the supply chain, with the hope of positively impacting retailers (with on-time delivery of just the right inventory to meet demand), manufacturers (with more accurate demand for items to be made) and distributors (with more precise item location and routing). This impact will be measured by how well each process can be optimized and the level to which each can be synchronized. The measuring requires that appropriate targets, benchmarks and thresholds be established, and that is where BI technology will eventually be of great benefit. However, let's not jump to any conclusions just yet.
In formal and informal conversations with a number of leading vendors including Business Objects, Cognos and SAS, along with application vendors OSIsoft and SAP, to find out what each is currently offering in terms of BI for RFID, we found that the feedback was slim and the news was not encouraging:
- SAP is focused on operations, not BI. SAP clearly has the most mature RFID approach based on its strength in the supply chain and RFID middleware product strategy; however, this does not include BI. Forrester expects there will be analytic applications built on SAP BW InfoCubes and ODS (Operational Data Store), but that is still to be defined by SAP.
- Leading BI vendors are waiting for a greater market demand. Not wanting to waste time, money and other resources on poorly defined supply chain analysis and reporting, the pure-play BI vendors are taking a wait-and-see approach. It will be 24 months before these vendors offer solutions based on RFID data. Consequently, look for the pure-play BI vendors to deliver solutions only after the middleware suppliers integrate RFID data with existing operational supply chain applications.
- Data integration vendors don't yet feel a need to support RFID directly. Consensus among extract, transform and load (ETL) tool vendors is that data integration for electronic product code (EPC) and other data generated by RFID systems is currently the arena of specialized middleware vendors, such as ConnecTerra, GlobeRanger, IBM, Manhattan Associates, Microsoft, OATSystems, Oracle, RF Code, SAP and Savi Technology. An ETL tool can pick up data that originates from RFID once the middleware stores it in a database or hands it to an operational application. However, there's currently no need for ETL reaching all the way to the edge into an RFID system.
A Chicken and Egg Scenario Revealed
The lack of definition of viable and necessary applications built to support BI and RFID has been misinterpreted by the BI vendors as "no good reason to invest in development." The reality is that user organizations are willing to invest in BI technology to leverage RFID data, even without a well-articulated, demand-side statement. What are some of the expected spending levels? It is estimated that a consumer packaged goods manufacturer will spend between $2 million and $20 million to meet retail mandates for RFID. But what does this mean to BI vendors? In early 2004, Forrester surveyed 235 North American IT decision-makers in relevant industries such as consumer packaged goods and retail, and 61 percent planned on increasing their spending on business intelligence (see Figure 1). Indeed, BI vendors are sleeping through the arrival of RFID and thus missing an opportunity to supply solutions and products when there is a well demonstrated demand for such solutions.
Figure 1: RFID Adopters Are Prepared to Invest in BI for RFID
Who Will Deliver Viable BI/RFID Solutions?
The vendors best suited to exploit the RFID growth opportunity, and in the process deliver viable solutions, are the independent software vendors (ISVs) that are closely related to supply chain, transportation and distribution applications providers.
Large enterprise resource planning (ERP) vendors. Oracle and SAP have significant market share and mind share around ERP and are most likely to develop RFID analytic solutions to complement their respective supply chain modules. Both Oracle and SAP have invested heavily in BI technology, with SAP BW data warehousing and SAP NetWeaver application integration, and Oracle Warehouse Builder data warehousing, Oracle Discoverer and OLAP functionality embedded within the Oracle9i AS and 10g AS components of its relational database management system (RDBMS).
Supply chain networks. Companies are spending less on functionally targeted SCM applications in favor of supply network solutions that allow for monitoring, managing and optimizing the entire supply chain network. Forrester projects that the total spending on supply network process improvement initiatives in the U.S. will rise from $2.4 billion in 2003 to $9.1 billion in 2008.
Pure-play supply chain management (SCM) suites. SCM vendors such as Manugistics and i2 Technologies are looking for ways to extend the reach of these suites to more users, via optimization and performance management solutions and deeper relationships with BI partners. To that end, they enhanced their applications to make supply chain insights, buried deep inside their complex systems, more accessible to the users or to integrate these insights with third-party data mining and analytics solutions. For example, i2 Technologies has reengineered its Performance Manager to serve supply chain information to a variety of products, such as those from Cognos, Business Objects and Informatica, to leverage the firm's current investments in business intelligence tools. As long as the SCM vendors can serve the data to the BI applications, this could prove to be the strongest short-term solution.
It's Way Too Early to Buy a BI Solution for RFID
With today's paucity of vendor solutions, pursuing these goals will mean building a solution yourself once RFID systems go online. Before building a BI solution for RFID, consider the following challenges.
- Best practices for RFID and BI are almost nonexistent. RFID can help supply chain visibility with deeper data granularity and more timely updates. To achieve this, companies must develop best practices for integrating this information into BI systems - data warehouses, operational data stores or multidimensional data cubes - where business users can benefit from it.
- BI costs shrink the margin even more. BI costs - including data integration, data modeling, data quality, reporting and analysis - would be over and above the estimated $9 million first-year cost for an RFID system. Companies should leverage existing BI analytic and enterprise reporting solutions and build proof-of-concept BI applications before investing in specialty RFID BI solutions.
- Standards are still emerging for the data generated or captured. Each instance of RFID means a unique approach to capturing different data, making it difficult to predefine what analysis can be done, what data models will represent and how the users can benefit from this data. Companies, however, should not wait for externally defined data modeling standards or approaches before developing BI applications using RFID data. Create new data models that represent first the RFID data available, and then look for ways to redesign existing data warehouses to accommodate the newer dimensions that RFID data will afford.
What's in it For Me?
Many suppliers, distributors and retailers see RFID looming on their horizons, but they don't necessarily see a return on the investment. Develop a value assessment such as the Forrester Total Economic Impact model to clearly define and measure the impact that adding a BI application will have on both bottom and top lines. Keith Gile is a principal analyst in Forrester's Information Delivery research group, covering information management, data management and business intelligence (BI). Gile may be reached via e-mail at firstname.lastname@example.org.
User Example: Collect Data Now, Analyze it Later - Much Later
Companies that are currently capturing RFID data - those under the Wal-Mart hammer - have enough pieces to develop proof-of-concept applications. The CIO of a distribution company acknowledged that his organization has included some RFID data in data marts, just to make it available for inclusion in ad hoc analysis of existing data warehouse data. Lacking a clear set of analytic applications - or a strong business driver - he was at a loss for exactly what to do with this data and, more importantly, how to alter the existing data warehouse, as the granularity of the data was at a lower level of detail than had been previously captured and the quality and validity of the RFID data is suspect.
For now, the goal of this company is to capture as much data as possible and sort it out later, after more real-world experience points to viable best practices. Many other companies that are adopting RFID are in a similar position, corroborating that there truly is demand for vendor solutions that provide reporting and analysis for RFID data.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access