To explore these issues, Food Engineering interviewed William R. Friend, group vice president for logistics and information services for the Food Group at J R. Simplot Co., the multi-plant, multi-product food processor based in Boise, ID. Friend manages the group-wide IS function which integrates manufacturing plants, distribution centers and group headquarters, and serves on Simplot's Food Group Executive Committee. Starting with the first plant site in 1995, Bill has managed implementation of the Marcam Protean ERP system throughout the Food Group at Simplot, and is currently integrating Protean ERP with Simplot's Wonderware Manufacturing Execution System (MES).
Food Engineering: What are the basic criteria for justifying investment in ERP?
Friend: Without a threshold return-on-investment (ROI), an ERP project won't get off the ground. Even for small plants, the cost of these projects can exceed seven figures. ROI must be based on a strategic position that the expenditure will deliver some ultimate value in the marketplace, through either higher quality or a cost structure that will produce more volume. Unlike a capital project for building a production line or expanding a plant, however, the investment horizon needs to be longer, because it typically takes longer to get value out of these projects. So you have to think strategically long-term about what direction the market is going in, and what the business requirements will be beyond the normal horizon. The investment must also deliver advantages which are sustainable over time.
FE: What do you mean by sustainable advantage?
Friend: Something that goes beyond just lowering production costs or reducing inventories. Sustainable advantages are those not easily duplicated by competition and which are valued in the marketplace. I think it comes back to issues like reliability, quality and brand recognition -- these things are more sustainable over a longer period of time than just a cost reduction. Since everyone has access to the same software, sustainable advantage must come from being able to anticipate future needs, and incorporate the ability to meet those future needs into the software today. Investment to enhance the MES level or incorporate ERP functionality must be based on future rather than current business scenarios. One strategy you can use to solve more immediate problems is tactical implementation of the software packages. You don't have to impose all the functionality at one time. For example, you can install ERP functionality related to scheduling and tracking product in-process without integrating with the MES layer. You just do the quick tactical things. From an ROI standpoint, there's some immediate return in cost savings. But you have to think longer term about why you're really doing this, and where it's going to take you in the marketplace as you roll-out the whole functionality.
FE: What current trends do you see anticipating future software needs?
Friend: Nutraceuticals are one. Control of ingredients will become much more important. New food additives will require more precision in mixing and blending. This is where the MES layer comes in, and traceability of ingredients will be very important. Packaging materials with sensors will make "smart labels" feasible, and will impact production scheduling and inventory rotation, especially of age-dated products. The ERP layer is where you track and hold that information. Internet shopping and added convenience will create the need for more assembly of food components into a common container, so we will become less a process industry and more a discrete industry. When you mix that trend with food safety issues and growing food imports, traceability and recall of components will become much more complex. ERP systems are really the only effective way of tracking raw materials from the field into a process and on through the supply chain.
FE: Integrating MES with ERP can generate more data than people want or can use. How will users be able to discriminate and get the information they need?
Friend: First, system outputs must be aligned with overall business strategies. Second, user needs must be addressed: You have to decide what information is absolutely needed in real time, to make process decisions, and what can be moved to near real time, for trend analysis. You have to react to real-time information, but it's hard to analyze if a lot is coming at you real fast. There are probably only 10 or 15 critical information points needed to know what's happening at the moment. If you separate real-time information from near real-time, you can manage the information better.
FE: In what formats should this information be presented?
Friend: Real-time information is generated from the PLCs and displayed from the MES layer, so it should be presented on screens. Near real-time is generated by the ERP layer and can be reported from either the ERP system or downloaded to a data warehouse, then analyzed and formatted into reports using tools such as Microsoft Access or OLAP (On Line Analytical Processing). Web-based data is now available directly from system databases. MES and ERP vendors are incorporating these tools into their products. Emerging technology will make real-time information available on hand-held devices, so you won't have to be married to a terminal.
