Has MRP run its course? A review of contemporary developments in planning systems

The onset of global competition in the 1970s began to change the manufacturing environment drastically. The advent of the computer created an opportunity for the developers of material requirements planning (MRP) concepts to automate many of the manual practices employed in manufacturing for acquiring and tracking of materials. This resulted in more efficient manufacturing operations in terms of labor for planning activities and better material control. Technology was expanded to include capacity planning and production schedule control. The acceptance in industry for the new `̀ tool’’ was monumental and soon a new industry was born. For some time MRP allowed manufacturers to perform at higher levels of proficiency. However, the demands and expectations of the customers have continued to change and manufacturers wishing to keep pace with competition are beginning to question if MRP is still a valid tool for production planning and control. As a result new concepts have begun to emerge. These new approaches call for an abandonment of some of the foundational components of MRP. being constrained by the ``system’’ which limits their ability to be agile, results in larger raw material and finished goods inventories, creates false manufacturing lead times and stifles efforts to become more integrated throughout their information systems. 2. Evolution of production planning and control Having discussed briefly the history of MRP and how its development has been lacking in an overall business view it is beneficial to see how current thinking in the area of supply chain management may impact developments in this area of this technology in the future. In order to provide an adequate understanding of the potential impacts, this paper will go into a fairly detailed discussion comparing MRP to two newer concepts in production planning and control, namely capacity, commonality and consumption (3C) and demand flow technology (DFT). A brief description of MRP terms will aid in the understanding of the following discussions: Material requirements planning (MRP) (sometimes referred to as little mrp). This was the acronym given to the initial methodologies developed for the material planning process that used a product billof-material (BOM) and a production forecast, by specific finished good, to determine future material needs and purchase timing. In this phase of development, capacity of the manufacturing process was assumed to be infinite for the purposes of calculations. The material was planned for according to the forecast timing, with no regard given to actual capacity. Manufacturing resource planning (MRPII) (sometimes referred to as big MRP). This was the next evolution of the methodology, which added manufacturing routings and work center capacities to the model. In this phase of development the finite nature of capacity was attempted to be modeled mathematically. This information was used for the determination of the production schedule, now based on capacity, and the timing of material acquisition activities to support the schedule. Forecasts at the finished goods level were still the basis for driving the system. Enterprise resource planning (ERP): ERP is third step in the evolution of these systems. This step added additional management tools for areas of a manufacturing operation not previously managed by the planning system. These areas include human resources and design engineering/product development. In addition ERP has broadened the view of the material planning function by extending the order tracking function out to the supplier. For the following discussion MRP will be used for generalization purposes to refer to the concepts described above forMRP, MRPII or ERP. In situations where it is necessary to be more specific, MRPII or ERP will be used for the specific aspects of the respective methodologies. 2.1. MRP Earlier in this paper it was explained that the concepts utilized in MRP were originally developed in the late 1960s and early 1970s. At this time several unique differences existed in manufacturing when compared to today’s business environment. During this period the USA was still going through a post-Second World War boom in terms of manufacturing capabilities and capacity. The manufacturing capabilities that had been developed to support the war effort had been shifted to consumer goods. The marketing experts on Madison Avenue had been very successful at expanding demand for consumer goods. As families grew larger in the ``Baby Boom’’ era, people were convinced that they needed more in order to enjoy modern life. Bigger homes, appliances, multiple televisions, and two automobiles were some of the desires that had been created. This demand helped to fill the capacity that had been developed during wartime and also helped to create a period of prosperity that had not been seen before in the USA. Additionally, global competition had not become an issue yet. As a result, manufacturers were able to enjoy a unique economic environment. They had a captive audience with an increasing desire for material goods. If demand outran capacity then lead-times (delivery times) were extended, which was very common during this period. Waiting for the delivery of a car or an appliance was expected. Model proliferation was much lower as well. These two issues alone significantly reduced the overall challenges of manufacturing by reducing the overall complexity of the product, as well as reducing the need for finished goods inventories. [ 454] Sameer Kumar and David Meade Has MRP run its course? A review of contemporary developments in planning systems Industrial Management & Data Systems 102/8 [2002] 453±462 2.1.1. Foundation of MRP It was during this time that Oliver W. Wight (Wight, 1984) and Joseph Orlicky developed the basics of MRP. Their work was focused at developing a better method for planning for the procurement of manufacturing materials to support the manufacturing process. The basis for their `̀ system’’ was a sales forecast. This piece of data was the driver for the entire system. With an accurate sales forecast the remaining work to be done was simply a series of calculations that were related to various time periods. These calculations produced the triggers to start the flow of materials into the operation to support the plan. This was all made possible due to the, then new, commercial availability of computers. The once laborious and inaccurate (or impossible) method of manually calculating material needs and tracking material movements could now be automated. This opened doors that had not been possible before due to the volume of calculations required. For the first time, planning accurate future purchases was possible by using a sales forecast (future needs) and knowing what the current on-hand inventories were. This was sufficient for a while and in fact did much to improve manufacturing capacities by eliminating part shortages, as well as controlling inventory costs by preventing the accumulation of too much of any component or material. Unfortunately, the consumer environment did not remain static. Some significant changes have occurred that have created major challenges to the successful application of MRP concepts in manufacturing. 2.1.2. The basic problem with MRP in today’s manufacturing environment Although many of the traditional drivers for manufacturing have shifted over the years, making past practices for production planning obsolete, only two such changes will be discussed here. The first is the requirement for immediate delivery of a product following a customer order and secondly, the increased complexity of the products being requested. Immediate delivery. In today’s instant gratification world it is no longer acceptable to consumers to wait for the delivery of a product as they did in the past. For larger purchases it may be acceptable to wait several weeks if some level of customization is being performed, e.g. special color for a car. In an ever-increasing amount of the cases, consumers are even expecting the customized products to be delivered immediately or within a minimal lead-time for a product, e.g. Marvin Windows in Warroad, Minnesota. This development has added the requirement to carry finished goods or near-finished goods inventories at the point-of-sale of the right model in the right quantities. In today’s market, wrong models or inadequate quantities equate to a lost sale. Increased complexity in products. Product proliferation, driven by consumer desires, has led to an ever-increasing need for additional component or material inventories to produce the various models. It has also led to the need for more records (electronic files) as to what components are required for a specific model (BOM). The various files are required by MRP to calculate material needs and therefore must contain current information and be accurate. Increased computing power to perform the calculations and deal with the expanding volumes of data has kept pace but the manual entry and maintenance of the records has become a growing issue. These two issues have led to the inability to accurately forecast future sales at a finished goods level and the inability to ensure adequate inventory levels of materials and components to support the changing production demands. Supporters of the MRP approach often present the mechanics behind the system as relatively simple. Successful operation of the production planning function merely depends on three pieces of information: 1 accurate production plan for final assemblies (forecast); 2 accurate BOM; and 3 accurate on-hand inventory information. With this information you can accurately calculate what you need and when you need it. Depending on a person’s point-of-view and experiences, the previous claim could either look like a sound, logical statement of a manageable function or it could look like complete sarcasm. It is at this point that the proponents of new methods for managing the procurement process and the established procurement world diverge. A dissection of the above claim will further define this point. The second and third points of this statement will draw argument from the non-MRP camp but not necessarily firm disagreement. Most will agree that it should be possible to maintain an accurate BOM. [ 455 ] Sameer Kumar and David Meade Has MRP run its course? A review of contemporary developments in planning systems Industrial Management & Data Systems 102/8 [2002] 453±462 As well, it should be feasible to maintain accurate on-hand inventory information. However, in practical application under the basic methods of MRP, this has rarely been achieved. This is due to the processes employed, back-flushing material out of inventory based on quantity of final assemblies produced, using BOM data for component quantities and adding scrap/reject information to account for components lost due to failures. Back flushing deducts the components and materials calculated to have been used from those materials that have been received into the operation and in this fashion keeps a running total of on-hand inventory. BOM inaccuracies and inaccurate reporting of lost, scrapped or substituted parts all contribute to inaccuracies in on-hand inventory information. Cycle counting programs are intended to capture these errors and allow corrections to be made. In the end however, stock outages still occur despite the manual effort and cost to prevent them from happening. Stock outages result in production stoppages and lost production, which in turn results in lost sales. As previously stated, in theory all this would work. Given the right information the computers can perform the necessary calculations accurately. All these variables are within the complete control of the manufacturer. Unfortunately it requires a certain level of attention of a great number of people to maintain accuracy. Add to this the fact that the mechanics behind MRP are abstract, meaning that the information is inside the computer and how it is used is unknown to most of the support people. This leads to a lower level of commitment from the necessary supporters for making sure they have done their part in maintaining accuracy. These alone create a significant barrier to the success of any MRP system. However, it is this first point, having an accurate forecast, which the opposition uses to dismiss altogether the viability of MRP in today’s business environment. This is the major point of divergence between the two groups. The lack of an accurate forecast is what many MIS managers and MRP providers and consultants have used in their defense when brought under attack for the poor performance delivered by their systems. Here the main issue for most manufacturers is that forecasting cannot be done with any level of accuracy or repeatability. This is the key realization that needs to be believed, in order to accept the arguments from the practitioners of a new approach. It is easy to see the evidence of this in all walks of life, not only manufacturing. Retail is a good example of this. End-of-season sales would not be necessary if the forecasting model had been correct. Grocery is another example of this. Often milk is sold at a significant discount at Meijers stores, a chain in the central Midwest of the USA, as the product approaches its expiration date. Restaurants discard food daily that was in excess of what was required for the day while at the same time running out of product that is in demand, requiring the customer to choose something else or go elsewhere. Despite all the hi-tech tools, it is no different in manufacturing. The reasons for this are due to: endless shifting in consumer desires, breakthrough developments in designs, materials or components, economic drivers, etc. . . . most of which are impossible to predict at all or at least what impact they will have on the market. History is also a poor indicator, but one that is often used for lack of a better approach. However, as is true in investments, past results are not necessarily an indication of future performance. There is any number of arguments to the inaccuracies of this approach. Lucent Technology, a world leader in communications equipment, with the power of Bell Laboratories behind them can generally forecast monthly demand for a particular product with an accuracy of 35 percent (Kopczak et al., 1998). As a result, in order to maintain high fill rates, Lucent has to either carry enough finished goods to cover variability or quote long lead times to their customers. The difficulty with producing a reasonably accurate forecast has increased exponentially since the creation of MRP. Globalization of the marketplace, technology and the rate in which it advances, continued shortening of the design process for new products and the insatiable desire of the customer for new and different as well as more material goods, have all been part of the problem. The more stable, slower-moving economic environment of the 1960s and 1970s