Although Kanban is an excellent execution tool, it is not intended as a planning tool. However, planning is where MRP-type systems excel. MRP systems work by taking a production forecast and turning it into a series of component forecasts. They do this by using bills-of-materials and routers, which break down each finished good into its basic components . MRP systems also use data on lead times and safety stocks to develop production schedules and raw material requirements. MRP systems excel at planning; however, they are data- intensive systems fraught with problems:
Every forecast is a prediction of future demand based on present information. And, like weather forecasts, the farther out the forecast goes, the less accurate it is. Thus, when you plan your production and material schedules based on these forecasts, you will frequently have the wrong amount of material. If the forecast is high, you will schedule too much production, resulting in excess inventory and unnecessary work. If the forecast is low, you will miss customer shipments resulting in excessive expediting and overtime costs. Since the consequences of stock outs are higher (and more visible) than the consequence of excess inventory, most materials professionals choose to increase inventory levels to protect themselves ("just-in-case" inventory management).
MRP systems are also very labor intensive. Bill-of-materials and routers must be kept up-to date on every part in the system. Since there are always going to be errors (either in the bills and routers or in scrap and production reporting), cycle counts must be taken to correct the on-hand inventory levels. Accurate inventories are critical in a traditional MRP system, as errors lead to stock outs or excess inventory.
Since MRP systems are computer systems, they require special (and expensive) expertise to manage. This expertise takes the form of:
Once all of the files and systems have been created to run an MRP system, they also act as a strong disincentive to change. Every change means that a significant amount of work needs to be done to update the system. Thus, strong resistance to change (in particular the small, incremental changes necessary to achieve continuous improvement) develops in the organization.
Still, there are situations where MRP is useful:
In cases where lead times are extremely long (such as when components must be shipped in from overseas), MRP forecasts may actually be a better solution than kanban. Generally, long lead-time parts also experience variance in transit times, and kanban systems rely on consistent transit times. Also, long lead-time items are usually accompanied by high freight costs. Since many MRP systems utilize EPQ models, they can be useful in keeping total system costs down (see Appendix E, EOQ vs. Kanban).
MRP systems offer an effective tool to turn production or sales forecasts into raw material component forecasts. The material components forecast is particularly valuable for companies with long lead-time items or complex supply chains. For example, tier-1 suppliers in the automotive industry are required to transmit material forecasts to all of their tier-2 suppliers. The tier -2 suppliers are likewise asked to transmit the same forecast data to their suppliers. When this is done throughout the supply chain, inventory can be minimized and the "bullwhip phenomena" minimized.
The "bullwhip phenomena" occurs when different firms in a supply chain modify forecast data before passing it down the chain. This distortion of data can result in huge swings in the amounts of inventory held in different stages of the supply chain, which leads to excess inventories, higher obsolescence costs, and slower responsiveness.
MRP systems can turn forecasts into capacity planning data. Thus, a firm can look into the future and resolve capacity issues (for both manning and equipment capacity). This information can also be used to make modifications to Kanban systems so it can support the new production requirements.
In order to track performance and report key financial information, an MRP system can be extremely useful. It can give detailed, real-time inventory information on raw materials, work-in-process, and finished goods that the accounting department needs to report plant performance. This information can also be useful to the plants' management team in tracking overall plant performance.
The shortcomings of the MRP systems, which we already discussed, are not as critical if the system is not used to execute production schedules and material purchases. Errors in the bill-of-materials, for example, will show up as inventory growth or shrinkage , but they will not lead to a stock out nor will they cost you sales. Thus, they still need to be addressed, but not with the same urgency. Also, the MRP system will not require as much detail in the areas of lead times, container sizes, and routings, since it will not be used to link operations or place material orders. Thus, it is possible to realize the benefits of MRP and Kanban at the same time as complementary management tools. The MRP system can handle the planning tasks , and Kanban can handle the day-to-day production scheduling execution, while preventing overproduction.
Preface