Читать книгу Demand Driven Material Requirements Planning (DDMRP), Version 2 - Carol Ptak - Страница 14
ОглавлениеMaterial Requirements Planning in the New Normal
As discussed in Chapter 2, the primary enemy of the protection and promotion of the flow of relevant information and materials is the bullwhip effect. The bullwhip effect exists largely due to the characteristics and configurations of conventional planning systems utilizing MRP. This chapter will describe these characteristics and configurations and highlight one key attribute as a core problem.
The APICS Dictionary defines Material Requirements Planning (MRP) as a:
A set of techniques that uses bill of material data, inventory data, and the master production schedule to calculate requirements for materials. (p. 103)
MRP is essentially a calculation hub. The master production schedule feeds demand signals to MRP, which in turn creates a synchronized list of supply orders based on current inventory records (on hand and on order) and product structure (bill of material). The supply orders have date and quantity requirements that define the elements of that synchronization plan. These date and quantity requirements are then fed to a manufacturing execution system. They are turned into transfer orders to distribution sites, manufacturing orders to be scheduled on the shop floor, and purchase orders to be relayed to suppliers. Figure 3-1 shows this conventional planning approach.
The requirements to run MRP are simple and straightforward:
The master schedule must be stated in terms of the bill of material.
Unique item numbers exist for every item.
The bill of material exists at the time of planning (product structure file).
Inventory records are available for all items (inventory record file).
When these requirements are implemented in the computer system, then the MRP batch program can be run. However, to be considered a Class A user or to expect some kind of reasonable result from the computer system, the following assumptions are made:
FIGURE 3-1 The conventional planning schema
File data are 100 percent accurate and complete.
Lead times are fixed and known.
Every inventory item goes into and out of stock.
There is full allocation; no order is started unless all the components are available. Components are discrete—things can be counted and measured (no “use as required”).
There is order independence, which means that every order can be started and completed on its own.
MRP was a huge leap forward because for the first time what was required could be calculated based on what was already there compared with what was needed, with the net result time phased. The objective of MRP was to precisely time-phase the requirements and replenishments to dramatically reduce inventory from the previous order point approach where some of everything was kept around all the time. This ability to calculate dependent demand through a bill of material was a significant development. It was no longer necessary to forecast dependent demand—it could be calculated based on the expected demand for the parent part. APICS defines dependent demand as:
Demand that is directly related to or derived from the bill of material structure for other items or end products. Such demands are therefore calculated and need not and should not be forecast. (p. 46)
MRP evolved because of the advent of the computer, and the age of marketing in the 1950s introduced more product variety and complexity than was managed previously. Order point (the previous method of materials management) clearly could not affordably handle these new requirements. To understand how planners deal with MRP on a daily basis, refer to Appendix A, where a simulated environment demonstrates the day-to-day difficulties associated with MRP.
Yet even if a company has 100 percent of the requirements and 100 percent of the assumptions validated, the conventional planning approach will still be ineffective. The remainder of this chapter will explain why.
Distortions to Relevant Information
The conventional planning approach actually creates the bullwhip effect and its inherent distortions to the flow of relevant information and materials. Some of the ways in which conventional planning creates the bullwhip is related to the manner in which convention chooses to use MRP. Other contributions to the bullwhip are related to hard-coded traits in MRP systems. All of these issues, however, are related to one key and fundamental attribute of MRP.
MRP is essentially a calculator. It needs three basic inputs to perform its calculation. One of those inputs is “demand.” Different demand inputs will produce different outputs. The APICS Dictionary defines demand as:
A need for a particular product or component. The demand could come from any number of sources (e.g., a customer order or forecast, an interplant requirement, a branch warehouse request for a service part or the manufacturing of another product. (p. 44)
By this definition, demand can be broken down into two different types: forecasted and actual. Both of the following definitions are from the APICS Dictionary:
Forecast. An estimate of future demand. A forecast can be constructed using quantitative methods, qualitative methods, or a combination of methods, and it can be based on extrinsic (external) or intrinsic (internal) factors. Various forecasting techniques attempt to predict one or more of the four components of demand: cyclical, random, seasonal, and trend. (p. 68)
Actual demand. Actual demand is composed of customer orders (and often allocations of items, ingredients, or raw materials to production or distribution). Actual demand nets against or “consumes” the forecast, depending upon the rules chosen over a time horizon. For example, actual demand will totally replace forecast inside the sold-out customer order backlog horizon (often called the demand time fence) but will net against the forecast outside this horizon based on the chosen forecast consumption rule. (p. 4)
The type of demand that is chosen to drive the MRP calculation is a primary determinant of how much relevant information can be produced from MRP. Remember, the flow of information and materials must be relevant to the required output or market expectation of the system. To be relevant, both the information and materials must synchronize the assets of a business with what the market really wants; no more, no less.
A hard-coded trait of MRP is that with a given demand signal, MRP is designed to net perfectly to zero. You make exactly what you need without any excess. In this regard it could be argued that MRP is the perfect JIT system. If the demand signal is perfectly accurate, then the MRP calculation will be perfectly accurate. Given that the math allows no tolerance for error, it seems obvious that MRP should only be given as accurate a signal as possible.
With that in mind, should the demand input to MRP be what a company thinks the market wants to buy or what the customers actually want to buy? Which will produce a more relevant result? As described in the definition of actual demand as well as Figure 3-1, the conventional approach combines both types of demand. Forecast is used to create planned orders, and then demand is adjusted as the picture becomes clearer with actual orders. Why is this problematic?
There are three truths about forecasts:
1. All forecasts start out with some inherent level of inaccuracy. Any prediction about the future carries with it some margin of error. This is especially true in the more complex and volatile New Normal.
2. The more detailed or discrete the forecast is, the less accurate it is. There is definitely a disparity in the accuracy between an aggregate-level forecast (all products or parts), a category-level forecast (a subgroup of products or parts), and a SKU-level forecast (single product or part).
3. The more remote in time or farther out forecasts go, the less accurate they get. Predicting the weather tomorrow is much more accurate than predicting the weather 52 days from today. Yes, history can be used as a basis for a prediction, but the margin of potential error is much higher. It is not uncommon that in many industries the accuracy of a forecast can drop below 10 percent beyond 90 days at the SKU level.
Today many forecasting experts admit that 70 to 75 percent accuracy is the benchmark for the SKU level. Figure 3-2 is the results of a 2012 survey conducted by forecastingblog.com showing the reported forecast error rates across various industries at the SKU level.
Unfortunately, when you start a serial, complex, and interdependent process with an error-prone input, the resulting output integrity must be suspect. Planned orders are derived from these forecasts, and very real commitments of cash, capacity, and materials are directly derived from a prediction that is subject to varying degrees of inaccuracy, sometimes with extremely significant degrees of inaccuracy.
As time progresses, the demand picture changes with the incorporation of actual demand, MRP is rerun, and subsequent changes occur. The result is that we end up with things that we do not need and desperately expedite things we have just discovered that we do need. These are the three effects of the bimodal distribution. Thus the bimodal distribution starts with the use of planned orders based on that forward-looking forecast.
FIGURE 3-2 Average forecast accuracy across industries
Rohan Asardohkar, August 22, 2012, http://www.forecastingblog.com/?p=423
This is a known and accepted routine in most industries despite the waste and performance erosion associated with it. Why would industry intentionally sabotage performance by using an input with known inaccuracy to drive activity and commitments when there is an obvious alternative? Why not just use only sales orders?
The most accurate form of demand input is a sales order. A sales order is a stated intention and commitment to buy from an actual customer in terms of both quantity and time. It is essentially an uncashed check. In this way it is a highly accurate and relevant piece of information. There should be no debate that sales orders are an order of magnitude more accurate than planned orders. So why don’t companies simply load only sales orders into MRP?
Using MRP with only sales orders, however, assumes something that does not exist in today’s New Normal—enough time. A basic attribute of MRP is to net to zero across the entire network of dependencies. This means that MRP by definition makes all activities dependent on each other. Thus, in order for MRP to be that perfect JIT system, there must be sufficient time to procure and make everything to the stated demand—called “cumulative lead time” (the longest stated chain of time in the bill of material including purchasing lead time).
This means that customer tolerance time would have to be equal to or greater than the cumulative lead time. Today’s supply chains, however, are characterized by shorter and shorter customer tolerance times and extended, elongated, and increasingly complex supply chains. There simply is not sufficient visibility to sales orders soon enough to properly plan for them using conventional MRP. Figure 3-3 conceptually shows the disparity between when companies gain visibility to sales orders (actual demand) versus the time that it takes to procure and produce the product (the time frame in which MRP makes it calculations).
FIGURE 3-3 Manufacturing and procurement times versus sales order visibility
FIGURE 3-4 Planning horizon depiction
With MRP’s characteristic of making everything dependent, the only way to find enough time is to attempt to predict what actual demand will look like so that an organization can attempt to ensure that the necessary materials are available in quantity and time as the market places its sales orders. A “planning horizon” extends into the future far enough to cover the cumulative procurement and manufacturing cycle. Figure 3-4 shows the planning horizon covering the cumulative procurement and manufacturing cycle in the example.
This explains the need to load MRP with demand that is largely derived from a forecast and then to make adjustments close in as sales orders become visible. Planned orders for end items are launched at the beginning of the planning horizon. The longer the procurement and manufacturing cycle, the longer the planning horizon must be. The longer the planning horizon, the less accurate the planned orders will be. The less accurate the planned orders, the more course corrections are required. This constant set of corrections brings us to another inherent trait of MRP called “nervousness.”
MRP’s nature of making everything dependent creates nervousness. Nervousness is the characteristic in an MRP system related to changes in parent demand transferring down and across bills of material. The APICS Dictionary defines nervousness as:
The characteristic in an MRP system when minor changes in higher level (e.g. level 0 or 1) records or the master production schedule cause significant timing or quantity changes in lower level (e.g. 5 or 6) schedules or orders. (p. 86)
Figure 3-5 illustrates the concept of nervousness. The figure illustrates the product structure for an end item called FPA. A timing or quantity change in FPA ripples down through the entire product structure, causing timing and quantity changes at every component position as the system constantly strives to net to zero. The dotted curved arrows depict that change. This creates a constant series of action messages for planners and buyers to review and interpret.
The challenge of system nervousness has been known since the earliest days of MRP. However, the system nervousness was manageable since plans were done once per month. Concepts like firm planned orders, the demand time fence, and the master production schedule were developed to manage the nervousness. But the complex and volatile environment characterized by the New Normal makes the issue a bigger challenge. Given the nature of MRP to make everything dependent, the only way to stop nervousness is to make no changes. Yet that would mean significant service challenges, as the forecasted orders will vary (many times dramatically) from what the market will really desire. What can be done to limit the impact of nervousness? MRP users are forced into compromises in order to slow down the rate of changes.
FIGURE 3-5 Nervousness illustrated
In most conventional environments, planning occurs in weekly buckets. This is a direct effect of the nervousness discussed above—nervousness that is directly related to the use of planned orders with MRP. Planning organizations know that if they ran MRP daily, or worse yet in real time, the resulting nervousness would create chaos. The amount of action flags and messages on the planning screens would be overwhelming.
Instead, a weekly interval is typically used to calm the waters on a daily level. This, however, comes at a price. First, it forces the planning horizon to extend even further (one week). This has a direct correlation to the level of signal inaccuracy at the end of the horizon. Second, it creates a latency that almost guarantees that the level of change between MRP runs will be dramatically larger. Instead of lots of little changes on a daily basis, there are massive changes (and signal distortions) on a weekly basis.
Figure 3-6 depicts the differences in net change impact between daily and weekly MRP runs. The upper left hand bar chart depicts MRP run each day. The level of each change is relatively small but each change ripples through all lower dependencies. The bar chart in the upper right portion of the graphic depicts a weekly MRP run. Days 1–7 are stable (no change) yet Day 8 introduces a significant change (40) that will ripple through the environment. The relative difference in changes is depicted in the chart in the lower left corner of the graphic.
Planning organizations are stuck between these two hard places because of MRP’s hardcoded trait of making everything dependent.
Flattening the Bill of Material
Another way to combat nervousness is to reduce the number of connections that MRP sees and calculates against. One way to accomplish this is to “flatten” the bill of material of a product. MRP from a planning and synchronization perspective then becomes blind to intermediate components. Figure 3-7 illustrates the difference between a full product structure (on the left) and a flattened one (on the right). The flattened structure has eliminated all intermediate positions.
While this reduces the amount of changes to intermediates (since there are none) and this reduces the total number of action flags, does it produce more relevant information or actually distort the picture further? The key to more relevant information is not to simply ignore dependencies. When we ignore critical dependencies, we risk oversimplification. Oversimplification means to simplify to the point of error, distortion, or misrepresentation.
The bill of material files used in a planning system should reflect how the product is actually made. Dramatically flattening bills like the example in Figure 3-7 effectively ends any capability to provide visibility to and plan for synchronization between the finished and purchased part levels. The price for this is paid by the manufacturing floor as scheduling and schedule execution become an order of magnitude more difficult.
FIGURE 3-6 Daily versus weekly MRP runs
FIGURE 3-7 The flattening of a bill of material
All of these factors combine to mean that MRP is producing plans:
With high degrees of known error (forecast input)
In a constant state of change (nervousness)
With a degree of latency (weekly bucket)
That may misrepresent the environment (flattened bills of material)
This means that the very nature of MRP combined with the way that it is typically used inevitably leads to distortions to relevant information. Furthermore, all of these distortions to relevant information have been related to one single attribute of MRP. Have you spotted it yet?
Distortions to Relevant Materials
The next consideration is the supply portion of the bullwhip—the distortion of relevant materials. As mentioned previously, MRP creates a synchronized and precise plan at all levels of the bill of material based on its required inputs and assumptions. This plan will happen only if everything in the entire dependent network goes precisely according to plan. In almost every modern environment, this is an impossibility for two reasons.
First, there is a basic and inherent level of variability in any environment, even one deemed to be in control. Deming called the normal or random variation that occurs in processes “common cause variation.” Normal or random operational variability results in a process that may be statistically within calculated control limits but still varying between those limits. Reducing the gap between the limits is a worthy goal. The elimination of the gap is an impossibility—it would require every process to be perfect.
We know that any process cannot be perfect. The collective effect of this imperfection must be examined. Figure 3-8 appeared in the first and third editions of Orlicky’s Material Requirements Planning. The figure has three columns. The first column is the number of components required to make a parent item. The second two columns are different levels of average component availability. The left column assumes all components have 90 percent availability, whereas the right column assumes 95 percent availability. For example, a parent item with 4 components that average 90 percent availability has a 65.6 percent (.9 × .9 × .9 × .9) chance that all components will be available simultaneously when required. A parent item that has 10 components that have an average of 95 percent availability will have a 59.9 percent chance that all components will be simultaneously available when needed.
Figure 3-9 shows how less than perfect material availability results in an erosion of the probability that all materials will be present when needed. Remember, MRP assumes full allocation—no order should be started unless all the components are available. In fact, even if many components have extremely high variability or arrive early, the parent order release is still at the mercy of any one missing component.
Figure 3-9 illustrates an environment in which four of the materials have high availability while one component has low availability on average. Components 1, 3, 4, and 5 have extremely high average availability (95 percent, 98 percent, 97 percent, and 99 percent, respectively. Component 2, however, has a relatively low average availability level (72 percent). The impact that component 2 has on the overall probability that all components will be available when required is significant; that probability drops to 64.4 percent. This translates to delays in the planned release.
FIGURE 3-8 Probabilities of simultaneous availability
FIGURE 3-9 One problematic material
Thus a simple rule emerges with regard to dependent structures that contain integration points requiring simultaneous inputs to advance to the next stage of the structure or plan. This is a valid description of the plans that MRP generates. This simple rule is “delays accumulate, while gains do not."
Figure 3-10 conceptually illustrates this effect. A dependent structure is visible at the bottom of the graphic. In this case that dependent structure is a synchronized plan based upon product structure. There are concurrent paths and integration points culminating in a finished item (FPA). Above the structure there is a graphical depiction of delay accumulation. The arrow steadily rises as activity progresses through the planned build. The arrow’s position at any one place depicts both how far along the planned activity path the build is (X axis corresponding to the structure) and the amount of accumulated delay (Y axis).
This effect is only partially impacted by signal accuracy. In other words, the demand signal could be perfect, but delay accumulation will still affect the environment if normal and random variation exist in the resources required to execute those signals. This delay accumulation results in an effect that is frequently referred to as “supply continuity variability.” This forces two profound realizations:
1. From an execution perspective MRP will never create a realistic plan in environments of even moderate complexity.
2. Any true solution to the bullwhip effect must address both demand signal distortion and the material supply distortion (supply continuity variability).
FIGURE 3-10 Illustrating delay accumulation
Amplifying the Distortions to Relevant Information and Materials—Batching Policies
The distortion to relevant information and material inherent in the bullwhip is amplified due to batching policies. Batching policies are determined outside of MRP and are typically formulated to produce better-unitized cost performance or are due to process restrictions or limitations. Batching policies dictate the way that MRP must perform its calculation (demand signal distortion) as well as influence the way in which materials progress through a supply chain (supply continuity variability).
The batching policies that dictate the MRP equation include order minimum—the amount that must always be ordered; order maximum—the largest quantity that can be assigned to an order; and order multiple—a rule that governs ordering between the minimum and the maximum. The order minimum and maximum should be evenly divisible by the order multiple.
An example: an intermediate component can have an order minimum of 100, a multiple of 50, and a maximum of 500. This means that if the intermediate component has a parent demand of 102 pieces, a minimum of 150 (the minimum plus the next multiple) of the component must be ordered to cover that demand. At some point later if the parent requirement changes to 99, the intermediate component requirement drops to 100. The parent changed by 3; the component changed by 50. The effect of this complication is devastating in any environment where ordering policies, particularly minimums and multiples, are dramatically different at each level of the bill of material.
Figure 3-11 is an example of the demand amplification in a more complex environment. An end item (FPA) has three components. All three components have minimums and multiples assigned. A demand of 115 for FPA will yield demands of 150 for Intermediate Component A (ICA), 250 for Intermediate Component B and 200 for Intermediate Component C.
Batching practices can dramatically affect the way that material moves in a supply chain, contributing to or amplifying the accumulation of delays. For example, delay accumulation could occur while an order waits on a truck for other orders to fill up the truck. The transportation batching policy dictates that only full trucks are allowed.
The logic and policies behind batching policies can be very problematic. Most batches are heavily influenced by an emphasis on protecting unit cost and have no consideration for flow. That emphasis on unit cost actually further distorts the flow of relevant information throughout most companies. The assumption that driving to unit cost performance equates to the best return on investment performance is unequivocally and mathematically proven false. Yet industry ignores this fact every day. This subject, however, is technically outside the scope of this text. For an in-depth look at this issue, refer to Demand Driven Performance: Using Smart Metrics by Debra Smith and Chad Smith.
FIGURE 3-11 Batching complications to MRP supply order calculations.
Are the challenges described in this chapter and Appendix A unknown to seasoned planning professionals? Absolutely not. These challenges are well known and common. They explain the existence of the poor asset performance, the work-arounds, the bimodal inventory distribution, and the bullwhip effect. Additionally, they leave most planning organizations in a huge dilemma: utilize MRP or ignore it. The answer to this dilemma is almost always the same; do both. Most organizations are simultaneously ignoring and utilizing MRP. Just how much ignoring and utilizing is something that tends to be specific to organizational functions and the individual users. There has to be a better way.
MRP enabled organizations to quickly calculate and synchronize total requirements given a set of demand inputs. This was of particular importance when the company had a deep bill of material or many shared components. The whole purpose of MRP was to synchronize connections and dependencies. In the New Normal there are undoubtedly more connections and dependencies than ever. Thus MRP should be more relevant today than ever. Yet MRP is failing in the New Normal.
MRP’s role in the modern supply chain is significant. Even in the New Normal, the heart of every supply chain is manufacturing, and at the heart of manufacturing is MRP—it is the conductor of the supply order symphony in every supply chain. Each node in the supply chain has an MRP system supporting a different manufacturing operation. Therefore, a primary limitation of any supply chain will be how well MRP systems perform not just individually at each node but also collectively throughout the web.
If industry wants more agile manufacturing and supply chains that protect and promote the flow of relevant information and materials, then industry will need a more agile form of MRP. As evidenced in this chapter, companies cannot simply expect to implement conventional MRP better to get the necessary protection and promotion of flow. The first building block of a more agile form of MRP will be explained in the next chapter. This building block will mitigate if not largely eliminate the bullwhip effect by simultaneously addressing both demand signal distortion and material supply distortion by dealing with the core problem driving the bullwhip effect. This building block is called “decoupling.”
