3. Yellow-Yet to Be Received (Y). This is a deeper hole in the buffer that should now be getting the attention of the personnel responsible for managing the buffer.

4. Red-Yet to Be Received (R). This is the deepest hole that we can dig without affecting the drum schedule. This zone alerts the appropriate personnel that if corrective actions are not taken, the drum schedule will be disrupted.

5. Late-Yet to Be Received (DR). The drum schedule has already been disrupted by this work order and it is still not present.

6. Early-Received (LB). The work order is physically present at the buffer resource and ready to be worked on by the drum ahead of the time horizon for which we are scheduled. This usually means that the standards we are using to generate the schedule may be over-estimated (very common since most companies' standards are highly inflated to try to combat Murphy and disruptions everywhere) or the work order was released ahead of schedule.

7. Green-Received (G). The work order was received within the scheduled time horizon with a relatively large amount of time to spare.

8. Yellow-Received (Y). The work order was received within the scheduled time horizon with moderate time to spare.

9. Red-Received (R). The work order was received at the constraint resource within the scheduled time horizon with little time remaining before it is scheduled on the constraint.

10. Late-Received (DR). The work order was received after the time it was scheduled on the drum. By definition, it has caused a disruption to the drum schedule.

In Drum-Buffer-Rope (DBR), "Early," "Late," and "Red" zone arrivals should require a reason code to be attached to the job explaining why this work arrived when it did. In the previous example of a buffer board, the reason code is forced in order for a work order to move from the "Yet to Be Received" status to the "Received" status. A red zone arrival is not necessarily a negative thing-in fact a good system should have approximately 20 percent of its work arriving in this zone-as it is pointing us to the work center with the greatest opportunity to apply improvement focus (i.e. Lean tools) to enable shrinking the buffer and cycle time (more on this in the following section "Metric 2: Local Improvement/Waste").

The reason codes for early and late are essential in removing the variation from inaccurate standards and routings resulting in a more accurate model for scheduling. The beauty of TOC is that it allows any company to start on a process improvement path regardless of the state of accuracy of their routings and standards. Buffers are initially sized to absorb the system's current variation. Entry into the received status of the buffer parts that have inaccurate routings or standards will fall outside the green and yellow zones. These part work orders will be captured in the red, late and early zones with a reason code denoting the standard is wrong or the routing is wrong. This allows a systematic method to correct those parts and remove variation. Ultimately, this allows for more accurate ropes, smaller buffers, and shorter cycle times.

Stock buffers according to the TOCICO Dictionary (Sullivan et al., 2007, 43) are defined as "(a) quantity of physical inventory held in the system to protect the system's throughput." ( TOCICO 2007, used by permission, all rights reserved.) In TOC, these stock buffers also have five zones for management and measurement. Figure 14-7a visibly depicts the typical stock buffer zones.

As you can see, light blue (some authors call this the white zone) depicts a position that is overstocked. Green indicates a position with ample stock, which requires no action. Yellow indicates a stock position that is in its rebuild zone. Red typically means danger or expedite, while dark red gives a visible signal of out of stock (some authors refer to this as the black zone). The total number of parts as well as the total number of days those parts have spent in red "stocked out" and "stocked out with demand" can easily be tracked over time. An example is illustrated in Figs. 14-7b and 14-7c. In both figures, the vertical axis simply represents various part numbers. In Fig. 14-7b, you can see that part 78df has been stocked out 58 days over a 180-day period. Within that 58-day period, for 34 of those days, the part has been stocked out with demand against it (represented by the dark red portion on the right side of the bar). Obviously, it is more damaging to be stocked out with demand. In Fig. 14-7c, you can see that part r643 has been over the limit of the green zone for 30 days over a 180-day horizon. This kind of clear visibility dramatically increases the reliability of a materials/inventory system over conventional tools like material requirements planning (MRP).

The expected behaviors of reliability-based metrics are quite simple. First, localities will perform work in an accurately prioritized sequence since buffer status is a direct reflection of that priority. Second, localities are encouraged to make or buy only what is necessary relative to the buffers. Since these buffers are highly visible, there tends to be little to no conflict about what the real priority is. This is effective Buffer Management (BM). Of course, this assumes that the buffers are set up properly. For more on setting up buffers properly (including placement and sizing), see Drum-Buffer-Rope, Buffer Management, and Distribution of the handbook.

FIGURE 14-7a Stock buffer zones.

FIGURE 14-7b Number of "stockouts" and "stockouts with demand" occurrences for parts over the past 180-day period.

FIGURE 14-7c Number of "parts over green limit" occurrences for parts over the past 180-day period.

Metric 2: Stability

The objective of this metric is to measure the amount of variation that is passed along through the system. A key factor in overall system performance is the amount of variability and volatility that the system experiences and how well that system absorbs or deflects it away from critical areas. In particular, these critical areas are the drums in TOC systems. Encouraging stability at drums is a must. Drums are the anchor point of an overall scheduling system, meaning that all other schedules are planned from the drum schedules. If this is the case, then obviously disrupting the drum schedule creates the effect that all other schedules are out of synchronization with what is deemed to be critical. Disruptions to the drum schedules can also erode their capacity. Drum utilization is defined as a measure (expressed as a percentage) of how intensively the constraint resource is being used to produce Throughput. Utilization compares actual time used to produce Throughput (setup and run time) to available time of the constraint (clock time). Utilization is 100 percent minus the percent of time lost due to the constraint starvation, blockage, and breakdowns. It is critical to measure to know what the overall potential of the system is (see the "Profit Maximization in TOC" section of this chapter) and what a company is leaving on the table every measurement period. This is a dramatically different focus than traditional accounting, which has no mechanism to measure lost opportunity. In reality, there are only a few reasons that cause us to lose potential at drums: 1. Starvation. Starvation occurs when the drum runs out of material on which to work.

2. Unnecessary or over-production. This is a waste of drum capacity on things that, quite simply, are not yet required.

3. Downtime. This is downtime of the drum due to unplanned (Murphy) or planned events.

4. Blockages. Blockages occur when the drum is prevented from running because an operation that it feeds is down. This usually occurs when there is not enough space to queue material between the resources or the resource is actually physically connected to the drum.

5. Poor Throughput rate product mixes. As explained earlier, a key to profit maximization is to make and sell products that produce the most Throughput per time unit on the drum. By making products with a lower Throughput rate, we squander the ability to generate additional cash. There are obvious caveats here as the market may require a company to make a full line of products (each with potentially different rates of Throughput) in order to win any business. (See Chapter 13.) Other critical factors that affect stability and thus should be measured are the amount of non-constraint overloading and the number of late releases. While TOC expects occasional overloads at non-constraints from time to time, it is important to be able to measure the amount of overload that has and is occurring. If it rises above a threshold (specific to the environment) in the aggregate and at the individual resource area, then the system's stability (and ultimately reliability) will be jeopardized as conflicting priorities and expedites rise. A late release is work that is released to the production floor after the scheduled released time based on the rope length tied to a drum or shipping schedule. Late releases exacerbate the non-constraint overloads to which we previously referred.

These measures are necessary to encourage localities to use good buffer management and roadrunner techniques (effective subordination) in order to ensure that work is available to the drum at the scheduled time and that drum utilization is protected. Additionally, it encourages problem solving and improvement initiatives in order to protect and bolster uptime on the drums and potential Throughput rates as well as communication to Sales and Management about those Throughput rates.

Metric 3: Speed/Velocity

The objective of this metric is to encourage areas to pass work on as quickly as possible. The time frame in which a system can respond is often a key factor in winning business and effectively managing capital requirements. The iconic basketball coach John Wooden often told his players, "Be quick, but don't hurry." Localities must be encouraged to perform work with maximum speed and minimal or no sacrifices to reliability, stability, and quality. If accomplished, it means that the buffer positions that these localities feed can be reduced or the system can be more responsive to potential demand. This metric often takes the form of something called cycle time. Cycle time measures the time that released material spends within an area rather than the standard machine or labor process time. By measuring cycle time, a locality is encouraged to enforce the roadrunner rules,10 encourage movement in time rather than batch, limit WIP inventory, limit early releases, and practice good BM. Conventional metrics like Lead Time, Cycle Time, and Stock/Inventory turns can also be used to reinforce this objective.

Metric 4: Strategic Contribution

The objective of this metric is to encourage areas to maximize the Throughput rate and Throughput volume according to the relevant factors of the environment and system. As mentioned previously, the relevant factors have everything to do with the defined constraints or leverage points. Key specific measures of Strategic Contribution will include measuring against a targeted Throughput rate as well as total Throughput. This metric is designed to encourage all areas to be proactive about participating in the generation of the company's opportunities (e.g., innovative ways to in-source or outsource based on market conditions as well as adding free products) or find ways to increase the Throughput rate (e.g., product or tooling innovation) by creating a feedback loop to measure how well we executed against our plan to exploit the constraint. This is simply variance analysis with a TOC twist and has four components: constraint rate variance (time), product mix, volume variance, and Throughput dollar variance.

The Throughput dollar variance is the budgeted selling price minus budgeted variable costs for a product family compared to the actual selling price and actual variable costs for the product family at the budgeted constraint volume.

The product mix volume variance is the budgeted volume of the product family versus the actual volume of the product family sold at the standard Throughput dollar rate for the product family.

The constraint rate variance is the standard constraint rate (planned time on the constraint) for the product family versus the actual constraint time spent on the product at budgeted Throughput dollars (selling price minus variable costs per product).

Variance analysis is not proactive; it is a forensic look at the past so we can understand how we used our constraint and judge our exploitation performance.

Remember, the constraint is the primary area where we are measuring utilization and is discussed under "Metric 2: Stability." While exploitation/utilization of the constraint begins in scheduling, its execution is ensured through BM by identifying effective actions on the shop floor. Having visual loading graphs that clearly show unused/overloaded capacity at the constraint is a proactive tool. The objective is to take actions to sell or make the decision to store the capacity in strategic stock buffers, offload if necessary, or have sales make the call on prioritizing the constraints, workload and communicating changes with the customer.

Metric 5: Local Operating Expense

The objective of this measure is to encourage areas to maximize the local metrics with a minimal or controlled spend. It essentially seeks to measure the amount of money that an area spends in order to convert raw material into Throughput. A local area should be judged against a targeted OE to Throughput generation ratio, which is defined by the relevant range of the TOC economic model demonstrated in Fig. 14-4b. The TOC break-even model is always governed by the impact or lack of impact on the constraint. These local OEs include things like labor, freight, outside processing, contracted or temporary labor, and expedite-related expenses like overtime and premium freight. With regard to this metric, localities will have to balance the level of local OEs with their other critical metrics identified previously. Certainly, a locality should be encouraged to improve flow and velocity with no additional expenditures. Along the same lines, a locality should not be penalized for increases in OE if they improve the ratio (this actually works in concert with its strategic contribution). This hints at a concept called variable budgeting. Variable budgeting allows areas to increase expenditures based on exceeding their relevant range of volume.

Metric 6: Local Improvement/Waste

The objective of this metric is to point out and prioritize lost opportunities. Specifically, it is measuring a locality's ability to identify an opportunity to move the other local and global metrics in the right direction with minimal or no conflict. Essentially, are we asking the right questions and getting the right answers? One very important aspect of determining this is with reason codes. As described previously in the section on Reliability, a buffer system must collect reasons when work orders enter in the red, late, and early zones.

The required transactional data from the execution of BM can be used to direct improvement efforts including Lean and Six Sigma events and capital application. By forcing reason codes when transactions (receipts) are made in certain key zones (late, expedite, and early) of the buffer and comparing them over time, we can get an amazingly clear picture of how to direct improvement efforts. Figure 14-8a shows an example of what this picture can look like.

Figure 14-8b shows some typical types of reason codes for work orders received in the late, red, and early zone receipt and what some potential recommended actions might be.

FIGURE 14-8a Reason code analysis.

FIGURE 14-8b Zone receipts with reason codes.

FIGURE 14-9 A summary of the six general local measurements.

It is important to note that we are directing improvement to the tails of the buffer zones to capture the largest outliers causing disruption and variation. Too early, our cycle time is too long and excess WIP inventory exists. Too late and we incur overtime and premium freight as well as jeopardize our market promise reliability. By focusing investment/improvement on the tails, we can eliminate the sources of the variation and safely shrink our buffers (time, stock, and capacity).

Figure 14-9 shows a summary of the six general TOC local measurements and their respective objectives as well as some specific examples in Operations.

Feedback and Accountability Systems

Now that we have laid down the foundation for a system of global and local metrics that should point the organization and its localities in the right direction with minimal conflicts, there is still one critical piece of the puzzle left to discuss. The APICS Dictionary (Blackstone, 2008, 97) defines a performance measurement system as, "(a) system for collecting, measuring, and comparing a measure to a standard for a specific criterion for an operation, item, good, service, business, etc. A performance measurement system consists of a criterion, a standard, and a measure" ( APICS 2008, used by permission, all rights reserved.) The performance standard can be the accepted, targeted, or expected value.

What is not evident in this definition is the need for steady feedback of system performance and regular adjustment to the actions needed to achieve the standard. The only certainty facing most organizations is that conditions do not stay the same. For example, a shift in the constraint due to changing market conditions or exploitation efforts will result in the need to modify activity dramatically. Without an effective feedback mechanism contained within the measurements, people tend to drive toward the target without recognition that the conditions of the measurement have changed. In other words, it will result in actions that, even though are believed to be the right thing for improving ROI, can actually be hurting the company. The problem is that yesterday the same actions may have been absolutely the right thing to do. Everyone throughout the organization must understand that the feedback mechanism drives measurement away from being "fixed." The problem that people can have in understanding this is that the target can stay constant, but the means to achieve it, and therefore the measurement, may change. BM clearly makes this connection for people. Although on-time delivery to the buffer is the target, very different actions are needed every day dependent on the real-time state of all of the buffers (time, stock, and capacity). Decisions on where to flex labor or where to direct maintenance, quality, or engineering efforts may change according to the status of the buffer.

Though an effective operational planning and control solution is a prerequisite to a proper measurement system, the operational system will fail to properly execute or sustain without an effective way to provide feedback on the current system status as well as help to synchronize decisions and actions.

So, How Is the Operational System Performing?

Two very different, and potentially conflicting, approaches to performance measurement exist for answering this question, although both are important. The first approach is using a performance standard. A set goal or benchmark is provided, which the employees collectively strive to meet over the course of some finite amount of time. For example, "decrease inventory by 30 percent company-wide in the next six months." The second approach is to obtain the everyday pulse with an exceptions feedback mechanism, analyzing the information and deciding if and what action needs to be taken to correct the situation or cause of the exception. The problem is that while both have a place in organizations, they are easily confused. Any growth opportunities will be minimized when they are up against the fixed performance target for employees' attention, unless the connection between the two is clear-which often is not the case, especially in larger organizations.

A performance standard will generally create a status quo that an individual being measured will be satisfied with attaining, often ignoring the other factors that are necessary to the optimization of the whole company ROI. This is not the behavior that the organization truly wants and needs because the standard is usually a subset of one of the five tactical objectives of ROI. In other words, the measure will drive organizational conflict (as discussed in the beginning of the chapter). Therefore, once the feedback system directs attention to the source of the problem, the key is to identify, define, and resolve the system conflict. (See conflict resolution in Chapter 24.) The local metric must be clear, aligned with the global target, deconflicted as much as possible with other measures, and must remain that way. Remember, despite having properly selected productive measurements to begin with if there is not an effective feedback and accountability system providing the current reality and relevance of all local measures to the goal, the system will often be desynchronized and conflicted.

Focusing on Improvement

In contrast to the fixed target, a feedback system does not have an end point but provides continual monitoring of flow to determine exceptions. The regions in BM used to monitor flow are set such that one can respond to an exception and react quickly enough to maintain the desired flow. Additionally, by conducting an analysis of these exceptions and identifying and eliminating the causes of exceptions, a process of ongoing improvement is achieved. Identifying, analyzing, learning, and improving the system is the only way to reach the goal of making the most money now and in the future. By definition, an effective feedback system considers all of the tactical objectives that determine ROI simultaneously rather than any single performance standard because understanding their interdependent nature is a necessity in the feedback system. Companies that understand this thrive on TOC and continue to grow regardless of the economic circumstances in which they are operating. Those TOC implementers who do not understand the significance of BM in providing a process of ongoing improvement will commonly see any improvement stagnate and decline after experiencing initial "brilliant" results and will ultimately end up discarding TOC.

If a manager gains visibility-through BM or any other mechanism-to a potential problem in advance of it affecting performance, this is a great sign that the system is working. Do not mistake this sign with an absence of problems. Companies and the humans who work in them have no shortages of problems. We want those problems to surface when they affect the company performance so that they can be clarified, understood, and resolved. Each problem seen and understood is an opportunity for improvement. Too many individuals of a "fixed" mindset view the presentation of the problem as an indication that the system is not working. To accept that identification of potential problems is vital to the measurement and execution system working effectively is to accept full responsibility and accountability. This thinking is not entirely comfortable for everyone. Without top management understanding and owning this view of the system, there is very little hope that the rest of the organization's management will be able to adopt the "right" mindset.

What Should a Good Measurement System Achieve?

A measure is simply a reading at any point in time of the state of the system relative to the standard the system was directed to execute. It is not used to reward or punish individuals. Cross-functional and interdependent parts of a supply chain can affect the same data but for very different reasons. A fair and productive performance metric will focus and coordinate the efforts of a team, department, process, etc., but a real-time exception feedback is needed to identify exceptions and their causes. Buffer metrics and rules are used to create an early warning system and provide a feedback loop to alert people when and how to act together to get the production system back on plan (to meet market demand). Strategic buffers and BM are used to identify and focus on local improvements most needed for organizational improvement and that have the highest ROI.

It is impossible to separate the measures from the system in TOC because the system is the decision-making tool and buffer status reporting is simply the feedback loop on the health of the system. The key to a successful measurement in a BM system is to generate the "hunger" to identify and learn from the problems. The performance standard (when properly aligned) will resolve itself naturally and should not require constant attention from the individuals executing the plan.

Physicist Niels Bohr defined an expert as "a person who has made all the mistakes that can be made in a very narrow field." Mistakes, problems, and disruptions in logistical systems should never be regarded as negative unless we do not resolve them, learn from them, and ultimately get better. These are opportunities. Managers should strive to be experts in what they manage. If things are running smoothly, the productive manager is going to push the system to ensure that the buffer is "stressed." This mindset will undoubtedly result in very short-term negative blips in buffer performance, but will ultimately trend upward and create the learning and thinking organization necessary for ongoing improvement.

The Key Feedback Information

The TOC information system has five necessary components: 1. Constraint and shipping buffer reporting that includes reason code analysis and constraint rate analysis over time.

2. Replenishment/Actively Synchronized Replenishment (ASR) buffer reporting that analyzes frequency of zone penetration and records stockouts, stockouts with demand, expedites, and the resulting impact on the shop floor.