What are measurement/performance dilemmas? For the purposes of this chapter, let's say that they are situations that pull people, departments, divisions, and companies in opposite or competing directions. For example, it is the purchasing agent who is torn between selecting the lowest cost supplier versus selecting the most reliable supplier; the shift supervisor who waffles on whether to authorize overtime; the salesperson who pleads for an earlier commitment date versus the scheduler who doesn't want to disrupt the schedule; the controller who wants to outsource versus the plant manager who wants to keep the business in-house; the CFO who wants to slash inventories versus the vice president of sales who wants to maintain or even increase inventory; and the engineering manager who wants to standardize products versus the sales manager who wants to sell customized solutions. These dilemmas often represent a constantly changing and frustrating series of daily, weekly, quarterly, and annual set of unsatisfactory compromises. These compromises can cost the organizations tremendous amounts of money as people and resources are whipsawed between two often extreme positions.

What is behind these dilemmas? Commonly these "extreme positions" represent the most apparent or obvious way to meet a particular metric. In our examples, it is purchase price variance versus material availability; overtime budget versus on-time performance; booked business versus schedule stability; product cost versus volume (which can connect back to product cost); cash versus availability/fill rates; utilization versus new business development.

Are the metrics always in conflict? Of course not but often they are. In most for-profit companies, the goal takes some form of return on investment (ROI) or return o average capital employed (RACE). The strategy to accomplish this goal almost always includes tactical objectives (Fig. 14-1) to: 1. Decrease inventory 2. Improve quality 3. Increase sales Copyright 2010 by Debra Smith and Jeff Herman.

FIGURE 14-1 Tactical objectives to increase ROI.

4. Decrease cost 5. Improve due date performance (DDP) Management assumes that improving these five tactical objectives will drive ROI in the right direction. Their assumption is absoltuely valid. The problem is that often the metrics and corresponding actions to achieve these seemingly straightforward tactical objectives will, and do, constantly come into conflict with each other.

Often when a company grows to a relatively modest size, it becomes necessary to segment the organization into areas of functional responsibility (i.e., Sales, Manufacturing, Finance, etc). Therefore it is logical that the tactical objectives are assigned to the functional managers to focus on and improve. However, can a drive to increase quality drive costs up and increase cycle time? Can a drive to decrease costs negatively impact quality and our marketplace? Can a drive to increase sales erode margins? Can a drive to increase on-time delivery or shorten our lead time, increase costs and inventory and erode quality? Can programs to decrease inventory starve the plant and result in decreased on-time delivery and increased overtime costs as well as increased cycle time and work-in-progress (WIP)? In reality the answer to these questions is, "YES!" Each local manager, measured on improving his or her functional responsibility, will drive the organization directly into conflict with itself. This, by definition, is extremely wasteful and limits the organization from having any type of dramatic improvement.

Is there no solution? For years, companies that have embraced the Theory of Constraints (TOC) have proven that there is indeed a solution. When properly aligned in a TOC system, moving all of the objectives in the right direction simultaneously and without conflict is achievable. Figure 14-2 shows the results of a review of the literature by Mabin and Balderstone (2000) of 82 TOC case studies from around the globe. This review showed that companies that implemented TOC were able to move these tactical objectives simultaneously in the right direction.

Do We Measure Too Much?

Once again, are metrics always in conflict? No, but often they are and when they are not in conflict the assumptions around how to achieve certain metrics can put people in conflict. One conclusion we can draw from this is that the more metrics an organization has, the more potential there is for those metrics or the assumptions of how to achieve those metrics to be in conflict.

Modern corporations have metrics everywhere and they devote tremendous amounts of resources and energy to maintaining them. What is interesting is that the number of measures, like the universe, always seems to be expanding (even accelerating). An analogy (with a direct connection to this topic, by the way) is with modern Enterprise Resources Planning (ERP) systems. Ask any ERP provider how many lines of code they had 10 years ago versus how many they have today (they may not even be able to give you a number). The irony or perhaps the lesson is that most of their customers will candidly admit (behind closed doors) that those systems have not really produced any better business results over that 10-year period; they are just more costly to operate. Are we making it harder than it needs to be?

FIGURE 14-2 Average improvement in measures of companies after implementation of TOC. (From Mabin, V. and Balderstone, S. 2000. The World of Theory of Constraints. Boca Raton, FL: St. Lucie Press.) Maybe in trying to control everywhere we end up controlling nowhere. Albert Einstein once said, "Any intelligent fool can make things bigger and more complex, it takes a touch of genius and a lot of courage to move in the opposite direction." He also said, "Everything should be made as simple as possible and not simpler." In these two statements, he cleverly lays out the criteria for effective problem solving and control. Solutions should be elegant, meaning concise and simple, but at the same time, all truly relevant factors must be considered. This is the direction of the solution for resolving measurement/performance dilemmas.

Ultimately, what is needed are measurements that contain a set of relatively simple, highly visible execution priorities to focus and align the entire team of functional managers around actions that have the greatest organizational ROI regardless of the impact on the tactical objectives. In other words, all of the objectives should be understood relative to their current impact on ROI.

The achievement of this solution will reduce the number of primary metrics and the corresponding potential for conflicts between metrics as well as better clarify the actions needed to meet those metrics.

Why Do We Have Measurements?

The point of any system of measurement should be to: Judge progression toward a specific goal or objective.

Drive behavior toward a specific goal or objective.

Highlight relevant factors in relation to achieving the goal or objective.

In a recent USA Today piece by Bruce Horovitz (2009), Douglas Conant, President and CEO of Campbell's Soup said, "You can't talk your way out of something you behaved your way into." Assuming an organization has a defined objective, one of the keys to reaching that objective (in any time frame) is to get all of the components of the system to behave in a manner that moves the system toward the objective. Why is there a link between measurements and behavior? The saying, "tell me how you measure me and I will tell you how I behave," has always been a cliche linking behavior to measurements. It is a cliche because, while true, it is an oversimplification of the relationship between metrics and behavior. Not only is it simply the existence of a metric that drives a specific behavior, it is also the lack of another directly conflicting metric that drives directly conflicting behavior under the same circumstances and a system of feedback and accountability that removes the cliche label from the phrase.

This means that metrics must be coordinated and constructed in order to induce local areas to work together to do what is in the interest of the whole and those metrics must be backed up by a robust system of accountability and visibility (that in itself needs to be measured). This is a basic building block to organizational synchronization and efficiency. If resources are not behaving in a synchronized manner, then some conclusions might be drawn: 1. Formal metrics are not in synchronization. There are potential conflicts or disconnects in the formal metric system. As mentioned previously, the more metrics a system has in place, the greater the chance of conflict.

2. There are no formal metrics or there are significant gaps in the formal metric system. This means resources tend to drive behavior around how they perceive they are measured or believe they should be measured. In the absence of a formal metric, this perception is often driven by a resource's own view of what the right thing is. This creates the opportunity for conflicts driven by interpretation or assumptions.

3. There are formal metrics, there are no conflicts in those metrics, and there are no significant gaps between those metrics but there is no effective feedback and accountability system. Many of us can probably remember coworkers who did whatever they wanted to do regardless of what they probably should do with little or no individual consequences. Additionally, many of us can remember a situation in which behaviors persisted according to a metric that was obsolete. Why was the metric still in place? There was no effective feedback system to point out that it needed to be changed or eliminated.

The question becomes, how do we set up a formal and coordinated system of metrics without significant gaps and conflicts with clear feedback and accountability?

This chapter is organized into three sections. The first section explores the basic global metric dashboard that companies should use to facilitate and judge progression in relation to the goal. The second section outlines the basic coordinated local measure dashboard that supports the global measures. The final section lays out how to build an effective feedback system to drive visibility and accountability in order to better resolve any remaining dilemmas and drive continuous improvement.

Global Metrics

This chapter assumes that a company has a defined goal and strategy. If there is no defined goal and strategy, then why measure anything? See Chapters 17, 18, and 19 in this Handbook for company goals and strategy.

Critical performance measures at the global level ultimately boil down to one basic measure of performance; some form of measure for return on equity. The specific return equation that a company uses is essentially irrelevant. Common measures are ROI, RACE, return on capital employed (ROCE), and return on assets (ROA). Essentially, two components come together to create that equation in whatever forms the company has chosen to use: FIGURE 14-3 The components of ROI.

A measure of profit. Profit can be derived simply by the equation of Throughput (T) minus Operating Expense (OE). Throughput is calculated both at the aggregate and product levels by taking sales dollars minus all direct variable costs. Direct variable cost (also called totally variable cost) is any expense that has a one-for-one direct relationship to the product or service, raw materials, freight, sales commissions, etc. OEs are all of the expenses of a business other than the directly variable cost of the product. This accounting approach eliminates the distortion in earnings between periods when the product produced is greater or less than the product sold by eliminating the allocation of fixed costs to inventory. This approach does not reward the building of inventory that does not protect either current or seasonal future Throughput. This approach better aligns cash flow of the period with the income statement of the period.1 A measure of investment or capital employed. Capital employed has many definitions. In general, it is the capital investment necessary for a business to function. It is commonly represented as total assets less current liabilities or fixed assets plus working capital. In most companies, there are two predominant factors in the investment equation. The first is the total amount of inventory. The second is commonly referred to as Property, Plant, and Equipment (PP&E).

Figure 14-3 shows the global metric hierarchy.

Effectively deploying these metrics, however, requires certain relevant factors to be understood and defined before any company's system of metrics can produce the meaningful and relevant information for good decision-making and a reduction in the number of measurement-related dilemmas.

Measurement is largely executed to judge performance. Unfortunately, even a judgment as seemingly straightforward as profitability could be disastrously flawed. One example is the common valuation of inventory as it relates to profitability within measurement periods. For decades now, we have known about the negative consequences associated with too much inventory-increased costs, decreased flow, damaged goods, etc. However, how do we judge it on the balance sheet? Not only do we value it as an asset, but we "add value" to it as we absorb labor and overhead into the inventory. This means that companies can build inventories and declare a profit without the sales to support it.2 This means that a dramatic attempt to decrease inventory for improvement in flow and Throughput can result in the punishment of these "go getters" for poor short-term profitability performance or low absorption rates. Yet, at the same time, inventory tends to be a critical measurement within measurement periods as well. Trying to balance these competing factors can lead to absurd behavior. We worked with a company a few years back that would refuse the receipt of incoming inventory at the end of every month (a measurement period) only to expedite it a week later and work overtime to attempt to meet on-time shipments. What impact does this have on cost? What is the impact on on-time delivery performance? They knew this behavior was painful to the organization and yet felt their hands were tied by the corporate measurement. Ironically, when you sit down with a CFO or Controller and explain the cause-and-effect logic of what is happening they often scream, "What!? That's not what we want them to do!" Is this a situation of conflicting measures? Yes. Is it a poor interpretation? Probably. Does it demonstrate a lack of an effective feedback and accountability system? Unquestionably yes.

The Constraint Is the Primary Relevant Factor

There is a single factor determining the flow through production and to the market. This same factor gives us the assumptions that underlie the cost and revenue opportunity of any potential action or investment. That factor is the constraint. The constraint could be a resource, raw material or purchased part, a skill set, a policy or procedure, a measure, etc. Information about the constraint is what is relevant. Information on the impact that any option has on the constraint is critical to a measurement system and it should point to actions that, when taken, will provide a bottom-line return.

Constraints change how we judge product profitability, short-term profit maximization, ROI, capital, inventory, and manpower. Constraints impact the rate at which an organization can make money-they are a system's leverage point. This is why having a TOC logistical system and its associated measurements is so vital for operating decisions and improvement in a logistics environment.

TOC is a methodology and a set of processes to maximize a system's ROI/RACE equation by employing solutions that identify, exploit, and manage a system through its leverage points and their interactions with each other.

When considering any alternative actions, plans, or improvement projects: 1. Consider the impact on the constraint's performance. This includes knowing if the constraint will move because of your decision and if so, where and what are the implications?

2. For every investment we must know how the economic return will be generated. Will the market buy more products or will there be any reduction in investment (e.g., strategic inventory buffers) or OE?

Cash inflow versus cash outflow should be one of the primary parameters of decision making. The true implications of any assumptions of the system should be judged against "cash in" and "cash out."

Using the simple performance measures defined previously and removing all other performance metrics prevents companies from making poor decisions3 driven by performance metrics heavily skewed by rolling up labor and overhead into the "cost" of a product. For example, most measures of unit cost would suggest that any reduction in run time or setup time at any resource would result in a lower cost product, or an improvement. TOC dictates that this logic is categorically untrue. Given that this type of localized "cost" thinking is embedded in the vast majority of decision-making tools, the savings or profits generated for most improvement projects are a mirage and never materialize.

Often appropriation requests are constructed to justify investment heavily weighted on reducing cost of the product. For example, if we can reduce the time it takes in one step of the process by 25 percent, then the product cost is reduced (due to less total labor content) and this will translate to an assumed bottom-line improvement. In reality, since these investments and process improvement initiatives (driven by the product cost equation) do not consider constraints or buffers, there is no effective way to judge whether any time reductions at any resource will result in bottom-line improvement. It is quite likely that without the consideration of constraints and buffers, most changes would not result in a net positive ROI position-and often negatively impact performance of the constraint and thus of the whole system as many of these changes require an investment or spending of some sort.

If we push for cycle time improvement at a non-constraint, generally what is the predictable outcome? Remember, a non-constraint means that this part of the system is not currently dictating the pace at which the company is making money. This means that if a non-constraint local resource were enabled to produce faster, the system would experience: 1. No increase in sales or shipments to the customer-no increase in Throughput.

2. A likely increase in parts produced that are not able to be immediately consumed-increased inventory. Increased WIP could also result in longer lead times, decreasing DDP and ultimately reducing Throughput.

3. Some investment was probably made to make the improvement. Potentially, there are additional space requirements or costs of borrowing associated with increased Inventory. In addition, most often the improved rate of production did not allow for any reductions in labor-no decrease in OE.

In other words, locally judging this "improvement" results in Throughput, Inventory, and OE moving in the wrong direction. Judging the potential action on its impact on the constraint would have resulted in saving this money in investment for something that would provide the opportunity for a real bottom-line return.

If one thing can be learned from this book it is the understanding of the impact the constraint has on the system. If the 25 percent improvement in velocity happens to be on a bottlenecked resource, the impact to the bottom line would likely be much greater than the small cost savings that the product cost measurement suggests. In addition to the savings, increased Throughput would result thus improving the organization's bottom-line. Thus, the product cost argument would dramatically understate the need for this improvement.

The discussion above helps explain why so often cost reduction projects approved by senior managers end up having no impact on the bottom line at all, and then why management have such difficulty understanding what went wrong.

FIGURE 14-4a TOC break-even chart of initial profit potential.

Profit Maximizing in TOC

Since most for-profit companies have a goal related to ROI or RACE and profit is a major component of those measures, we need to understand the basic strategy for profit maximization for each company. Remember one of the most relevant factors is the location of the constraint, the company's major leverage point. TOC goes back to fundamental economics as the basis for management accounting information (Horngren et al., 1993, 156) to maximize profit as demonstrated by a quote from a popular management accounting text: The criterion for maximizing profits when one factor limits sales is to obtain the greatest possible contribution to profit for each unit of the limiting or scarce factor. The product that is most profitable when one particular factor limits sales may be the least profitable if a different factor restricts sales. When there are limitations, the conventional contribution or gross margin per sales dollar ratios provide an insufficient clue to profitability.

Figure 14-4a demonstrates the cost and revenue potential of a system through a simple break-even chart.4 Note that the "fixed costs" in the diagram include all OE as defined previously. The "total cost" line is all of the direct variable costs that are added on top of the fixed cost baseline that is associated to the sale of product. This company's revenue potential is determined by the intersection of the relevant range and the total mix revenue. The company profit potential is the revenue potential minus the total cost at any given point above the break-even point.

FIGURE 14-4b TOC break-even chart of volume exploitation.

If top management's attention is focused on investments directly impacting the bottom line as opposed to being distracted by endless requests in the name of localized improvement, then relatively rapid and significant organization improvement is close at hand.

The limitation of the constraint is the key factor that determines overall capacity and therefore the initial relevant range potential in Fig. 14-4a. By identifying and exploiting the constraint-the organization's leverage point-and by applying all brain power and focus to squeezing more out provides a great opportunity to have an immediate and long-term impact to the bottom line for minimal or no investment.

Exploiting the constraint has two levels. The first level is first and foremost about increasing volume. This increase in volume can happen from two primary avenues, both of which require knowledge of the position and status of the constraint or drum. The first avenue is through volume that can be increased by squeezing more out of the constraint itself. This can be accomplished through a number of methods including improvements in its run rate, eliminating its starvation, or minimizing or reducing its setup time. The second avenue that volume can increase is through driving the sale of free products-a product that is free from passing through the constraint. Free product volume must be carefully managed so that it does not create an additional constraint. Careful management often means some sort of governing mechanism to adjust volume in relation to the total system's effectiveness to support the constraint. If there is too much free product volume, it can often mean that the resources involved in making it will have less overall sprint or protective capacity5. This means that they are less responsive to the constraint or (once downstream from the constraint) the customer. The danger of this is obvious. It can cause disruptions to the constraint, late shipments, expedites or bigger buffers (time or stock) impacting lead times, DDP, or cash in inventory. Figure 14-4b shows the impact of volume-based exploitation techniques. Increasing the volume has expanded the relevant range of the system, which translates to higher total potential revenue and profit.

FIGURE 14-5 TOC break-even chart of rate-based exploitation of the resource constraint.

The second level of exploitation is about rate. Now that capacity/volume has been maximized, decisions must be made about which products create more profit relative to that available capacity/volume. The primary metric here is the rate at which products generate Throughput across the constraint. Using our clients as a benchmark over the past 15 years, we have seen the rates of Throughput generation per unit of constraint time by product differ by as little as $3 to $1 and as great as $20 to $1. Of course, free products do not cut across the defined constraint and thus their relative rate in comparison to each other is simply the calculated Throughput margin (Selling Price Direct Variable Costs) per unit of product. Notice this "new" company in Fig. 14-5 has a much lower break-even point and is less at risk in a downturn as well as in the best position to capitalize on a market upturn. This happens because constraint productivity is improved by product selection of the highest Throughput per constraint unit in defining product mix, resulting in an upward thrust in the total mix revenue as seen in the steeper slope of the total mix revenue line in Fig. 14-5.

Of course, these rates can be a moving target as prevailing prices and the costs of material and components can change, which in turn can change the Throughput per unit of product. This requires the use of a pricing indifference model. A pricing indifference model is a tool to show at what point companies become indifferent to which product the limited capacity will be dedicated to producing, for example, product A versus product B as relevant factors for each change. Relevant factors include any significant changes to Throughput rates or capacities. Many companies use a targeted aggregate Throughput rate in the annual budgetary process and judge progress and action around maintaining or exceeding that rate.6 These exploitation techniques provide a simple and level playing field (in replacement of traditional product cost and margin) to assess products against each other on the rate at which they generate cash, the selling price that different products need to capture, and the proper mix to seek in the market for the newly exposed capacity derived from constraint exploitation and free product emphasis. This global performance metric, ROI, helps focus management on the fundamental factors that influence the company's goal and strategy and significantly reduces the number of management dilemmas.

Local Metrics

Once again, metrics need to encourage the right behavior. When dealing with an organization of size and complexity, it always seems to be a challenge to construct a system of local metrics that: Encourages the local parts to do what is in the interest of the global objective.

Provides relatively clear conflict resolution between and within the local parts.

Provides clear and visible signals to management about local progress and status relative to the organizational objectives.

A relatively simple set of six general measurements for localities is given. It is important to note that these local metrics assume that a valid TOC model has been implemented.

1. Reliability 2. Stability 3. Speed/Velocity 4. Strategic Contribution 5. Local OE 6. Local Improvement/Waste Depending on the organization and the functional responsibilities of each organization, these metrics will be translated to very specific forms. Due to space limitations, our examples of specific metrics will be oriented toward Operations only.

Metric 1: Reliability

The objective of this metric is to measure execution compliance to a plan or schedule. When localities (resources, work centers, processes, departments, etc.) and systems are less reliable, it requires systems to hold excessive buffer positions (time, stock, or capacity). Time,7 stock,8 and capacity9 are all interchangeable investments in production capacity. All three are simply stored time. Conversely, when localities can reliably perform within planned time horizons, it reduces the amount of buffer required. This reliability is pivotal to moving global metrics in the right direction. In TOC, reliability metrics are easily implemented by tracking service levels. There are obvious types of service level metrics that are important. Conventional metrics like on-time delivery and fill rates are still very important and relevant. In TOC, however, other critical service level metrics must be installed and tracked. These metrics are performance to time and stock buffers. Remember that resources feed buffers. If those resources are more reliable, it often means that buffers can be reduced. The reduction of buffers is a critical improvement objective in any TOC system.

FIGURE 14-6a Stratification of a time buffer.

With regard to time buffers, this commonly means that early, expedite, and late zone penetrations into the time buffer are noted and managed at two levels-first, to direct execution actions to keep the constraint and delivery schedule stable; and second, to capture information that identifies the source of variation for future improvement activities to increase system stability. Figure 14-6a shows the stratification of a time buffer into different zones ranging from an early (far left, sometimes referred to as the light blue (LB) or white buffer zone) to expedite (red), to late delivery (far right, sometimes referred to as the dark red (DR) or black zone). The general zone color designations are provided in the figure. Notice there is a released work order (48709-01) that has not entered the time horizon that the buffer represents in front of the drum. One of these work orders will enter 9 hours out. If this facility works 24 hours a day, then that entry will occur at 10:00 A.M. on Wednesday.

We will have to reconcile the work order's actual presence in the buffer by recording when it entered the buffer and judge that against its scheduled entry to create a view about what, if any, corrective actions need to taken. When a work order is not ready and in the buffer at the start of the green zone (scheduled buffer entry), a penetration is created in the buffer. This hole can be caused, for example, by missing materials, tooling, specs, etc. The severity of this penetration will ultimately determine when we have to act and the priority of work orders on which we have to act.

This means that we have to think about the five zones from two perspectives, "Yet to Be Received" (at the drum) and "Received" (at the drum). These are the two situations that can occur. When something is "Yet to be received" the clock is still ticking on the time it has to travel to the constraint. When something has been "Received," the hole has been filled. In Fig. 14-6b, a real-time buffer board that reconciles a released work order against their buffer status is shown.

Notice that when we account for the same time horizon from the two different perspectives, it actually creates 10 status zones. Those zones are: 1. Early-Yet to Be Received (LB). This zone actually represents all released work orders that are on the way to the buffer.

2. Green-Yet to Be Received (G). This is a hole in the buffer. Not a serious hole, but a hole nonetheless.

FIGURE 14-6b Reconciling released orders against buffer status.