Different definitions of terms undoubtedly have caused much confusion. While it may be impossible, at this late date, to change TOC terminology, people trained in accounting call Throughput by the name contribution margin,24 with the same definition-revenue minus totally variable costs. Inventory is a highly controllable subset of Investment in total assets. Operating Expenses, in accounting terminology, would be fixed costs, including manufacturing fixed costs and general, selling, and administrative fixed costs.

Where to Focus Quality Improvements

The same example shown in Fig. 13-1, along with identification of the internal constraint (Resource 2) can be used to focus quality improvements. Assume the company experiences a scrap problem at Resource 4 resulting in 4 percent of units (3.6 units) of Product X being scrapped, 7 percent (0.7 units25) of Product Y, and 8 percent (6.4 units) of Product Z (see Fig. 13-2). The quality team can correct the problem on only one product at a time, the cost being approximately equal for each product. Which product should they work on first?

FIGURE 13-2 Scrap problem at Resource 4.

Provided the quality problem data shown in Fig. 13-2, most people immediately select Product Z for first attention because (1) it has the highest percentage of scrap, (2) it has the most units being scrapped, or (3) it is the company's most profitable product. With traditional accounting, even if the cost of the time lost (45, 35, and 40 min, respectively, for Products X, Y, and Z), at $0.50 per minute26 is included in the analysis, along with the cost of materials ($60, $50, and $45 for Products X, Y, and Z, respectively) and variable manufacturing overhead ($8, $5, and $2 for Products X, Y, and Z, respectively), the priorities remain Product Z (with a total cost of $428.80), then Product X (total cost of $325.80), and then Product Y (total cost of $50.75).

Because there is sufficient time to replace the work lost on Resources 1, 3, and 4, however, the Throughput approach includes lost variable costs (materials and variable manufacturing overhead) plus the cost of lost time only on Resource 2. Units lost of Products Z and X will be replaced, resulting in fewer units of Product Y being produced and sold. First, we safely may ignore the option of fixing the quality problem on Product Y because only 1 unit is lost per week, leaving only Products X and Z for improvement consideration.

Since only whole units may be sold, if the quality problem on Product Z is corrected first, Product X's remaining quality problem will result in 4 fewer units of Product Y being produced and sold. Product X production losses will be replaced, but 80 min (4 unit 20 min) on the constraint (Resource 2) will have been lost. Thus, after Resource 2 has been used to produce 80 units of Product Z (still the best product, requiring 400 min) and 90 units of Product X (the next best product, requiring 1800 min), leaving only 120 min to produce and sell 6 units of Product Y, resulting in operating income of $12,126.

However, if the quality problem on Product X is eliminated first, Product Z's remaining quality problem will result in 7 units having to be replaced, at 5 min per unit, meaning 35 min of Resource 2's time will be lost, leaving 2,365 min available. Once again, 80 units of Product X, requiring 400 min on Resource 2, and 90 units of Product Z, requiring 1800 min on Resource 2, leave 165 Resource 2 minutes available for the production and sale of Product Y. With 165 min available, 8 complete units of Product Y [(165 min)/(20 min per unit)] can be produced and sold. With this product mix, operating income will be $12,492, $366 higher than if the quality problem is fixed first on Product Z.

While looking only at the relevant resource minutes lost on each product, multiplied by Product Y's Throughput (contribution margin) per minute of Resource 2 production time, plus the variable costs involved will provide the same answer (fix the quality problem at Product X first, then go to Product Z's problem). This shortcut method (see the Throughput Examples spreadsheet, AR29: AZ48) risks failing to consider all variables. The total effect on operating income (shown in cells AQ50:BB69 in the Throughput_Example spreadsheet) is by far the safer method and the one recommended by experts.

Adding a New Market Segment

The Throughput solution often has been compared with a linear programming solution of a one-constraint problem (Dopuch and Birnberg, 1969) and has been described as a step-wise linear programming analysis, dealing with one constraint (the worst one) at a time. Unfortunately, the Throughput solution, like a linear programming solution, is extremely sensitive to deviations from an equilibrium solution, one in which a best solution is found using current assumptions concerning resource availabilities, product demand, and so forth.

For example, suppose a salesperson returns from China with an order for 30 units a week for each of the three products, X, Y, or Z, or any combination thereof, with agreed selling prices equal to 80 percent of the U.S. prices. Should the company sell any of its products in China? Facing this decision, the company must be very careful not to make the easiest of mistakes: assuming the constraint will not shift to another resource.27 After computing Throughput per minute of Resource 2 for each of the three potential China products, suppose the company decides to sell 30 units of Product Z in China (called China Z in the spreadsheet) with a Throughput per minute of Resource 2 of $18.80 ($156$62 = $94 5 min on Resource 2), prior to filling orders for Products X and Y, and will not be interested in selling Product X (China X, with a Throughput of $7/min) and Product Y (China Y, with a Throughput of $6.55/min) in China. Following this strategy, however, will cause the company not to make a higher total profit ($14,214), as it expects, but to make $12,448-$1,766 less than expected-and $410 less than its previous best performance with no sales to China. The deterioration in operating income will occur due to Product Z's (and, therefore, China Z's) high usage of Resource 1, causing it to be in tighter supply and resulting in an interactive constraint with Resource 2. (See "Throughput_Example" spreadsheet, cells BD2: BS82.) Controls should be in place to prevent actions that will reduce operating income. The following examples illustrate how traditional accounting can lead to nonoptimal decisions.

Purchasing Decisions

Even though materials are not often an organization's constraint, rapid expansion in 2007 and 2008 saw raw materials prices skyrocket. Of course, the recession in late 2008 and 2009 brought material costs back in line. When materials prices change, Throughput and Throughput per unit also change. Therefore, product priorities also may change. In a TOC world, any time any Throughput metric input changes, its impact on priorities must be computed.

Less obvious purchasing decisions involve opportunities to acquire materials from a lower-cost supplier or to outsource certain portions of the productive effort. Potential acquisition errors can occur based on both accepting and rejecting outsourcing proposals as well as on initial material purchases. Each of the following decisions should be considered independently. That is, the starting point is the current most profitable combination of 80 units of Product Z, 90 units of Product X, and 10 units of Product Y.

Acquisition Decision Purchasing has found a new supplier who is willing to provide Raw Material #7, for $2.50, saving the company $200 a week. Figure 13-3 illustrates this opportunity. If Purchasing primarily were evaluated based on cost savings, they would like to make the deal. After trying a sample of the new material, however, the production manager states that Resource 4, Task 3 will incur approximately 10 percent scrap. Since Resource 4 has plenty of idle time, Purchasing assures the production manager, they can easily make up the 80 min lost due to scrap. Further attempting to seal the deal, the purchasing person tells the production manager that since Resource 4's utilization will increase, efficiencies may increase, offsetting any scrap.

The production manager, having been trained in TOC concepts, states (not too patiently) that since the scrap occurs following processing on the constraint (Resource 2), each lost minute on Task 4 means that fewer units of other products can be produced and sold. Additional materials will be processed to make sure all demand for Product Z will be filled, so the lowest priority product, Product Y, will take the hit. Because Resource 2, Task 4 requires 5 min of processing time per unit, 40 min of Resource 2 time (8 units 5 min) will be lost. Because Product Y requires 20 min per unit on Resource 2, two units of Product Y will be eliminated (at $183 Throughput per unit). Therefore, this material cost "savings" of $200 will cost the company $36628 in lost Throughput every week! (See the original "best" operating income versus the operating income if the proposed change were accepted in the Throughput_ Examples spreadsheet, cells BU1: CF40.) FIGURE 13-3 Proposed acquisition of a cheaper material.

FIGURE 13-4 Product X, Resource 1, Task 1-make versus buy decision.

This "opportunity," if accepted, would result in a decrease in operating profit of $166 ($366 $200) each week. Fortunately, the production manager rejected this "cost-saving" proposal.

Outsourcing Proposal #1 Suppose another person in purchasing has received an offer from a new supplier to provide a component that will include Raw Material #1 and processing through Resource 1 for a cost of $21.75. No variable overhead is incurred for this operation. Should the offer, shown in Fig. 13-4, parts a (make) and b (buy), be accepted? Traditional accounting would say, "Absolutely." The unit cost through that point in production is $22.67 ($20.00 for the material and $2.67 for 5 min of processing at a cost of $0.5333 per min); resulting in a savings of $82.80 each week ($0.92 a unit times 90 units) or over $4100 each year.29 However, TA would respond, "No way!" Resource 1 is not a constraint and already has 50 min of unused time each week.30 Accepting this outsourcing offer would result in incurring an additional cost of $1.75 a unit ($21.75 $20) for the 90 units needed-$157.50 each week31; almost $8000 a year. Meanwhile, Resource 1 would incur additional idle time of 450 min each week. In addition, the company would lose direct quality control and incur the risk of unavailability of the component when needed.

FIGURE 13-5 Products X and Y, Resource 1, Task 2 and Resource 2, Task 3-make versus buy Of course, traditional accounting would respond that Resource 1 should be put on a 4-day workweek since they now have over 8 hours of idle time. Sometimes this makes sense, but not normally. Cutting the pay, effectively, of one worker does not inspire high worker morale and job commitment. In addition, Resource 1 is the most likely constraint candidate should Resource 2 capacity's be elevated.

Outsourcing materials and Resource 1 work, in the situation described, would not be a good decision.

Outsourcing Proposal #2 Purchasing also has an offer from a supplier to provide a component that would include Raw Material #3 and processing by Resource 1 and Resource 2 for a cost of $40. Variable manufacturing overhead for the operations involved is $2.50 a unit. Figure 13-5a shows the current arrangement, and Fig. 13-5b shows the "buy" proposal. Should the company accept the offer?

Traditional accounting would say, "No." The cost to make is only $29.97 ($20 for the material, $7.47 for the labor-14 min at $0.5333,32 plus $2.50 variable overhead). Buying the proposed component would increase the company's costs by over $10 per unit and $100033 each week.

Rather than comparing costs, however, a person who is aware of TA would look at the impact on the company's Throughput. Resource 2, Task 3, requires 10 min. Recall that the constraint in this system is Resource 2. Saving 10 min of Resource 2 time on Product X (90 units) and Product Y (10 units) amounts to 1000 extra minutes. With the extra time, additional units can be produced for the unfilled demand of Product Y. All unfilled demand for 40 units of Product Y can be produced and sold, adding an additional $183 per unit for a total of $7,320 additional Throughput, and adding $4520 to the bottom line ($7320 $2800-the added cost of outsourcing 140 units34 at $20 incremental cost). Compared to the example company's previous best performance of $12,858, this change represents about a 35 percent increase.35 Even if the company incurs additional overhead to track the supplier's quality and dependability, the outsourcing (buy) offer should be enthusiastically accepted. In addition, everyone in the company should be made aware of the fact that marketing is now the organization's constraint and management should formulate plans to increase the capacity of both Resource 2 and Resource 1 when product demand increases.

FIGURE 13-6 First engineer's engineering change proposal.

Engineering Change Proposals (ECPs)

Engineers have been studying production operations and two new engineers have submitted engineering change proposals (ECPs).

First Engineer's ECP A young engineer has read about a new process that can reduce the time on Resource 1, Task 3, from 15 to 5 min (see Fig. 13-6). New tooling costing $5000 would have to be acquired, however. Should the proposal be accepted?

Traditional accounting typically would value this opportunity as favorable since saving 10 min on 80 units would produce a savings of 800 min. At the applied labor rate of $0.5333 (or the actual rate of $0.50), cost savings would total $426.64 or a minimum of $400 a week. Thus, payback would occur in $5000/$400 = 12.5 weeks, at the longest. This is a very quick return on investment (ROI).

Of course, traditional accounting information provides support for an incorrect decision. Resource 1 is not the constraint and the "cost savings" of $400 or more per week would never occur. Resource 1 would just have more idle time and the company would be out $5000 for the tooling. TA correctly and immediately would reject this proposal.36 Because Resource 1 might someday become a constraint (it has the highest loading after Resource 2, the current constraint), this proposal might be kept on file for action later, but not now.

FIGURE 13-7 Second engineer's engineering change proposal.

Second Engineer's ECP Another engineer has submitted an ECP affecting Product X and Product Y. Figure 13-7 shows three changes: (1) an increase in materials cost from $20 per unit to $30, and (2) a 2-min decrease on Resource 2, Task 3, and (3) an increase from 4 min to 9 min for final processing of Product Y by Resource 4, Task 2. Oh, and the change will require an additional $8000 investment.

Of course, the accounting department is shocked by the $10 increase in material cost and the 3-min increase in net processing time for Product Y (from 35 to 38 min) that is partially offset by the 2-min decrease in Product X processing time. Considering the additional investment required, accounting might even suggest that this engineer should work for a competitor.

By now, you know that since Resource 2 is the company's current constraint, this change should be evaluated further using TA concepts. Saving only 2 min on Resource 2, Task 3, for 100 units (90 for Product X and 10 for Product Y) means an additional 200 min of availability but costs an additional $1000 in raw material cost ($10 per unit 100 units for the original quantities) in addition to the $8000 investment. With this additional time, however, more units of Product Y can be produced and sold. How many additional units are possible? Not 10 (200/20), but 11 (200/18). With an original Throughput per unit of $183, 11 additional sales of Product Y will bring in $2013 each week. This amount, minus the $1000 in additional cost for Raw Material #3, means an additional $1013 in operating profit. Resource 4 has sufficient available time not only to use 9 min of processing time on the 11 additional units, but for the entire market demand of 50 units. The payback period for this investment would be $8000/$1013, or less than 7.9 weeks. Assuming the cost of the investment can be amortized over 52 weeks, operating income will increase by $3986 each week.37 This is a great investment, but it would have been turned down using traditional accounting metrics!

The Problem of Identifying Decision-Relevant Costs and How to Avoid a Disaster

Throughput (contribution margin), Inventory (investment), and Operating Expense (fixed costs) changes are always relevant. However, it is extremely difficult to accurately select the relevant costs and revenues (including those associated with lost opportunities) of many, if not most, management decisions. For example, when multiple changes are occurring to more than one element of a process (the second engineering ECP is an example), keeping everything straight for a correct analysis can be difficult. The advice provided throughout the book, Supply Chain Management at Warp Speed (Schragenheim et al., 2009), is to consider any change (product mix, investment, make versus buy, special orders, rationalization of product lines, etc.) in terms of its impact on total amounts of Throughput, Inventory, and Operating Expenses (contribution margin, investment, and fixed costs, in accounting terminology). This same advice, couched in terms of the dangers of allocating fixed costs, is included in virtually every cost and management accounting textbook (see, for example, Hilton, 2009, 600601, 612; Garrison et al., 2010, 588589), and should be followed without exception to avoid costly errors.38

Inventory Changes and GAAP Accounting

Basic goals of TOC are for Throughput to increase, Inventory to decrease, and Operating Expense to decrease. Throughput increases and expense decreases will be reflected favorably on external reports that conform to GAAP. Inventory reductions, however, will be reflected unfavorably on GAAP statements by reducing both assets and operating income. Therefore, inventory reductions should be handled with special care.

Because some accounting and other people have trouble understanding exactly how reducing inventory results in decreased income, I have developed several examples over the years to validate this result.

For example, assume a company that has no beginning inventories of WIP or finished goods, produces 20,000 units and sells 15,000 units for $20 each. There is no ending WIP inventory. Budgeted costs (as traditionally prepared) include the following: A traditional (absorption costing) income statement and a Throughput (variable or direct costing) income statement (both assuming costs are the same as those projected) are shown in Fig. 13-8.

As shown in Fig. 13-8, the traditional income statement shows net operating income of $71,250, while the Throughput income statement produces net operating income of only $51,250. The difference of $20,000 ($72,250 $51,250) can be reconciled solely by the change in inventory fixed costs. That is, the increase of 5000 units in finished goods times the fixed manufacturing cost of $4.00 per unit equals the $20,000 increase in traditional (GAAP) income over the Throughput income of $51,250.

FIGURE 13-8 Traditional and Throughput income statements.

A more detailed example involving materials and WIP changes, as well as changes in finished goods, can be found in a spreadsheet entitled "InventoryReductionExample" located at www.mhprofessional.com/TOCHandbook. In this example, everything other than WIP inventory and finished goods (FG) inventory are held constant over a 4-year period. Year 1 sets up a baseline performance, the resulting income, and balance sheet. With no changes in inventories (Year 1), both methods (traditional and Throughput) produce the same net income before income taxes. WIP and FG inventories then are reduced by 50 percent in Year 2. Three spreadsheets are included in this file: "GAAP Accounting," "Throughput (Variable) Accounting," and "Reconciliation of GAAP & T. Inc."

In the example, a normal year (Year 1) income is $200,000 and return on sales is 10 percent. The 50 percent WIP and FG inventory reductions occur in the second year (Year 2) and GAAP income drops to $66,667 and return on sales drops to 3.33 percent. (See the "GAAP Accounting" spreadsheet, cells A48:Y80.) The third year of the example, GAAP return on sales recovers to 6.67 percent (cells A83:Y113) with income of $133,333, but is not back to the full 10 percent (income of $200,000) until Year 4 (A116:Y147). The reduced income occurs because, with traditional (GAAP) accounting, WIP contains a portion of fixed manufacturing costs, depending on the percent complete, and FG contains its fair portion (100%) of full fixed manufacturing costs. In the environment established in this example, the only way to reduce inventory is to cease entry of raw materials into the system.39 The decrease in production activity required to lower inventories means that all fixed costs of the current period, plus the fixed costs in units in beginning FG and WIP, are charged to cost of sales in the inventory reduction year.

The TA approach, shown on the second spreadsheet of the "InventoryReductionExample" shows that income and return on sales for the entire 4-year period remain constant at $200,000 and 10 percent, respectively. The third spreadsheet in the example file reconciles GAAP and Throughput income for the year of the inventory reduction and suggests general journal entries to adjust from internal Throughput reporting to external GAAP statements. Table 13-4 shows the reconciliation followed by the general journal entries.

Because the inventory reduction is permanent, other things remaining equal, reported GAAP income would remain $200,000 less than that reported under TA.40 Given that this inventory reduction permits the opportunity to increase future earnings (lower WIP means faster processing that permits additional production with no increase in fixed costs), the potential "sacrifice" in reported earnings is necessary and must be undertaken. Careful planning and communication with relevant stakeholders, especially employees, creditors, and owners, can minimize potential negative effects.

Table 13-4 illustrates how the difference between GAAP income of $66,667 and Throughput income of $200,000 in Year 2 (a difference of negative $133,333) may be explained totally by the change (reduction) in fixed costs in beginning and ending WIP of $53,333 plus the change (reduction) in fixed costs in beginning and ending FG of $80,000.

Following Table 13-4 are all the year-end adjusting general journal entries to convert all income and balance sheet accounts from Throughput to GAAP. This example clearly indicates that keeping accounting records using Throughput concepts during a period quite easily can be converted to GAAP accounts at the end of a period.

Difference explained by change in fixed costs in inventories: TABLE 13-4 Reconciliation of Traditional (GAAP) Costing Operating Income and Throughput (Variable) Costing Operating Income for Year of Inventory Reduction *

End of period adjusting entries to convert from Throughput to GAAP accounting:

Value Metric Used to Track Performance

To provide timely feedback to managers and operations personnel, TOC has some unique metrics that reveal both what should be done and what should not be done. These metrics support standard TOC policies and are designed to encourage appropriate behavior. Many use a TOC concept called value days that aggregates the value of amounts invested or delayed using the following formula: Vn= Vn-1+ Value where Vn = the value for a current time period (e.g., day or week) Vn-1 = the value for the previous time period Value = the total net value ($ in the U.S.) invested or realized in the current period (e.g., days, weeks, months) This formula basically says that every amount of currency invested for a day results in a lost opportunity to use that amount for some other purpose, and every day the amount is not recovered repeats the lost opportunity. Therefore, the value is not fully recovered until sufficient amounts have been received to cover the entire deficit. The basic idea is to age investments analogous to the aging of accounts receivable.

Inventory Value Days For example, if $100 is invested in inventory on Day 1 and the inventory is not sold until Day 10, inventory value days would equal $100 10 days or $1000. In this way, slow moving inventory is highlighted and the information provided encourages the quick sale of older inventory. It also provides a way to project demand saturation so that additional units are not manufactured or acquired. With thousands of products, the process becomes more complex, but can be accomplished either with enterprise software or with spreadsheets. A one-product example that includes acquisitions and sales over a period of 33 days ("VALUE-FORMULAandEXAMPLES.xlsx," second spreadsheet, "Inventory Example") may be found at the following Web site: www.mhprofessional.com/TOCHandbook. Figure 13-9, based on the spreadsheet example, shows the difference between the absolute amounts invested, as traditionally recorded (dark color) and inventory value days (lighter color). The rapid growth in inventory value days beginning on Day 23 signals management that inventory is building too rapidly and gives advance warning to cut acquisition of this item, which happened on Day 29. While the increase also is signaled by the traditional inventory values, upon close inspection, it is not nearly as noticeable and dramatic.

FIGURE 13-9 Traditional inventory value versus inventory value days.

Throughput Value Days The same file ("VALUE-FORMULAandEXAMPLES.xlsx") has a third spreadsheet, "Throughput Example," that illustrates how the same formula may be used to track orders that are delayed (missing from the critical area of a buffer for quality or other reasons or returned by the customer due to a quality problem).

Transfer Pricing

Organizations regularly go through restructuring programs where they go from a centralized organizational structure, where divisions are cost centers and major decisions such as product selection, pricing, and investments are made at headquarters, to a decentralized structure, where divisions are profit centers or investment centers. As their names imply, a cost center is responsible only for controlling costs; a profit center has responsibility for generating revenues and controlling costs, and investment centers are encouraged to behave as entrepreneurs and are responsible for making investments, generating revenues, and controlling costs. Organizations also regularly go through the process in a reverse direction: from decentralized profit and investment centers to centrally controlled cost centers.

Top management would like to control all decisions (a centralized structure), but they realize that they are too far from the action to act sufficiently to quickly compete in a meaningful way. However, when organizations are decentralized, incorrigible transfer-pricing issues arise if any products are transferred from one division to another, which they usually are.

Additional complications arise because division managers usually are in competition with each other. The selling division wants a higher price; the purchasing division wants a lower price. Transfers between divisions in different countries bring complex tax issues into the mix.

Addressing transfer pricing in detail is beyond the scope of this chapter. However, if you are interested in a nice treatment of transfer pricing in a TOC context, without international transfers, I recommend, "The 'Transfer Prices' Problem," in Approximately Right, Not Precisely Wrong (Eden and Ronen, 2007, 241258).

Other TOC Metrics

An ideal TOC operation has minimal fire fighting and expediting, no chaos or brutal overtime, and sufficient orders with reliable promise dates entering the system. In order to accomplish this desirable environment, it is critical that sufficient raw materials are on hand,41 minimal (but sufficient) WIP and FG inventories are held, orders are delivered on time, appropriate buffers are established, buffer penetration (consumption) is tracked, and performance is continuously improving. In addition to the previous metrics, some general measures that organizations have found useful include the following: On-time deliveries-1 minus the ratio of late orders, weighted by days late, divided by total orders.