Throughput Economics Chapter 3: Throughput – a key concept pointing to a sensible direction

Bake-T, the same company as the leading case of chapter 1, receives an offer to sell 800 donuts every day to a children’s hospital. Instead of paying the regular price of 78 cents per donut, as other customers do, the hospital offers to pay only 70 cents, a total of $560 per day.

Assuming that the additional 800 donuts do not cause any capacity problems for Bake-T how will this additional deal contribute to the bottom-line of Bake-T?

Baking 800 donuts requires materials, bought from suppliers. Let’s assume that $220 is paid to suppliers for all the ingredients.  The cost of labor, equipment and space do not change due to the extra deal.  Some minor energy costs are included in the $220 sum, which comprises all the variable costs.  The difference between the selling price minus the variable costs, $560-$220=$340, is attributed to the added value generated by Bake-T in this deal.  While the common accounting methods recognize this difference, it is usually called the ‘contribution’ of that particular deal, even though ‘contribution’ includes direct labor, which is not truly variable in most cases.  ‘Marginal costing’ is a cost accounting method that uses the contribution number for decision making.  However, marginal costing is harshly criticized because of its focus on local and short-term considerations only.  It is argued that considering just the variable costs would influence sales people to give big reductions in price, with the unavoidable result being the inability to cover fixed-costs, which are not considered at all in the marginal-costing calculations.

The case of the drugstore chain in chapter 2 considers the accumulation of the margins of three brands. In Retail, the margin is usually based on the selling price minus the cost of purchasing an item.  As with the Bake-T example, no other costs are involved, so the margin is the added-value of selling one item.

Throughput (T) is a key concept in the Theory-of-Constraints. It is similar to ‘contribution’ but deals with the concerns regarding marginal costing, so that the overall approach and usage are wider than in marginal costing.  Throughput was defined by Dr. Goldratt (The Goal, The race) as:

The rate at which the system generates money through sales

This definition fits all business oriented organizations. Non-profit organizations should state “goal units” instead of “money”.

The definition refers to the whole system, usually an organization, and this marks a basic difference from “contribution”, which focuses on a specific deal or item. While it is possible to measure and monitor the aggregation of all contributions in a period, this is not the common approach. In few large organizations the aggregate contribution of a particular business unit is measured, but it is not a prime performance measurement.

Instead, the closest common measurement is ‘gross profit’, which usually includes additional operational costs that do not vary with every sale.

The emphasis on added-value is a breakthrough insight. Of course, the total expenses that do not vary with every sale, and the capital of all the assets, should also be considered when assessing the performance of the organization as a whole.  However, the trend of throughput (T) through time is the most important holistic measurement. It gives an excellent indication of the direction in which the business is heading.

The above definition of T requires interpretation. “Rate” refers to the holistic throughput measured for time periods, like a month or a year.  Generating money is the periodical difference between revenues and the truly-variable-costs (TVC) attributable to the revenues.  The adjective “truly” emphasizes that those costs occur with every sale, even the smallest one.  Thus, unless labor is paid by piece or by hours, the cost of labor is not included in the calculation of throughput.

The monthly T for Bake-T is the summation of all sales in that month minus the costs that vary with every sale. Assuming ongoing activity of Bake-T, the main truly-variable-costs are the materials used for producing the products Bake-T sells.

T can be applied to a specific item as long it has a selling price. It is also easily applied to a deal. When this is done, the T is not per time, but per item or deal. The meaning of T-per-item is that any additional sale of that item increases the monthly and annual T by that amount.  Thus, accepting the order from the hospital Bake-T increases the monthly T by 30*$340 = $10,200.

Throughput (T) is NOT the profit!  The total T is not the profit and the T-per-item is not “the profit of an item sold”.  In order to generate T the organization has to spend costs that are not truly variable. These costs are denoted Operating-Expenses (OE).  On top of those costs is all the money invested by the organization (I) or the money held within the organization in assets or capital.

The connection between T, I and OE and the regular financial performance measurements of the organization is simple:

Net-Profit (before tax) = NP = Revenues minus Expenses.

The expenses can be split into two parts: TVC and OE.  So, we get the following connection:

NP = Revenues – TVC – OE = T – OE.

The money held in the organization, I, is part of the return-on-investment (ROI):

ROI = NP/I = (T-OE)/I.

This simple connection raises the question:

What is the advantage of using T, I and OE over the common NP and ROI?

The simple answer is that by separating the truly-variable-costs (TVC) from all other expenses, and setting them off against the revenue, relevant focused information on the full impact of sales is revealed, which is critical for supporting better decisions. The big difference between marginal costing and T, I and OE is that full attention is given to the total T, rather than just on individual opportunities.

There are two reasons for the special power of the T, I and OE information:

  1. Throughput contains the net information about sales: the overall contribution to the organization.   There are many decisions that impact only sales and do not cause any change to the capacity costs. In such cases the delta-T, the change in total T due to the decision, expresses the net economical value of the decision. This is demonstrated by the decision of Bake-T to sell the extra 800 donuts at a reduced price, but without any impact on OE, because no overtime or any other additional capacity is required, as assumed in this case. The only other impact that might be problematic is if the deal would somehow impact other sales, like causing other clients to demand price reductions. Such effects are caused by internal dependencies within sales, so the question is still focused on the impact of the deal on the total throughput. So, the clear distinction between Sales and OE allows focused questions to reveal the information relevant for making the decision.
  2. The role of OE becomes much clearer after the definition of T. OE contains all the expenses that do not vary with every single sale, but are required to maintain the available capacity of the various capabilities needed for various sales. Every single expense in the OE is about capacity. For that matter the CPA and banking services are part of the capacity of required capabilities for generating throughput, even though these services are very generic and not influenced by specific transactions or items. As discussed in chapter 1 the behavior of the total cost of capacity is not linear, it is the accumulation of step functions. This fact clarifies the behavior of the OE when sales go up or down. Certain changes in the sales might not cause any impact on the OE and then the decision depends only on the additional throughput. In other cases the impact on OE could be very significant, depending on the new steps of one or more resources. In order to estimate delta-OE there is a need to identify the resources for which the capacity must be increased above the current available level and the cost of those increases. It is also necessary to check the feasibility of increasing the capacity in the given time frame.

As a general comment the usual “Revenues minus Expenses” approach, without distinguishing total variable cost and OE, does not help to determine the impact of an increase in revenues on the profit. This lack of confidence often leads to focus on known and easily identifiable expenses, with the hope of not causing any damage to the revenues. On the other side, T and OE come with easy-to-use method how to approach and calculate the net impact of the bottom line.

The basic formula, NP = T-OE, makes it much easier to evaluate the net impact of a decision on the bottom line: delta-NP = delta-T – delta-OE.

Delta-T is the outcome of the sales analysis, including all the internal dependencies within the sales. Recalculating the T after the changes and subtracting the total T before the changes yields delta-T.  Of course, delta-T can be calculated directly by considering just the additional T minus the T lost due to the proposed changes.

Delta-OE is the net change in the gross OE, considering all the true changes in the cost of capacity due to the changes in the sales.  When the resulting changes still leaves spare capacity on all resources then delta-OE = 0.

Considering the basic example assume now that there is another hospital, which currently buys 300 donuts for the regular price of 78 cents. It could be expected that this hospital will now insist on paying the same price as the other hospital of only 70 cents per unit.

delta-T = added throughput of the new deal minus the lost throughput of the old deal = ($560-$220) – 300*(.78-.70) = $340 – $24 = $316.

Suppose that producing the additional 800 donuts every day would require 5 hours of overtime, at the cost of $20 per hour. In these circumstances, accepting the deal would add delta-OE of $100 every day.  Under these assessments, the deal of the hospital would add to the current profit of Bake-T:  $316 – $100 = $216 per day, or 365*216=$78,840 net profit for the whole year.  Practically it means that under the above assumptions this is a good deal.

Is it possible that the same deal would cause a loss due to causes that are not part of the deal itself?

A lack of capacity could have the effect that a deal that generates good delta-T could eventually cause a loss.  The lack of capacity is caused by all the other existing and potential demand plus the considered deal.  It is enough that one resource would lack capacity to cause substantial delays in delivery to clients. If this shortage of capacity persists then loss of market demand is inevitable.  So, possible lack of capacity creates dependencies between the deal and existing sales that is caused by the interaction between sales and operations.

What can be done when one resource lacks capacity? One answer is to do whatever it takes to exploit that particular capacity and subordinate everything else to that exploitation.  One of the means to exploit the capacity of one resource is to give up sales that generate less T per capacity unit of that particular resource.  When, because of a certain significant marketing and sales move there is a concern that more than one resource would limit the potential T, there is still the option of reducing sales. This could free enough capacity to enable Operations to deliver reliably, and still generate more T than the current state.  Note, this way of dealing with capacity limitation is focused only on the T side, as no change of available capacity is considered for now.  When there is no change in the regular available capacity then there is no change in the level of OE.

There is also another way of dealing with a lack of capacity of a resource. It is a rapid increase of capacity. This is not always possible.  When it is possible to increase the required capacity within the right time-frame, then it generates delta-OE.  It may be necessary to increase the available capacity of a specific resource by only 1%, but the minimum quick capacity increase is 5%.  For instance, the agreement with the union could allow overtime only when several employees stay together.  So, to realize the full delta-T only 1% of capacity increase is enough, but reality requires much higher level of delta-OE.  Eventually the objective result is expressed by delta-T – delta-OE. It is preferable that this be greater than zero.

So far only T and OE have been dealt with. This leads to the following question:

How changes to ‘I’ are considered by this direct approach to predict the impact of a decision on the bottom line?

‘I’ used to be defined as “Inventory”. However, it should rather be used to represent all the money held within the organization.  Both ‘I’ and OE are about maintaining capacity of capabilities.  Inventory of materials required for generating throughput can be viewed as special type of capacity, even though inventory behaves in a more linear manner than most other resources.

The difference between ‘I’ and OE is the expected time-frame for generating value. OE includes expenses that are required for generating T in this year.  Expenses that are supposed to generate T for several years are considered an investment.  These expenses should be apportioned to the relevant years.

This has the effect of treating an investment, like the purchasing of a production-line, as a stream of yearly expenses. Investment justification should compare the basic investment with the stream of annual T.  Again the special value of focusing on T rather than on revenues can be seen.  It is meaningless to use revenues to justify an investment.  When the investment is stand-alone, then a projection of annual profits could be used. However, when the investment is part of the overall activity of an organization, then delta-I can be compared only to a stream of T and possible delta-OE that is fully dedicated to the objective of that investment and is not part of the capacity used for other activities.

The key insight of this chapter is the simplification of dealing with two different aspects, the T side and the OE side. Eventually there is a need to compare the two impacts. Focusing on the issues of each one separately, before combining them, simplifies the process and makes it possible to analyze the decision impact on the bottom line.