There has always been a strong desire to be able to calculate the overall performance of a system (or business) by summarizing the contributions of each individual component to the system (or products and services).

This aspiration led to the invention of product costing near the beginning of the 20th century. At that time, expenditures in most manufacturing companies were dominated by raw material and direct labor, while the amount of management and indirect labor (e.g., white-collar employees) involved was negligible. Generating a value for product cost per piece of product was reasonably accurate, since the largest portion of the Profit and Loss (P&L) statement (material and direct labor) could be allocated to the individual products.

This situation has now changed dramatically. While material and direct labor are still important aspects of a company’s spending, the amount of "overhead" has increased significantly. Consequently, the traditional approach to product costing is no longer accurate nor viable.



Piece part cost calculation in our ERP age is typically based upon:

With this information, the individual product cost per piece may be calculated as follows:

(1) Material Consumption M = å (BOMk * Pk)

where BOM is the amount of material "k" used

P is the purchasing price per unit of material "k"

k is the product index.

(2) Fabrication cost per piece to produce the product is:

Fabrication cost F = å (Cl * Rl * 1/n)

where Cl is the net cycle time per production operation "l"

Rl is the hourly rate for operation "l"

l is the operation index.

n number of products produced per cycle time (e.g. cavities)

After enumeration of material consumption and fabrication, the total cost incurred per reporting period may be computed as:

(3) Total Cost TC = å (Mi + Fi)

where i is the index of products produced in that period.

This approach is logically valid as long as the assumptions underlying formulas (1) and (2) are well founded. One fundamental assumption

is that one can "itemize" spending. This assumption translates into

the ability to assign spending in "currency units per product quantity" [e.g. dollars per pound, or $/#]. Checking this further,

(1) M = å (BOMk * Pk)

BOMk is definitely measured in "physical units per piece of product", typically [kg/#], [m/#], [l/#], [#/#] or similar units.

Pk is also known as "dollars per unit of measure" such as [$/kg],[$/m], [$/l], [$/#].

So this translates very easily in a material cost [$/#] per unit of product produced and the above assumption is valid for this portion of the equation.

(2) F = å (TCl * Rl * 1/n)

TCl is derived by measuring the net cycle time for each individual operation necessary to produce (one piece of) the product or [sec/#]. While in principle this is a discrete value, in reality it is very difficult to establish an exact value. Cycle times are subject to statistical variation with standard deviations often in the range of approximately plus or minus 10%.

Rl is a value derived from a very doubtful calculation based on general ledger (GL) information [$] and projected net production run times per reporting period [e.g., hours]. The reporting period could be an accounting period (month) or a fiscal year.

n represents the number of products produced during one cycle. In injection molding "n" represents the number of cavities; for assembly operations n = 1 (typically).

Rl = å (SPENDINGm) / HRSm

where SPENDINGm is the ALLOCATED currency units spent per

production resource "m"

HRSm is the projected net production run time per

production resource "m"

å (SPENDINGm) is the total of expenditure and basically is the summation of expenses such as:

[$/m2] rental fee

[$/kWh] electricity

[$/month] wages for overhead personnel

[$/m3] water

[$/bill] for miscellaneous types of spending


that need to be allocated to each production resource.

Translating and allocating these diverse types of spending into an amount of "currency spending per production resource and reporting period" has kept numerous accountants and clerks busy for many, many decades. It is obvious that the accuracy of this allocation process includes an error of at least plus or minus 20%. Complicated approaches to find and apply "cost drivers" in order to allocate spending more accurately, such as Activity Based Costing (ABC costing), do not increase accuracy substantially. In fact, they still attempt to allocate, through a more complex approach.

So in calculating fabrication cost (2), cycle times and hourly rates are used that are each individually subject to considerable errors of +/- 10% to +/-20%. The end result of the computation of (2) is therefore prone to error of significant magnitude and consequently the calculation of the final product cost according to formula (3) is of questionable validity.

The basic assumption of assigning material and fabrication cost to spending per unit of product produced is therefore invalid, due to the inherent inaccuracies included in cycle time estimates and to approximations of hourly costs. Consequently any further measures, such as margin per product or profit per product, derived from this inaccurate product cost calculations are equally flawed. The assumption that the performance of a system or company can be calculated by totaling the performances of individual products is wrong as well.



The cost calculation model was developed to support the management decision-making processes.

What is necessary in order to make "good" decisions ?

First of all, the goal of the system for which decisions are to be made needs to be well defined. The goal in "for profit" organizations is usually to make money and to be profitable.

Next, measurements to evaluate performance relative to the goal need to be chosen based on accurate information that is available.

The goal to make money may be expressed as:

(5) NP = å (Si - Mi) - å SPENDINGj

where NP is the net profit of the total system

Si is the revenue generated by selling product "i"

Mi is the material consumed per sale of product "i"

SPENDINGj is the spending per type of spending "j".

This may be shortened to:

(6) NP = å Ti - å SPENDINGj

where Ti is the throughput per product "i"

Both Ti as well as SPENDINGj are readily available. Together they constitute the Profit and Loss (P&L) of the total system or company.

Profitability can be calculated as:

(7) ROI = NP / å In

where ROI is the return on money invested into the system.

In is the money invested per type "n" of assets.

NP is available and In is taken from basic balance sheet information.

Good decisions will increase NP and Return on Investment (ROI) of the system. All that is required is:


Availability of cash (nCF = net cash flow) per reporting period may be

calculated as:

(8) nCF = NP + delta (I)

Every system is subject to constraints such as labor, machines or external restrictions such as the market or suppliers. Good decision-making will attempt to maximize NP and ROI improvements by exploiting these constraints while maintaining positive nCF.

These thoughts and concepts are the core of the Throughput Accounting aspects of the Theory of Constraints. A suggested Internet resource as a starting point for further investigation is:


Hans Peter Staber

August 2001