What is Should Costing?

A pair of compasses sitting on top of some paper.

Defining Should Costing

Should Costing is the practice of determining the "SHOULD COST" of a product, component, or service prior to issuing a Request for Quote. Should Costing is similar and relies on some of the same principles of the typical Cost Estimating, but is entirely different.

Should Costing is an attempt to predict the cost/price of your supplier based on Best-in-class processes. Once you have this established, you then use Fact-Based Negotiations for your purchased components, assemblies, or services.

"Should Costing" is not the be-all and end-all to all procurement issues. It is merely a means to start having intelligent, cost-based discussions. A good negotiator will see where to ask questions and how to understand a supplier's quote if they understand a cost model.

When developing a SHOULD COST, certain ground rules need to be set first. The idea is to develop a defendable, repeatable, possible but unobtainable "SHOULD COST." Let's look at each qualifier in more detail.


Why must the SHOULD COST be defendable? Imagine going into a negotiation trying to defend a fictitious process or material. You have no credibility. You would lose all respect, and anything you say going forward would be meaningless.

When developing a "SHOULD COST," one should use as much verifiable data as possible. There are numerous places to gather data. Various specialized "SHOULD COSTING" software will come preloaded with things like labor rates, machine costs, energy costs, etc. If you are developing your own homegrown system, you can still get this data. For example, you can get labor rates from places like the BLS (Bureau of Labor Statistics) and or ILO (International Labor Organization).

A drawing of a house with tools and equipment on it.
A person is using a calculator and writing on paper.


A strong organization should guard against the possibility of two or more Cost Engineers coming up with drastically different SHOULD COST. This may seem obvious, but it is more difficult than it sounds. Each engineer will come to the job with his or her own knowledge levels, experiences, and expertise.

Being repeatable is important. The level of repeatability could be questioned. If a part, process, or service is "modeled" today or tomorrow, you should be able to get very close to the same SHOULD COST. There will be slight differences. An organization should strive to minimize the differences by setting guidelines, policies, and templates to minimize user discretion. For example, using standard government labor rates is a good practice. However, in these rates are numerous labor categories. Organizations should clearly define when to use one category over another. For example, when do you use Unskilled vs. Skilled Labor? Which one would you use for a CNC operator? Neither is correct nor incorrect, but consistency is the key here.


If the SHOULD COST is based on the impossible, it is no good to either the internal organization or for use in the negotiation phase. Building a "SHOULD COST" will lead the internal organization to believe the "GAP" is much larger than in reality is ever possible. This will cause considerable issues internally. Unrealistic goals for the buyer, unrealistic business cases, etc. Externally, if the SHOULD COST is based on the impossible (unattainable materials, material that does not work in the product/process, process that does not exist, etc.), it will be difficult or near impossible to have a credible fact-based discussion with a supplier.


This may seem to contradict the "Possible," but in reality, it isn't, and it plays an important role in "SHOULD COSTING." A SHOULD COST is designed to stretch the supplier. The SHOULD COST is based on a highly efficient factory, zero or near zero scrap rates (bad part scrap not engineered in scrap), LEAN or batch (whichever drives a better cost), and correct cycle times vs TAKT time.

There are multiple reasons for these pushes. The most obvious is, if you assumed a higher cost than a supplier was experiencing, do you really think they would tell you and, therefore, reduce their selling price just to be nice? A second example, and sometimes controversial, is the 0% scrap. Similar to the above, if your model contained 5% and the supplier was actually experiencing 2%, why would they tell you? An even better explanation for why 0% is correct would be why you are paying and effectively incentivizing them to create scrap. They should always be striving to reduce it.