Should Non Finished Goods Safety Stock be Carried?

Executive Summary

  • Safety stock is how stock is kept to account for variability.
  • George Plossl discussed how safety stock ruins the credibility of MRP, and whether safety stock should be applied to raw materials and components. Safety stock changes for make to stock versus assemble to order.

Introduction to Safety Stock

Safety stock is the quantity of stock carried to manage supply and demand variability. Cycle stock is the stock carried between the order cycle, and safety stock + cycle stock = the total stock.

Setting Safety Stock for Different Material Types

Any supply planning system has a safety stock setting at the product location combination. This means that all materials, be they finished goods, raw materials, components, etc.. have the ability to have safety stock set for them at every location where they are created in the system.

A question which many companies have, and which is infrequently answered is how safety stock be set at each location? Should one apply the same rule of safety stock policy for each material type?

George Plossl was one of the deepest thinkers on MRP and inventory management. He had some interesting observations on this topic, as can be seen from the quotation below:

Plossl On: Does Safety Stock Ruin the Credibility of MRP?

“Safety stock has little, if any legitimate role in MRP, however, which provides forward visibility of planned requirements and orders for raw materials and components and can replan easily. When safety stock is added to MRP, the resulting overstated requirements and false timing of order release and due dates destroy credibility. Factory personnel quickly discover order due dates, believing that safety stock is available and that orders are scheduled for completion before they are needed. Missed delivery dates without customers’ quick and strong reactions give suppliers clear evidence that their purchase order due dates are padded.”

That is a curious statement and one which is not well known. This is quite strange as it is unheard of to use MRP without safety stock.

Plossl On: What is the Primary Purpose of Safety Stock?

“The primary purpose of safety stock is to provide cushions against fluctuations in demand (forecast error) and supply (upsets in production). Demand for raw materials and components are calculated by MRP and therefore, are not subject to forecast errors. Experience with safety stock in MRP has shown clearly that the cushions are rarely found on items affected by users, nor are they large enough to be useful. Better understanding of the importance of smooth, fast flow of materials and flexibility of operations led to the realization that the causes of upsets had to be eliminated. Efforts to do this proved successful and yielded enormous benefits far beyond the costs in incurred.”

This paragraph requires rereading several times (at least it did for us). But it seems to be saying that the safety stock of the finished goods is not large enough to be useful or worthwhile. And the quote seems to take the view of Lean proponents that it covers up the reasons for the variability.

Plossl On: Should Safety Stock be Applied to Raw and Component Levels?

“Safety stock is properly only to inventory items subject to independent demand. These include stocked finished products, service parts and MPS end items….(safety stocks) should not be duplicated at raw materials and component levels.

Another function of safety stock is protection against uncertainly of supply. It can be useful sometimes on an item whose supply is erratic and beyond the control of people in the plant, characteristic of some purchased items. Safety stock will be wasted when production of manufactured items is erratic and unpredictable; planning cannot be sound in chaotic environments.” – Orlicky’s Material Requirement’s Planning

The synopsis of this quotation is that safety stock should only be carried at the finished good. And for firm make to stock environments this makes sense — especially where there is not overlap or shared raw materials and finished goods. The table below is an example of this type of scenario along with the scenario of assemble to order.

Make to Stock and Assemble to Order Scenarios

Material TypeMaterialMTS Scenario 1 - Used by Which Finished Good BOM?ATO Scenario 2 - Used by Which Finished Good BOM? 
Finished GoodTable Type 1N/AN/A
Finished GoodTable Type 2N/AN/A
Raw MaterialWood Type 1Only for Table Type 1All
Raw MaterialWood Type 2Only for Table Type 2All
Raw MaterialWood GlueAllAll
ComponentLegsNot Interchangeable Between Table 1 and 2.All
ComponentJoinsNot Interchangeable Between Table 1 and 2.All
ComponentTable TopNot Interchangeable Between Table 1 and 2.All

Make to Stock Scenario – No Interchangeability

Notice that in the first scenario – it makes no sense to carry safety stock at the raw material. This is because all of the variability is accounted for in the finished good.

Assemble to Order Scenario – Interchangeability

In the second scenario, things change because the components and raw materials are now interchangeable. This means that the company can follow an assemble to order strategy. This means they do not have to forecast the demand for both Table Type 1 and Table Type 2 but need to create a combined forecast.

It means they can also choose to create safety stock at the raw materials and components — and reduce it at the finished good. How much they can do this depends on the final assembly lead time.

There are two options:

  • If the Final Assembly Time is — Combined with the Delivery Time <(Smaller Than) the Customer Lead Time
  • If the Final Assembly Time is — Combined with the Delivery Time >(Greater Than) the Customer Lead Time

Let us get into each option in detail.

If the Final Assembly Time is — Combined with the Delivery Time < the Customer Lead Time

The company can switch all of its safety stock from finished goods to raw materials and components – removing the safety stock that is carried at the finished good.

If the Final Assembly Time is — Combined with the Delivery Time > the Customer Lead Time

Some safety stock can be funding can be re-allocated from the finished goods to the raw materials and the components. The longer the final assembly + delivery time is versus the customer lead time, the less safety stock can be re-allocated from finished goods to raw materials and components.

The movement and interchangeability of components change the inventory that must be held. This can be seen by looking at the how forecast error changed when the demand for finished goods can be combined when there is component interchangeability.

Material TypeMaterial January ForecastJanuary DemandIndividual Forecast ErrorCombined Forecast Error
Finished GoodTable Type 120017512.5%2.8%
Finished GoodTable Type 215016510%
Average Error11.25%2.8%

Different individual forecast scenarios lead to different error improvements. For instance, if the sales move in the same rather than in opposite directions, as I have shown above, the forecast error reduction is less. 

Material TypeMaterial January ForecastJanuary DemandIndividual Forecast ErrorCombined Forecast Error
Finished GoodTable Type 120017512.5%11.46%
Finished GoodTable Type 215013510%
Average Error11.25%11.46%

In this case the error seems higher with the aggregated value — however, I just took an average rather than weighing the forecast error by the forecast size. If I weighted the forecast error, the forecast error for the divided forecast would be higher. 

The Reality of Make to Order Versus Assemble to Order

Assemble to order manufacturing is preferable to make to stock.

Make to order is the most preferable, but make to order is not feasible. That is it is not feasible except for a small portion of the overall product database of most companies. This is true outside of specialized industries like aerospace and defense.

No company makes a fighter jet without a firm order in hand. Make to order has a very high bar. The customer lead time must be greater than the overall manufacturing lead time + the procurement lead time + the delivery time. The bar for assemble order is much lower. It is the final assembly time + the delivery time.

Many companies speak of moving to make to order but assemble to order a much more realistic goal. And as can be seen, it has beneficial effects on inventory management. To implement assemble to order two important prerequisites must exist:

  1. Assemble to Order Lead Time: This should not be confused with the overall manufacturing lead time — but this value + the delivery time must be shorter than the customer lead time.
  2. Interchange-ability: The finished goods must have a degree of interchangeability between raw materials or components. However, it is important to note that not all raw materials and components in the bill of material need to be interchangeable. If less expensive raw materials or components are not interchangeable, infrequent orders, high stock, and safety stock can be carried on them at little cost. If on the other hand the more expensive raw materials or components are not interchangeable — then it makes it less likely that the assemble to order strategy can be followed.


Where the safety stock is set in the bill of materials goes hand in glove with the type of environment that is being modeled.

  • For a pure make to stock environment, George Plossl makes a good argument that safety stock should be allocated to the finished good.
  • Something that is little discussed is that most companies cannot carry all of the safety stock that the standard calculations calculate they should.

This means taking a limited amount of safety stock and allocating it across a product location database in the best possible way.  So it also means that safety stock should not be calculated for each product location independently.

Brightwork MRP & S&OP Explorer for Safety Stock

Safety Stock Calculation

The way safety stock is set and sometimes calculated is highly imprecise and problematic. What is needed is a way of calculating safety stock dynamically that is performed in an integrated manner with both constraints and service levels. The Brightwork MRP & S&OP Explorer does exactly this, and it is free to use in the beginning until it sees “serious usage.” It is permanently free to academics and students. See by clicking the image below:



Lean and Reorder Point Planning Book

Lean and Reorder Point 2

Lean and Reorder Point Planning: Implementing the Approach the Right Way in Software

A Lost Art of Reorder Point Setting?

Setting reorder points is a bit of a lost art as company after company over-rely upon advanced supply planning methods to create the supply plan. Proponents of Lean are often in companies trying to get a movement to Lean. However, how does one implement Lean in software?

Implementing Lean in Software

All supply planning applications have “Lean” controls built within them. And there are in fact some situations where reorder points will provide a superior output. With supply planning, even within a single company, it is not one size fits all. The trick is understanding when to deploy each of the approaches available in software that companies already own.

Are Reorder Points Too Simple?

Reorder points are often considered to be simplistic, but under the exact circumstances, they work quite well.

There are simply a great number of misunderstandings regarding reorder points – misunderstandings that this book helps clear up.

Rather than “picking a side,” this book shows the advantages and disadvantages of each.

  • Understand the Lean Versus the MRP debate.
  • How Lean relates to reordering points.
  • Understand when to use reorder points.
  • When to use reorder points versus MRP.
  • The relationship between forecastability and reorder points.
  • How to mix Lean/re-order points and MRP to more efficiently perform supply planning.


  • Chapter 1: Introduction
  • Chapter 2: The Lean versus MRP Debate.
  • Chapter 3: Where Supply Planning Fits Within The Supply Plan
  • Chapter 4: Reorder Point Planning
  • Chapter 5: Lean Planning.
  • Chapter 6: Where Lean and Reorder Points are Applicable
  • Chapter 7: Determining When to use Lean Versus MRP
  • Chapter 8: Mixing Lean and Reorder Points with MRP-Type Planning