How to Best Figure Out Ordering Quantity and Ordering Frequency

Executive Summary

  • Ordering quantities are often the focus of supply chain planning as they directly impact the frequency with which things are done.
  • We cover determining ordering quantities.

Introduction to The Safety Stock and Ordering Quantity and Frequency Scenarios

While not often discussed, the topic of ordering frequency is linked to all of the other subjects that are frequently discussed. These subjects include items such as:

  • EOQ
  • Safety Stock
  • Min Lot Size

That is quantities are much more often a focus of discussions in supply chain planning. It directly affects the frequency with which things are done.

An excellent place to start off this analysis is to look at a straightforward ordering pattern. This occurs with a level demand as is shown in the graphic below.

Safety Stock

Safety stock is one of the best-known concepts in supply chain management. Every MRP application and advanced planning application on the market has either a field for safety stock or can calculate safety stock.

Notice that safety stock is not required because of lead-times or because of the volume of a forecast – it is because of the variability of either of these two components. The second most important thing to understand about safety stock is that variability is projected – it is probabilistic and therefore subject to error. If the variability was predictable, a lower level of safety stock could be maintained – however, variability is generally not predictable.

Economic Order Quantity

The reorder point tells the system when to reorder, while the economic order quantity tells the system how much to order; as such they are necessarily highly integrated values. EOQ is one method for performing what is generally known as lot sizing. The lot size is the quantity in which the item is produced or procured and therefore it is set at the production location combination in the product master. Here it is in Demand Works Smoothie on their Policies Tab.

Min Lot Size

The creation of lot sizes (discrete quantities of a good to be produced or purchased at a given time or per order) is based on the selected lot-sizing procedure. The exact sizes matter because, in each production level, the lots are usually produced completely before being passed on for further processing.

Scenario 1: No Forecast Error or Supply Variability

  • Because the demand is entirely level and there is zero forecast error, a mere economic ordering quantity if (EOQ) of 750 units can be placed once every three months. This is true even though the lead time is only two weeks. This ordering quantity is represented by the orange line. This quantity is received into inventory every three months.
  • Because there are zero forecast error and zero supply variability, there is also zero safety stock.

Now that we have covered a scenario with no variability, let us add variability and add see what happens to safety stock.

Scenario 2: Forecast Error and Supply Variability

Now we will change the scenario to include variability. The forecast error for this product is now 14.6% and the lead time variability has been modified to 7.4%. Both errors apply over the procurement/production (supply) lead time. However, safety stock is only necessary under the following scenarios:

  1. When actual demand exceeds the forecast.
  2. When the supply leads time exceeds the actual lead time.

If the reverse happens, then we are only left with excess stock.

Notice how the forecast error changes when we are only concerned with negative forecast error.

Both errors or variations are averages.

If we put together a sample of forecast error and procurement/production lead time history, it might look like this.

History of Forecast Error and Lead Time Variability

 JanFebMarAprMayAverage ErrorAverage Error for Safety Stock Purposes
Forecast Error History20%-25%+15%+10%-3%14.6%-5.6%
Lead Time Variabiltiy5%-15%0-10%25%7.4%-5%

  1. The error is much smaller when we do not take into account errors that do not pull from stock.
  2. This is what we will increase our starting stock level for, which is -5.6% + -5% = -10%, or 10%.

The safety stock here is not statistically calculated in this scenario. Instead, it is a mean error — or a service level at exactly the 50th percentile. This means we need to add the twin variabilities that sum up to 10%.

This is how much we have to increase in stock. The stock required to cover the possibility that both the supply is late and the forecast is in error. If both of these events do not occur, then we will have carried excess stock.

The next question is what should this increase be calculated over.

There are two options:

  1. Lead Time Demand
  2. Demand Between Replenishments

Let us discuss each in detail:

Lead Time Demand

The standard answer is that the bump in stock should be over lead time demand. The order frequency is once every three months — but the lead time is 1/2 of a month. Enough safety stock is necessary to cover the two-week lead time — if demand spiked, we could place an order before the next regular order date.

  • The lead time is 1/2 of a month, and the average monthly demand is 250. So 250 units is one of the parts to the formula.
  • The 50th percentile sums to 10%.
  • Therefore 10% should be multiplied by 1/2 * 250 or 125. This brings a safety stock of 12.5.

Demand Between Replenishments

The EOQ was already calculated, and it was determined that this product at this location should be ordered once every three months. With a 1/2 month replenishment time, the flexibility does exist to keep a little safety stock, but only by taking on a risk that an order will have to be placed on the next standard order date or reorder date, and that is less economical.

To enforce the EOQ, the 10% increase would be applied to the three-month demand of 750 units, which would bring the safety stock to 75 units. Because the safety stock is 75 units, the plan is for that 75 units to be carried throughout the duration, and therefore this bumps up the ending inventory of each month by 75 units.

Now we are ready to make the safety stock statistical, which means using a service level.

Scenario 3: Forecast Error and Supply Variability + Service Level

Scenario 2 would only offer the appropriate safety stock for up to 50% of cases, or what is called the 50th percentile on a normal distribution curve. However, in most cases, companies have a service level, and this means moving up the distribution curve, which of course gets more expensive. 95% is a very common service goal, so let us apply this service level to our safety stock.

This will mean converting 95% to an inverse value — which happens to be 1.644.

  • So now we take the 75 units and multiply them by 1.644 which yields 123.3.
  • Because our error is a composite value of both demand and supply variability, this level of safety stock should cover us for 95% of the scenarios.

Conclusion

This article was all about showing the trade-offs on order frequency, order quantity (EOQ) and safety stock. The order date is as important as the order quantity. The order date is the timing of the order, and the order date accuracy can be tested and checked historically.

This article used statistical safety stock, but not the standard formula — because we prefer our own as is explained at the following link.

Also, learn about the limitations of EOQ at this link.

Search Our Supply Planning Articles

Remote Supply Planning Consulting

  • Questions About This Area?

    The software space is controlled by vendors, consulting firms and IT analysts who often provide self-serving and incorrect advice at the top rates.

    • We have a better track record of being correct than any of the well-known brands.
    • If this type of accuracy interests you, tell us your question below.

Brightwork MRP & S&OP Explorer for Order Optimization

Order Sizing and Optimization

Order optimization is necessary in order to get the predicted value from ERP and other supply planning applications. 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:

References

Plossel, George. Orlicky’s Material Requirement’s Planning. Second Edition. McGraw Hill. 1984. (first edition 1975)

Safety Stock and Service Level Book

Safety Stock

Safety Stock and Service Levels: A New Approach

Important Features About Safety Stock

Safety stock is one of the most commonly discussed topics in supply chain management. Every MRP application and every advanced planning application on the market has either a field for safety stock or can calculate safety stock. However, companies continue to struggle with the right level to set it. Service levels are strongly related to safety stock. However, companies also struggle with how to set service levels.

How Systems Set Safety Stock

The vast majority of systems allow the setting of safety stock by multiple means (static, dynamic, adjustable with the forecast in days’ supply, etc..). However, most systems do not allow the safety stock to be set in a way that is considerate of the inventory that is available to be applied.By reading this book you will:

  • Understand the concepts and formula used for safety stock and service level setting.
  • Common ways of setting safety stock.
  • Service levels and inventory optimization applications.
  • The best real ways of setting both service levels and safety stock.

Chapters

Chapter 1: Introduction
Chapter 2: Safety Stock and Service Levels from a Conceptual Perspective
Chapter 3: The Common Ways of Setting Safety Stock
Chapter 4: The Common Issues with Safety Stock
Chapter 5: Common Issues with Service Level Setting
Chapter 6: Service Level Agreement
Chapter 7: Safety Stock and Service Levels in Inventory Optimization and Multi-Echelon Software
Chapter 8: A Simpler Approach to Comprehensively Setting Safety Stock and Service Levels

Software Ratings: Supply Planning

Software Ratings

Brightwork Research & Analysis offers the following free supply planning software analysis and ratings. See by clicking the image below:

software_ratings