Figure 1

Supply chains are subject to uncertainties from demand and supply sides

The safety stock level and method depend on the specific industry condition and, more so, on a company’s supply chain philosophy. Two variables that affect safety stock decisions are ownership of the inventory and the metrics adopted to evaluate planners and supply chain organizations. In some companies, the supply chain group is responsible for the inventory, while in others, the responsibility lies with business groups, units, or divisions. If the business is responsible for the inventory and the supply chain organization is evaluated on the basis of its ability to meet the demands from the divisions, then the supply chain organization tends to go for fixed quantity-based safety stock methods. If the supply chain team is responsible for the inventory and is measured on the basis of supply availability to the divisions, then I have noticed better and more sophisticated safety stock practices are deployed to minimize the inventory while meeting service levels.

This article will help you determine which method of safety stock is right for you. In it, I examine five common variations of safety stock:

• Fixed quantity safety stock
  
  
• Days (weeks) of supply
  
  
• Maximum of fixed quantity and days of supply
  
  
• Safety stock as forecast/demand
  
  
• Statistical method

Then, I will detail a sixth method I developed during the course of many implementations, which I call the flexible and cost-effective safety stock method. You can model these variations using SAP Advanced Planning & Optimization’s (SAP APO), demand planning (DP), and supply network planning (SNP) to help you decide which method is best for your company.

When you begin determining safety stock, you need to consider factors that sometimes conflict. Your guiding principle should be to reduce the overall channel inventory without compromising customer service levels. Figure 2 shows a comparative analysis of all the methods. Keep this comparative model in mind as you read through the various methods.

Figure 2

Safety stock methods offer trade-offs between cost, service level, and ease of use

Method 1: Fixed Quantity Safety Stock

Fixed quantity is probably the most common method of maintaining safety stock. You simply assign a numerical value for an item and supply planning engines suggest the production, distribution, or procurement plans accordingly. This method, often referred to as “stock in the pocket,” owes its popularity among planners and supply chain organizations to the sense of comfort it brings with having the physical inventory in the supply chain. In case of an emergency or demand upsurge, supply chain planners can meet the demand without affecting customer service levels.

Figure 3 shows how the fixed quantity method works. The demand is represented with the blue line while the supply plan is represented with a green line. In a perfect world, supply plans should mirror the demand pattern and therefore overlap. However, such patterns are not common. Figure 3 shows the overlap of supply and demand plans from week 2 onward. The supply in week 1 is more than the demand. The difference is due to the fixed quantity safety stock of 200 units. The bi-directional arrow represents the safety stock quantity .

Figure 3

Fixed quantity safety stock method results in extra physical units

Setup in SAP APO

The fixed quantity method requires a clear understanding of the items and the locations where your organization intends to manage safety stock. Then, maintenance of safety stock numbers at that item location level needs to be carried out either by supply chain organizations or the master data team. In stand-alone SAP APO implementations, you can maintain the safety stock values at the product location level. In implementations with SAP R/3, you can maintain the values in R/3 and bring those to an SAP APO product location using Core Interface (CIF).

At the product location level, safety stock values can be maintained in the Lot Size tab and then in the Quantity and Date Determination sub-tab. Access these tabs by using either transaction code /SAPAPO/MAT1 or menu path APO > SAP Menu > Master Data > Product > Product. Selecting the change mode after inputting product and location takes you to the screens where you can maintain product location level details. Select the Lot Size tab and then the Quantity and Date Determination sub-tab to access the safety stock related fields. The Safety Stock Method field determines which method is activated for the particular product and location combination, and the Safety Stock field determines the level of safety stock to be maintained. Select SB from the drop-down menu in the Safety Stock Method field and enter value 200 in the Safety Stock field. This activates the fixed quantity safety stock of 200 units for product L5-01080-07A at location 7610 (Figure 4).

Figure 4

Product L5-01080-07A at location 7610 has a fixed quantity safety stock value of 200

As noted earlier, the fixed quantity method owes its popularity to the ease of covering any demand surges. The supply chain organization is measured on the customer delivery performance but important internal measures, such as inventory and safety stock, are often ignored. The fixed quantity method is prevalent in high tech and consumer packaged goods (CPG) industries where demand patterns are not stable. In the high tech industry, high levels of inventory (including safety stock) are maintained for some key customers and key products. Any compromise on service levels due to lack of inventory could have serious ramifications on the business.

On the other hand, this method also exposes companies to higher obsolescence, inventory, and carrying costs. It is common to see companies in this industry have 25% to 35% of their current assets locked up in inventory. If the supply chain planners are not judicious about maintaining this level, huge inventory obsolescence and write-offs can result. Another drawback of this approach is the lack of scientific methods applied to derive the safety stock values. Primarily, it is based on planners’ experience. For example, I know of a consumer electronics company that used this method, but wound up with two-thirds of its warehouse stocked with inventory. The inventory had not moved in years and had no possibility of gaining momentum in the future since the products had become obsolete. The company was stuck with the inventory and carrying costs. This is not to suggest that the entire inventory resulted from poor safety stock practices, but it is the widely accepted culprit.

Method 2: Days of Supply

Unlike fixed quantity, the days of supply method offers the benefit of not having too much obsolete inventory. Planners can define time buffers based on their experience and supply plans are created using these buffers. As the name suggests, the approach builds a buffer of time between demand and supply. Supply is pulled in before the expected demand by the defined time buffer. In Figure 5, the green line of supply is offset by three weeks from the blue demand line. In this case, the supply for week 4 demand is created in week 1 and the supply for week 4 is created for the week 7 demand.

Figure 5

Supply plan mirrors demand patterns but with an offset

If the demand of week 7 reduces from 500 to 400, the supply plan adjusts itself, as seen in Figure 6.

Figure 6

Supply plan adjusts to the changes in demand patterns

Setup in SAP APO

Similar to fixed quantity, days or weeks of supply also requires that you have a clear understanding of the items and locations where your organization intends to manage the safety stocks. There needs to be clarity around lead times and deviations in those lead times. That is a vast topic in itself, but once it has been agreed on, the maintenance of safety stock numbers at that item location level needs to be carried out. In stand-alone SAP APO implementations, safety stock days’ values can be maintained at the product location level. In implementations with SAP R/3, values can be maintained in R/3 material master and brought over to an SAP APO product location using CIF.

Similar to the fixed quantity method, product locations can be accessed by using either transaction code /SAPAPO/MAT1 or by following menu path APO > SAP Menu > Master Data > Product > Product. At the product location level, you can maintain the safety stock days’ values in the Lot Size tab and Quantity and Date Determination sub-tab (Figure 7). The Safety Stock Method field determines which method is activated for the particular product and location combination. The Safety Days’ Supply field determines the level of safety stock to be maintained. Select SZ from the drop-down menu in the Safety Stock Method field and enter value 14 in the Safety Days’ Supply field.

Maintain the Safety Days’ Supply field with the number of days in terms of calendar or work days depending on the selection of the supply planning engine. The SAP APO planning engine Capable to Match (CTM) uses this field in calendar days while Optimizer engine treats this as work days.

Figure 7

Product L5-01080-07A at location 7610 has 14 days of safety stock maintained

The days or weeks method helps to reduce costs by adjusting the supply with changing demands. It is particularly useful and suitable for companies facing uncertainties from the supply side but have relatively stable demand patterns. In this age of outsourcing, companies are more and more leaning on partners to manufacture goods. These partners are subject to different geographical, political, and economical situations that affect supply chains. Days of stock helps reduce the cost of carrying the extra inventory. However, building the supply earlier than the demand offsets some of the benefit.

The biggest disadvantage to this approach is that it does not allow the supply chain to meet any demand upsurges within the manufacturing lead time. Say, for example, it takes two weeks to manufacture the finished goods in Figure 7. With the days of supply method, a supply of 100 for week 3 demand would be available in week 1. If the customer demand comes for a quantity of 200, then it isn’t possible for the planner to meet the increase in demand without affecting customer service. This not only adds to performance deterioration but also makes the supply chain organization nervous. This method does not provide much insurance from any disruptions in the supply side. If the vendor has a plant breakdown for a few weeks, then the supply chain dries up. In my experience, I have found that planners like this method theoretically, but still want to have some physical stock in place. Many companies often inflate supply lead times instead of using days of supply to imitate this method to build buffer in the supplies.

Method 3: Maximum of Fixed Quantity and Days of Supply

This method combines the first two methods – fixed quantity and days of supply – and considers the greater of the two values as safety stock value. Let me illustrate this method with an example shown in Figure 8.

The demand pattern is shown in line 1 in Figure 8. At the product location level, fixed quantity safety stock of 200 units (line 2) and days of supply of 14 days (line 3) are maintained. Fourteen days of safety stock would mean safety stock value of 120+140=260 units (line 4). The greater of two safety stock values 200 and 260 would lead to final safety stock value of 260 (line 5). With the safety stock of 260, the supply plan would be as shown in line 6.

Setup in SAP APO

Product location can be accessed by using either transaction code /SAPAPO/MAT1 or by following menu path APO > SAP Menu > Master Data > Product > Product. Follow the instructions as I have outlined in the first two methods until you reach the Change Product for Location Screen (Figure 9). Then, select SM from the drop-down menu in the Safety Stock Method field, maintain 200 in the Safety Stock field, and enter value 14 in the Safety Days’ Supply field.

Figure 9

Maximum of Safety Stock and Days of Supply method activated for product L5-01080-07A at location 7610

This method is a hybrid between the fixed quantity and days of supply methods and offers insurance against the fluctuating demand by adjusting the supply plan. It assures a certain stock level for any demand upsurges. However, like the previous two methods, you still can run into the issue of estimating the appropriate safety stock numbers. I have seen this method deployed by a high tech company for the product lines during the ramp-up production phase. During this phase, the company either launched a new product or qualified a new production facility or combination of the two. This led to increased uncertainty from the supply side as well as from the demand side. The company used a fixed quantity safety stock number from a similar product and assembly lead time as the days of supply. This allowed the company to meet the uncertainties over the production ramp-up phase. After the ramp-up phase, based on the product and its demand, the company reverted to a different safety stock method.

Method 4: Safety Stock as Forecast/Demand

Another way to deal with safety stock is to artificially increase the demand or forecast to accommodate the safety stock. This approach takes away the pain of calculating the safety stock values and resulting accountabilities due to the fact that any residual obsolete inventory is passed on as the result of fluctuating demand. In Figure 10, you can see that the planner has upped the demand in the first week from 100 to 300 to accommodate the safety stock.

Figure 10

Forecast has been increased in week 1 to accommodate safety stock of 200 units

The safety stock as forecast/demand method is common across industries and companies. Many companies do not have a properly streamlined approach to monitor forecast accuracies, but have stringent customer-facing and inventory matrices such as on-time delivery performance and inventory turns. This situation leads planning departments to increase the forecast itself to accommodate any safety stocks. This does not lead to any accountability issues.

Setup in SAP APO

Use the DP book to manage this method. Access the planning book with transaction code /SAPAPO/SDP94 or by following menu path APO > SAP Menu > Demand Planning > Planning > Interactive Demand Planning. Once in the planning book, rows or key figures are used to capture the type of data, while columns are used to capture the data itself. As shown in Figure 11, the key figure Market Final Forecast captures the forecast for a specific product at a location. The forecast number in W24 has been increased from 100 to 300 to accommodate the safety stock.

Figure 11

Safety stock number is built in the Market Final Forecast row for week 28

The safety stock as forecast/demand method is treated like any other forecast, hence absolving planners from accountabilities up to a certain extent. This also takes away the pain of deriving the safety stock values. However, the biggest disadvantage of this method is that it leads to erroneous calculations in forecast accuracies which in turn could skew future forecasting. Many companies want to differentiate the supplies proposed on the basis of demand type to help them properly execute, in case of any supplier issue such as capacity constraints. In Figure 11, for example, there is no indication that 300 units were created for forecast or forecast and safety stock. Had this been known, in case of capacity constraints, it would have been easier to assign the priorities (e.g., build first to meet the forecast and then for safety stock).

In one company I worked with, the supply chain organization discovered that no supplies were being created for most of the products. Upon investigation, my team and I discovered that instead of entering 5,000 as the safety stock in the planning book, the planner has added a few more zeroes and it became 50 million units of forecast for one specific product. As a result, all the capacities in the supply chain were diverted to this product, depriving other products of capacities.

Method 5: Statistical Method

The statistical method of determining the safety stock quantities is based on the demand/ supply variability and customer service level targeted. The demand variability is measured in terms of demand (forecast and sales order) variation and the supply variability is measured in terms of lead time variations. This method can be applied to products with regular or sporadic demand.

The formula used to calculate the safety stock is:

Safety Stock = K* Square Root of [(Average Demand * square of lead time variation) + (Average Lead Time * square of demand variation)] (Figure 12).

Figure 12

Normal distribution is considered while calculating service factor K

Various variables in the above equation are as follows:

K is a service factor that depends on the service level targeted by the company. Service levels can be defined as the percentage of requirements fulfilled with the warehouse stock. Based on the context, the definition of the service level could be different. Sometimes it is taken as the number of periods with no shortfall while other times it is defined as expected demand fulfillment in terms of units. For example, if out of 10 periods the expectation is to meet nine periods of demands completely, then the service level would be 9/10 = 90%. If the demand for one period is 100 and the expectation is to at least meet 90% of the service level, then 90 units of stock have to be maintained to meet the demand. Based on the industry, company, and measure adopted, the definition of service level could be different. In the CPG industry, it is common to use the second definition where a 90% service level means carrying 90 units of stocks for every 100 units of demand. In the high tech industry, the first definition is more prevalent where the focus is on meeting demand 90% of the time.

For regular demand patterns, the formula determines K based on a normal distribution curve shown in Figure 12. For reference, the service factor table has been included in Table 1. From the previous example, a service level of 90% would translate into a service factor of 1.28.

Although the method has been much debated in the academic world, in my experience, I have seen few companies adopting the statistical approach to safety stock. There are plenty of reasons for such low adoption. First, the method is difficult to comprehend. It involves variables such as service factors that could have different meanings based on the context. Secondly, the results can be counterintuitive and may not provide adequate coverage of the demands from some important customers and for key products. It does not take things such as relationships between business entities into the equation. Finally, this method does involve lot of work. Let me explain.

I once saw this method deployed in a CPG company where the demand was volatile and the chances of stock-outs while carrying inventory for the wrong product mix were very high. Usually in the CPG industry, customer loyalty is a rare commodity. If a favorite brand is not available, the customer switches to the next available brand. Therefore, CPG companies require very tight control over their inventory policy. In order to compete, they must excel at capturing Point of Sales data, which leads to better forecast and safety stock calculations. The method was deployed successfully, but the company ran into some problems. It could not capture the point of sales data and, rather, had to bank on the forecast patterns provided by customers.

Setup in SAP APO

Similar to the fixed quantity and days of supply approaches, the variables related to this method are also maintained at the product location combination level. You access product location by using transaction code /SAPAPO/MAT1 or by following menu path APO > SAP Menu > Master Data > Product > Product. To activate this safety stock method, maintain the following fields: Safety Stock Method, Service Level (%), Demand Fcst. Err(%), RLT Fcst Error (%), and Replen. Lead Time. In Figure 13, you can see that Product 106494552 at location 7000 has a service level of 90%, a demand forecast error of 40%, an average replenishment lead time of 44 days, and a replenishment lead time (RLT) forecast error of 30%. The demand forecast error is the same as demand variation and RLT Fcst Error is the same as Lead Time variation.

Figure 13

Statistical safety stock method is activated for product 106494552 at location 7000

Transaction /SAPAPO/MSDP_SB or menu path APO > SAP Menu > Supply Network Planning > Planning > Safety Stock Planning carries out the safety stock calculation. The Safety Stock Planning screen (Figure 14) carries out the safety stock calculation for Product 106494552 at Location 7000 considering demand from the Key Figure for Demand Forecast field and other maintained values.

Figure 14

Transaction /SAPAPO/MSDP_SB carries out the safety stock calculation

The statistical method is a scientific method and takes a broader comprehensive view of the supply chain since it considers the demand, lead time variability, and service levels. SAP APO offers this functionality to carry out such calculations for a large number of products in a quick and efficient manner. Having planners carry out such calculations is very time consuming and inefficient. Also, the results are not in line with the supply chain planners’ expectations, making them difficult to comprehend and follow. For example, based on all the variables involved, safety stock planning could suggest carrying 60 days of supply as the safety stock, while the average lead time is only 44 days. This makes it difficult for a planner to grasp other factors, such as service level variables, that the system didn’t consider when suggesting 60 days of supply. Also, there are always some variables that are not known to the system but only to the human intelligence, such as trends in the market, ground level realities, or relationships with customers. All these factors make this method not very popular. However, it can and should be used to do a sanity check on safety stock numbers.

Method 6: Flexible and Cost-Effective Safety Stock Method

This brings me to this final safety stock approach. This simple but effective method achieves the objective of calculating safety stock values within standard SAP APO without resorting to any modifications or enhancements. In contrast to the above methods, this method offers the following benefits:

• Captures the supply chain planners’ intelligence and experience
  
  
• Adjusts the supply plan if demand changes
  
  
• Helps build physical stock in the warehouse to accommodate any demand upsurges
  
  
• Intuitive, user-friendly, and easy to use
  
  
• Helps to assign the responsibility of inventory and measure performance

This method is based on the premise of how many weeks of average demand you want to carry as safety stock in your supply chain and over how many weeks you want to spread this safety stock. Planners provide three key components of this method: how many weeks of demand to consider for average demand calculation; how many weeks of average demand to carry as safety stock; and how many weeks to spread the safety stock. The first component captures the period of demand variability based on planners’ experience and knowledge. The second captures the lead time aspect of the product and the third captures the period of slow buildup without hogging precious supply chain resources such as capacities.

Let me illustrate the method by an example. The demand pattern for a product is as shown in Table 2.

Supply chain planners decide that, for this product, four weeks of demand should be considered for average demand calculation (Table 3).

The average demand calculation is carried out. The average demand comes to 387.5 (Table 4).

Now, supply chain planners can determine how many weeks of average demand should be carried as the safety stock (Table 5). In this example, it is two weeks.

Safety stock calculation is carried out. Total safety stock = Average Weekly Forecast * # of Wks S.Stock required = 387.5 * 2 = 775 (Table 6).

Companies can follow different approaches after this step. Some companies place all the safety stock demand in the very first period or any one period while others tend to spread the safety stock over a period of time. Spreading the safety stock requirements over a period of time helps companies to build the safety stock slowly without tying up costly supply chain resources such as production and warehouse capacities. It helps to reduce the working capital by not locking money in the safety stock and also reduces the inventory carrying cost. In my example, planners can decide to spread the safety stock over a period of four weeks (Table 7).

Based on the allocation of the weeks defined, the safety stock can be spread over a period of time. The final safety stock values are displayed in the last row entitled Safety stock (Table 8).

Setup in SAP APO

You can use the planning book functionality along with the macros available in SAP APO DP and SNP modules to set up this safety stock method. Access the planning book with transaction code /SAPAPO/SDP94 or by following menu path APO > SAP Menu > Demand Planning > Planning > Interactive Demand Planning. The planning books can be configured with different rows or key figures as shown in Table 9. The names of the key figure can be changed by the planners or users to suit the business requirements. For example, in Figure15 the key figures have been renamed by planners to suit business requirements (e.g., Average weekly forecast or Safety stock allocation weeks).

Figure 15

DP/SNP planning book is designed to capture all the steps of safety stock calculations

Macros can be defined in the system by users to carry out the mathematical calculations such as averaging or multiplication as outlined in Tables 2 through 8. You can build macros by using a macro builder. Access your macro builder by using transaction code /SAPAPO/ADVM or by following menu path APO > SAP Menu > Demand Planning > Environment > Current Settings > Macro Workbench. Figure16 shows an example of a macro builder.

Figure 16

Macro is defined to do average demand calculations in the planning book

The safety stock calculations outlined in Tables 2 through 8 can be carried out systemically after the planning book and macros are defined. For example, Figure 17 shows that the average weekly forecast of 388 is calculated and populated by a macro based on Forecast and Safety Stock Average Calculation Weeks values. Another macro uses 388 and number of weeks stock required to calculate Total Safety Stock of 775. Yet another macro can utilize values of Total Safety Stock and Safety Stock Allocation Weeks to spread the safety stock of 775 over 4 weeks –194 each.

Figure 17

Steps of safety stock calculation are performed in the SAP APO DP/SNP planning book

This approach can be applied across industries. Not only does it allow planners to control the safety stock levels, but it also provides the provision to capture human intelligence. The variables such as average calculation weeks and allocation weeks can be reviewed and managed on a regular basis. Thus, the planner has greater flexibility to decide the safety stock levels. Also, it considers the demand fluctuation on a dynamic basis.

However, there are some cons to it as well. For one, it is not as scientific as the statistical method in the sense that it does not calculate the demand and lead time variability to arrive at the safety stock number. Instead, the method banks on the judgment of planners. In my experience, I have seen this approach working very well for three clients. This method helped a consumer electronics company manage some key products and customers. Furthermore, it helped the company realize savings in inventory while exceeding its customer service targets. The company was able to cut down the safety stock carried in the supply chain by 30% while it maintained targeted customer service levels of 90%.