Use of Supply Chain Modeling to Examine Total Supply Chain Costs

Author(s):

Lee Buddress, Ph.D., C.P.M.
Lee Buddress, Ph.D., C.P.M., Assistant Professor, Portland State University, P.O. Box 751, Portland, OR 97207-0751, 503-725-4769, leeb@sba.pdx.edu
Alan R. Raedels, Ph.D., C.P.M.
Alan R. Raedels, Ph.D., C.P.M., Professor, Portland State University, P.O. Box 751, Portland, OR 97207-0751, 503-725-3728, alanra@sba.pdx.edu
Jim Burrus
Jim Burrus, supply chain manager from Hewlett Packard, 18110 SE 34th St, Camas, WA 98607
Scott Culbertson
Scott Culbertson, supply chain manager from Hewlett Packard, 18110 SE 34th St, Camas, WA 98607
Scott Swenson
Scott Swenson, supply chain manager from Hewlett Packard, 18110 SE 34th St, Camas, WA 98607

85th Annual International Conference Proceedings - 2000 

Use of Supply Chain Modeling to Examine Total Supply Chain Costs

Introduction/Problem Definition. Many manufacturing industries are facing technological advances that have drastically reduced product life cycles. If an organization is not ready with the latest advances, the competition will be. Attempts to gain customer loyalty in this fast paced market has lead to an ever increasing variety of products and the phenomena of mass customization (the ability to provide individually designed products on a mass scale). This environment continues to drive customer expectations for variety, availability, and quality, not to mention ever-lower prices. Managers understand that the only way to stay alive in this game is to be nimble. Too much inventory late in the product life cycle may need to be deeply discounted and, more often than not, must be written off as obsolete. On the other hand, being too lean on inventory quickly leads the customer base to select from a wide variety of alternatives that the competition is both able and willing to provide. In the face of these pressures, managers have been trained to increase inventory "velocity". Remaining flexible and responsive while efficiently moving inventory through the supply chain has become the mantra of the industrial world.

There are many ways to shorten cycle times in the supply chain. Postponement provides for the final changes in the product to occur late in the supply chain allowing inventory to remain flexible to customer demand without holding large stocks of multiple SKU's. More conventional means include moving transportation from ocean to air, giving suppliers a financial incentive to locate close to your assembly points, or investing in information tools that reduce the decision cycle times in the supply chain. All of these efforts seek to reduce total supply chain cycle time. However, the continued focus on lean production and distribution means that purchasers must make educated investments. More often than not, the supply chain manager does not have sufficient tools to calculate the expected return on cycle time reduction investments. For example, a purchaser may have an option to source the longest lead-time component locally for $.05 more per component. If it takes two weeks out of the supply chain cycle time, is this a good investment? What if the cost increase was $.10 per component?

The following article demonstrates an approach to evaluating the value of time in the supply chain. The application of these concepts is most appropriate for industries with relatively long total cycle times, significant forecast error, and intense competition. As mentioned earlier, the correct focus for this type of business is to remain fleet of foot and incorporate high velocity to their inventory management. The question this paper addresses is "how much is time worth in an organization's supply chain?" How much should the organization be willing to spend to reduce a supply chain's total length whether the change comes from shorter decision cycles or reduced transportation times?

The Fundamentals. To illustrate the mathematical approach to the valuation of time, the first step is to define cycle time. Total cycle time in this analysis is the total time required of a supply chain to respond to a demand surprise. In the case of an unexpectedly hot market, assume that the increased demand exceeds on-hand inventory stocks in both finished goods and raw parts. For simplicity, assume sufficient capacities are available. Exhibit 1 illustrates the various components of cycle time. The first component is a decision cycle. Traditional business operations must first notice the change from forecast, understand that it is real, and react with a modified forecast. This information is than fed into the ERP system and time is added to secure components that enable the increased production. Finally, the increased production is brought on line and delivered through the channel to the end customer. Most managers are unpleasantly surprised to find that the total cycle times for their products can easily reach half the product life cycle!

Figure not available in text-only version of this article.

To appreciate the impact of long cycle times in the arena of unpredictable demand, consider Exhibit 2, which illustrates inventory performance in the face of an unexpected demand increase. The top diagram represents demand while the second chart represents the impact on inventory. The x-axis for both charts is time.

Figure not available in text-only version of this article.

Inventory is shown to be running at a constant level until the demand increase occurs. At this time, inventory begins to decrease at a rate equal to the difference between the increased demand and the rate of supply. This decrease will continue until a decision to increase the build rate can be executed and the increased rate delivered through the supply chain. This is noted as "response time" on the chart and is equal to the total cycle time of the supply chain (30 weeks in Exhibit 1). Note also that upside capacity is required to recover from this scenario. Much like a runner who is tied for the lead and falls down, the runner must run faster than the lead runner to catch back up. The same principle applies to recovering inventory levels. If sufficient capacity is not available, sustained stockouts will result. Finally, note that adjusting inventory levels back to a desired state takes time and the rate of change is equal to the difference between the higher rate of supply and demand. This period is labeled "recovery time".

In contrast consider Exhibit 3, which illustrates the impact of an unexpected drop in demand. In this scenario, inventory continues to increase through the "response time". Should end of life occur during the response or recovery time, excess inventories and the associated costs will result.

Figure not available in text-only version of this article.

Modeling the Financial Value of Time. To begin to understand the valuation of time in the supply chain, the organization must first identify the impact of long cycle times. Exhibit 4 helps to illustrate the impact of the inventory excursions demonstrated in Exhibits 2 and 3. The chart maps the cumulative contribution margin (Y-axis) over a product life cycle (as represented by time on the X-axis). Contribution margin is defined as net revenue less variable costs.

Assume for a moment that an organization had the perfect supply chain and a perfect forecast. In other words, as soon as a unit is rolled off the production line a customer was waiting to pay for the product and take it home. With no inventory excess or shortage, the organization would experience no lost revenues due to shortages or end-of-life costs due to excess. This scenario is depicted in Exhibit 4 as the straight line moving up and to the right. Of course no organization actually has perfect supply chains and certainly doesn't have perfect forecasts! The magnitude of the costs associated with imperfect supply chains and forecasts are demonstrated with the lower line. To begin with, the organization incurs the capital costs of holding inventory. However, the greater financial impact associated with inventories is not having sufficient stock and having too much at the end of life. Stock outs reduce revenue as the customers turn to the competition for product. An even greater impact can be excess inventories at end-of-life that results in deep discounting through fire sales or high obsolescence costs. Understanding the magnitude of these costs, we can now turn our attention to the value of time in the supply chain.

Figure not available in text-only version of this article.

Recall the inventory movements associated with long cycle times and forecast errors that were depicted in Exhibits 2 and 3. Through dynamic modeling, the resulting inventory position incurred when long supply chain cycle times are combined with forecast errors, can be represented. Exhibit 5 depicts this relationship with resulting inventory levels on the y-axis and relative forecast error on the x-axis. Moving left from zero error indicates the market demand is greater than forecast. Moving right from zero, there is less demand than expected. The various lines on the chart depict the inventory excursion associated with the forecast error. As we learned from Exhibits 2 and 3, longer total cycle times result in greater excursions of inventory. Through a dynamic simulation of the supply chain, the peak inventory positions associated with a soft market as well as the shortage of inventory associated with greater than expected demand can be mapped.

A fundamental finding from Exhibit 5 is that the shorter the total cycle time, the more robust business is to forecast error. A five-week supply chain does not stock out even at a 30% error. In contrast, a relatively small forecast error results in stock-outs for the 30-week supply chain. Forecast error in the other direction has a similar financial impact with inventories going out of control for the longer supply chains.

Exhibit 5

Inventory Excursions vs. Demand Shifts: The Impact of Long Flow Times

Figure not available in text-only version of this article.

The question remains as to how to put a value to this relationship. Dynamic modeling can provide the expected inventory performance over time associated with a range of forecast errors. A fairly quick historical analysis will provide a relationship between excess inventory levels and associated write-offs. Coupling this information with the impact of lost contribution margin associated with lost sales, the organization can now apply dynamic modeling to the value of time. Exhibit 6 uses forecast error for the x-axis. However, the y-axis has been changed from inventory to cumulative contribution margin. As in Exhibit 5, each individual line represents a different total supply chain cycle time. Note that at zero forecast error, the length of the supply chain is not a significant determinant of cumulative contribution margin. However, as forecast error grows, there is a dramatic effect on the cumulative contribution margin. The reason for this is the inventory movement identified in Exhibit 5. All we are doing now is including the financial impact on the inventory excursions. Longer total supply chains incur greater lost sales and greater end-of-life write-offs associated with their slow response. In short, longer total cycle times result in supply chains which are less robust to forecast error.

Finally, the value of reducing the supply chain cycle time can be calculated by comparing the various lines on Exhibit 6 at specific points of forecast error. Assume for a moment that the total cycle time was 30 weeks (represented by the lowest line). With a demand surprise of just 20% above forecast, a total cycle time reduction of just 5 weeks is worth $6 per unit sold. The same analysis could be performed for lower than forecasted demand levels. Revisiting the initial question asked at the beginning of this paper, "Would the organization pay 5 or 10 cents more to reduce the supply chain by two or three weeks? If you believe the business represented in chart 6 has any significant forecast error, the answer is a quick "You bet!"

Exhibit 6

Impact of Supply Chain Flow Time on Life-Cycle Product Contribution for Unexpected Volume Shifts

Figure not available in text-only version of this article.

Final Points. First of all, note that this analysis is not intended to capture dime accuracy. It is intended to make relative comparisons and focus the organization's attention on the factors that count. Further, it is important to note that this type of analysis is very different from the traditional cost accounting approach. It does not add up all of the historical costs associated with inventory and apply these with a rate, like peanut butter, to the current inventory levels. Inventory has a purpose in the supply chain and its costs and risks are very much product-life-cycle dependent as demonstrated by this analysis. Don't fall into the seductive mental trap of labeling inventory as evil. Inventory serves a purpose in buffering a business against uncertainty. Although stock-outs, unlike excess inventory, do not show up on any particular manager's report card, it is extremely expensive to your business. Managing inventory properly by applying sound analysis to its value and its relationship to total cycle time is a necessary step in making sound business decisions.


Back to Top