By Ted Schaefer and Alan Kosansky | Supply Chain Quarterly
Most companies calculate total delivered cost (TDC) based on inaccurate and outdated assumptions. Using optimization technology to more accurately forecast TDC by product and customer will help them to improve both their supply chain planning decisions and their costs. Profitability is the engine that drives all successful businesses. To manage profitability, a company must understand and have good control of both its revenues and its costs.
For a long time, companies have had a good understanding of the revenue side of the business at a detailed customer and product level. It is only in recent years, however, that they have begun to understand their costs at the same detailed level by customer and product. To gain that insight, many companies use total delivered cost (TDC)—the complete cost of sourcing, producing, and delivering products to customers. TDC, in turn, has become a critical metric in guiding supply chain planning decisions.
Total delivered cost is indeed an important tool in supply chain planning. However, we have observed that many companies do not make the best use of TDC in their supply chain planning process because of three common errors or oversights. First, many of them use a historical TDC metric, rather than a metric that is based on the supply chain costs they will incur now and in the future. Second, they often make oversimplified assumptions, ignore certain components of TDC, or use average accounting allocations rather than product-specific operational data when computing TDC. And third, companies often fail to use optimization technology to simultaneously make, for each product, the interdependent decisions in sourcing, production, warehousing, transportation, and distribution that will have an impact on TDC and yield the best overall future plan.
We believe that a well-organized business process for collecting and computing TDC that includes detailed, future-looking data and optimization technology will enable supply chain managers to accurately assess the trade-offs among the different components of cost; identify which products and customers should be prioritized; and determine at what price customers should be served.
Benefits of total delivered cost
Total delivered cost is the complete cost of producing and delivering products to your customers. It includes the cost of sourcing raw materials, manufacturing bulk and intermediate products, packaging of finished goods, inventory holding costs, transportation, distribution, and final delivery to the customer. Some elements of TDC are built from per-unit variable costs associated with the specific product and customer, while others are allocated fixed costs from production lines, plants, or other fixed assets used in the manufacturing, storage, and delivery of the product. TDC typically is reported as a per-unit cost for each product at each customer. However, for customers that purchase a large number of products, the aggregate delivered cost or aggregate gross profit are better measures to use to determine the attractiveness of a given customer/product portfolio.
Many companies compute some variant of TDC using historical data. This is a very informative exercise to help a business determine how its profits and losses developed from recent activities. Based on our experience, however, we believe there are a number of flaws in some of these types of analyses. First, we have seen TDC computed at a “product family” or “customer hierarchy” level without diving into the details at the individual product or customer ship-to level. This type of analysis can point the business in the direction of large groups of products or customers that are not meeting profitability targets, but they don’t adequately identify which products or which customers are the real drivers of business profitability. Second, this historical TDC tells a company where it has been, but it doesn’t necessarily tell it where it should be going.
Leading companies have begun to integrate this computation into their supply chain planning processes, especially their network design and analysis activities. By using TDC as a metric in analyzing supply chain network configuration and operations, companies are able to determine the profit margin on each sale at the customer and product level of detail. This enables them to evaluate both the costs of supplying products through their supply chain network and the revenue those products produce, across a number of different network options. Conversely, pairing TDC with expected pricing allows a company to evaluate the impact of imposing a minimum gross profit-margin target on the structure of the network. Thus, a company can consider important business decisions like:
- What is the appropriate price for a particular product/customer location combination to meet profitability targets?
- What is the true walk-away price during a competitive negotiation?
- What is the best allocation of production capacity among different product families?
- What is the impact of increasing volume and dropping price (and vice versa) for a given customer?
- For which products, industries, and regions should we be looking to sell additional volume?
While TDC is not the only criterion in determining the answer to critical business decisions like those listed above, it is an essential factor in arriving at the best strategic decisions for a company. In addition, the ability to identify profitability at a product/customer level provides critical information to effectively maximize profit by ensuring that:
- Constrained capacity resources are directed toward the most profitable products and customers
- Negotiated prices meet target rates of profit
- Data-based decisions are used when sales or marketing managers suggest “strategic reasons” for serving a particular customer
- Growth is directed toward the best products and geographies
Focusing on the product/customer level of detail also improves a company’s ability to manage aggregate profitability across existing and potential portfolios of products and customers. For example, when a large-volume customer is threatening to walk away during tough price negotiations, the real decision for the seller usually is not what will happen if all that volume is lost, but rather how it will impact the business if that volume (or capacity) is sold to a different customer. Having that product/customer level of detail allows the seller to develop a pricing strategy informed by a set of alternatives across the entire network, thus it will be able to forecast the impact of a potential change like the one just described.
Calculating TDC: Do’s and don’ts
As is true with all things measured, precision is overrated, accuracy is essential. Such is the case with computing TDC. Calculating TDC is not a precise science, because there are too many fixed costs involved in the manufacturing and delivery of products that need to be allocated to products and customers. That is no excuse, however, for avoiding the work required to ensure that a TDC calculation is sufficiently accurate.
The consequences of not doing the necessary work are clear. Many companies find it too difficult to get accurate costs or to implement a sound costing methodology, so they end up using oversimplified proxies for TDC that more often than not lead to faulty decision making that is based on sloppy data and/or analysis. Almost unbelievably, we have seen companies build up a “cost to serve” by applying a single average freight rate for all of the products that are sourced from a given plant to any of its customers. Likewise, we have seen many instances when the “average manufacturing cost” is used across all products, regardless of significant differences in processing times or manufacturing difficulty.
There is no question that accurately calculating TDC is challenging. It’s even harder to make sure the analysis produces data that will be useful for planning future business. To help in this regard, the following sections outline the key components of TDC, some common pitfalls associated with each one, and suggestions for the best way to compute each component. Following these recommendations will provide the forward-looking estimate of the TDC that is necessary to make the best supply chain planning decisions.
Manufacturing, packaging, warehousing, and distribution costs: Financial systems often store some form of variable and fixed costs for manufacturing, packaging, warehousing, and distribution operations. Typically these are based on some sort of averages and/or standard costs that do not reflect the true marginal cost of incremental activities. For example, consider a 10-truck distribution fleet that is expected to log 400,000 miles in the coming year. If the combined cost of the truck leases, maintenance, fuel, drivers, insurance, and other factors is budgeted at US $1 million, then the average projected cost would be $2.50 per mile. However, assuming the fleet could serve additional customers, which would require another 50,000 miles, without adding more trucks or drivers, the actual added cost would be significantly less than $125,000 (50,000 x $2.50).
Thus, it is important to properly split the fixed and variable costs in a supply chain to provide the best information to develop future plans. People tend to use the same basic approach when calculating the appropriate planning costs for manufacturing, packaging, warehousing, or distribution operations, but the unique characteristics of those activities should be taken into account. For example, to appropriately capture the TDC of manufacturing activities, it is best to determine which costs are truly variable in the manufacturing operations across the volumes that are “in play” in the analysis; that is to say, costs that would not be incurred if the manufacturing operation does not occur. These are the volume-related variable costs. To illustrate this, let’s look at two different examples.
Plant A has three production lines, each capable of producing 1 million units per month. The decision has already been made to operate all three production lines in the coming year, so there is some level of expense for items that typically are categorized as “variable” in accounting analyses—personnel, maintenance, utilities, supplies, and so forth—that will be incurred just because each line will be in operation. There is also likely a minimum volume level under which a given line cannot be run efficiently. Thus, the total cost to run the line at its minimum volume should be thought of as a fixed cost when it comes to decision-making. Then, the projected variable cost per unit becomes the incremental cost to produce each additional unit on the line over the range from the minimum volume to the maximum capacity of the line. Therefore, we suggest that the variable operating cost should be calculated as:
(total cost to operate at maximum volume – total cost to operate at minimum volume) / (maximum volume – minimum volume)
Depending on the operation, this calculation may be quite different from the typical accounting definition of “variable cost,” as shown in Figure 1.
In the second example, Plant A has the same three lines, but the question is how many lines must be in operation in order to maximize profit or minimize cost. In this example, we suggest making the same calculation for the variable cost and also calculating the total cost that would be saved if a given line were to be shut down. This would represent the “line fixed cost.” Again, this cost may be very different from the costs calculated by a company’s accounting department.
Additionally, those costs that truly qualify as overhead—that is, they occur whether the manufacturing (or distribution) activity happens or not—represent fixed costs for either the asset or the location being used. It is important to get these distinctions correct in order to accurately project the TDC by product and customer.
Finally, when determining any of these costs, consideration should be given to the volume of product that can be moved through the plant, line, warehouse, or distribution fleet on a single shift or straight time versus the volume that can be moved by changing to two shifts or with overtime. Adding a shift will increase capacity as well as the fixed cost, and it may or may not change the variable cost per unit of throughput, whereas adding overtime will generally increase only the variable costs.
Fixed costs: In order to capture the total delivered cost of a specific product to a specific customer, TDC must include allocated fixed costs associated with the assets used to make (and possibly to ship and warehouse) the product. However, there are some challenges companies commonly encounter when making those calculations.
The first is how to properly allocate manufacturing costs. Many companies allocate manufacturing costs by volume, but in many cases that is incorrect. More often than not, capacity is a function of time, and therefore the fixed costs should be allocated by time consumed on the asset. This way, slow-running products will, appropriately, be allocated more of the fixed cost than will fast-running products. For example, a chemical product that has a batch cycle time of six hours for a 5,000-gallon batch should receive twice the asset allocation as a product that has a cycle time of three hours per 5,000-gallon batch. It is important to identify the unit of measure in which capacity constraints exist, and allocate based on that unit of measure.
The second challenge is how to correctly allocate fixed costs when assets are underutilized in some time periods and not in others. Should the fixed costs be allocated by time periods to only those products made on the asset within the given time period, should they be allocated to all production across the entire production horizon, or should it be something in between? There is no one-size-fits-all answer to this question. Making sure the approach is accurate and understandable should be the priority. Most often, the most effective way to make this allocation is to use the same time bucket that’s used in the planning process, because the business will already have a sense of the cost to run the asset for that amount of time.
Inventory carrying costs: The activities of sourcing raw materials, manufacturing intermediates and finished goods, and then transporting them to a distribution center or warehouse all accrue costs that are included in the final TDC calculation. But what happens when the product is held in the warehouse for an extended period of time as part of a planned inventory build-up? The total delivered cost associated with those units of inventory increases because the holding cost of that inventory becomes part of the TDC. And then when it finally comes time to draw down the inventory and deliver those products to customers, that accrued inventory holding cost must be accurately reflected in the product delivered to the customer.
This carrying forward of TDC from the time the product enters inventory until the time it is pulled from inventory and delivered is one of the most overlooked aspects of TDC analysis, especially when used as part of a planning process. Unfortunately, this oversight can have a big impact and distort the TDC picture. A key to using TDC successfully is to ensure that the methodology is able to account for the costs of both creating and holding inventory, and then allocating those costs to products and customers that are served from inventory.
Raw material supply costs: Raw material supply costs are often integrated into standard costs in financial accounting systems, which can lead to poor decisions in a traditional cost-to-serve analysis. It’s not a big problem when each raw material is supplied from a single source to a single location in the manufacturing footprint; in such cases the practice of using standard costs works just fine. However, for materials that are sourced from multiple suppliers and have different costs based on where they enter the manufacturing supply chain—because of differences in pricing across suppliers, freight, duty, or other factors—using average supplier costs will result in inaccurate TDC calculations. Those companies that are able to most accurately compute TDC are breaking out supply costs from the largest-volume raw materials based on bill-of-material consumption, by supplier and by the locations where they enter the manufacturing network.
Transportation costs: On many occasions we have seen businesses understate their total transportation cost by ignoring the cost of the rolling assets, such as rail cars or specialized trailers and containers. We have also seen companies ignore the cost of product inventory that is tied up in a given move. This is of little consequence when a shipment moves from warehouse to customer via parcel express or to a location within a few hours’ drive from the warehouse. For ocean or rail shipments, however, such an oversight can represent a significant inaccuracy in the total cost. As an illustration of the potential magnitude of this error, we found that one of our clients was including only the line-haul freight cost for decision making; as a result, it was understating its total cost by 35 percent.
To avoid these pitfalls, we recommend building “all-in” transportation costs for each lane that will be considered in the analysis. These costs should include the line-haul, rolling-asset lease cost, inventory cost, and any assessorial costs that are routinely incurred on a movement. For example, a rail movement should include the following elements:
- Line-haul freight
- Fuel surcharge, switching, and any other fees paid to the rail carrier
- The cost of the rail car (car lease per day x the sum of the loaded transit, the dwell time at destination, the empty transit, and the dwell time at the loading point before the next reload of the car)
- The cost of the inventory tied up in the movement
- The cost of cleaning the car
TDC and decisions about the future
Up to this point, we have defined the elements of total delivered cost and have suggested an approach for developing estimates of the individual costs that will determine the projected customer/product-level TDC. Now, we will explain how optimization technology, combined with computing TDC at the product/customer level, can produce better supply chain planning decisions.
A leading-edge planning process will investigate and compare a number of different future options or scenarios in order to come up with the best future plan. Different scenarios may include an upside or downside demand forecast, changes in cost or price, plant relocations or shutdowns, or the launch of a new product. In each of these scenarios, an optimization model is used to simultaneously make the interdependent decisions in sourcing, production, warehousing, transportation, and distribution to generate the best path through the supply chain for raw materials to be transformed into finished goods and delivered to the customers.
By definition, optimization is the choice of the best set of decisions to either maximize or minimize a mathematical function that defines the objective of the system. In fact, the “objective function” in traditional network optimization and in production and distribution planning models minimizes the total cost to deliver the forecasted demand—or, put slightly differently, to minimize total delivered cost. At a conceptual level, then, we have been using TDC for years in the network design process. What’s new is the ability to generate TDC at the customer/product level of detail. The ability to view projected TDC at this level of detail opens up a whole new discussion during the supply chain planning process, as the following examples demonstrate.
An example using TDC in the strategic network design process: In a recent engagement, we worked with a client that wanted to reduce the cost to serve its customers while strengthening the specialty segment of its business. Because management had decided not to invest in new manufacturing capacity, any new volume gained on the specialty side of the business would require a reduction in volume on the commodity side. However, because the specialty grades of the product tended to have a higher TDC than the commodity grades, the client needed to ensure that any changes to the split between specialty and commodity grades would result in increased profitability. Thus, our optimization model was developed using projected demands, prices, and costs, and it constrained the manufacturing plants to their maximum achievable capacities. To understand the impact of growing the new business segment we included deliveries to a number of potential specialty customers at a projected price per unit.
We developed a number of future scenarios to compare against the original business plan, which did not include growth in the specialty segment. In addition to identifying significant cost savings by changing product sourcing and by changing the warehouse footprint, we were able to show which customers would be dropped for any given amount of growth in the specialty sector. We also were able to determine the unit price that those customers would need to pay in order to stay above the “cut line.” Those prices were then used by the sales organization as input during the annual bidding events executed by major customers’ procurement departments.
During the scenario analysis, managers on the commodity side of the business argued in favor of using the commodity volume to absorb the plant’s fixed costs, thereby helping to keep the cost per unit across all products at a reasonable level. Because we were able to compute the TDC for each product delivered to each customer, we were able to have frank, data-based discussions and were able to demonstrate that, within the volume shifts we were suggesting, that approach was irrelevant and would not improve profitability.
An example using TDC in the sales and operations planning process: Scenario analysis has become a critical part of a good sales and operations planning (S&OP) process. Often the S&OP process identifies short-term shortfalls of product, and a company must choose which customers to serve, and at what cost. One method used by most companies to address near-term supply shortfalls is to utilize a premium freight service, like air or expedited trucking, to curtail long lead times and deliver enough product to last until adequate supply is once again available. But such freight services are costly, and to make optimal decisions about using premium freight because of capacity constraints in supply or manufacturing, it is necessary to have accurate TDC information by product and customer.
In one case we know of, a large technology company that sources most of its product from Asia into North America and Europe uses weekly TDC cost calculations to manage its premium freight spend. By using TDC to analyze the trade-off of inventory and premium freight, the company is able to more effectively assign manufacturing capacity to the right customers and products. As a result, it has reduced premium freight costs by more than 10 percent—all with no negative impact on overall customer service.
A prediction is a forecast of what is expected to happen, and as we all know, for many reasons forecasts rarely (if ever) turn out to be truly accurate. To minimize forecast error, we recommend investigating a number of different sensitivities around the key assumptions in the model.
There is no single best answer in a strategic network analysis. Rather, there are a number of good solutions, each of which comes with a certain level of risk. Our job as supply chain professionals is to understand which sensitivities to run and how to choose the “most good” scenario that contains an appropriate level of downside risk if the forecast should turn out to be wrong. Thus, we should ask ourselves, “What if my demand forecast is off by ±10 percent?” or “What if the cost of a key raw material increases or decreases?” and other, similar questions before making a final decision.
Leverage the data
Every business wants to both satisfy its customers and achieve the right level of profit on each and every sale. Understanding the specific costs associated with each sale is a key step in managing both the product portfolio and the customer portfolio. Until recently it has been very difficult to estimate the true delivered cost by product and customer, so businesses have been willing to make and work with crude estimates, a practice that likely leads to decisions based on inaccurate or incomplete data.
In today’s world, there is no need to settle for rough approximations that may be far off the mark. The necessary data are readily available, and improved analytics allows companies to avoid the typical pitfalls associated with calculating TDC. Following the recommendations in this article will help companies make more accurate decisions about what they should sell, to whom, and at what price, thereby allowing them to leverage maximum competitive advantage from their supply chain planning processes.