By Bill McBeath and Colin Kessinger

The Shortcomings of Traditional Procurement

In today's twenty-first century global outsourced business world, the traditional and somewhat simplistic approaches used to measure cost for sourcing decisions of direct materials fall short. Procurement history is laden with penny-wise, pound-foolish decisions where the low cost supplier ultimately costs tens or hundreds of millions in lost revenue, lost market share, expedite fees, or write-offs. This goes beyond total landed cost calculations and strategic sourcing metrics. Firms that fail to evolve their method of calculating total cost will find themselves with an increasingly uncompetitive and profit-eating total cost structure.

Before the advent of strategic sourcing, procurement focused primarily on material cost, usually with some consideration of transportation costs. As companies moved from vertically integrated business models to virtual integration (increased out-sourcing, design and build responsibilities spread out across the supply chain) they found themselves buying increasingly complex components and services from an increasingly dispersed global supply base. This has placed much more importance on drivers of total cost beyond the material costs, such as the cost of quality, manufacturability, serviceability, flexibility of supply, ease of manufacturing, etc.

Global Strategic Sourcing and the Total Cost of Supply

As a result, global strategic sourcing evolved as a more sophisticated approach to selecting and managing the supply base and the procurement of direct materials. In strategic sourcing, the relationships are longer-term and there is a drive for continual improvement along many dimensions. Most companies that have adopted strategic sourcing practices have developed in-depth supplier scorecards for tracking performance and driving improvements. However, tools and methodologies for measuring the true total cost of supply have lagged behind. An ideal approach goes beyond supplier performance ratings to accurately calculating the effect of the supplier’s performance on all aspects of the total cost, as illustrated in Figure 1.

Figure 1 - Total Cost of Supply Elements

Total Landed Cost Calculators and In-house efforts

In spite of these challenges, progress has been made in specific areas of total cost of supply calculation. For example, total landed cost calculators are available, often as part of Global Trade or Transportation Management or Strategic Sourcing software systems^[1]. Total landed cost calculators typically factor in things like transportation (looking at different modes, quantities, etc.), customs, duties, tariffs, taxes, fees, and insurance. Some more sophisticated total cost calculators attempt to model other elements of total cost, such as quality, the cost of capital, tooling costs, etc.

In addition, a few companies have made efforts to build their own total cost calculators. Typically, these contain rough estimates and simple formulas for calculating the cost of things like late deliveries and quality failures, which are added onto the material cost in attempt to calculate total cost of supply. While this is a big step over just measuring material costs, these tools lack the depth that typically only comes from many man-years of effort put into best-of-breed software, where significant resources have focused specifically on tackling the issue.

Risk-Adjusted Total Cost Calculation

One of the most promising recent advances in total cost calculation is the availability of off-the-shelf software for calculating the cost of demand/supply matching risks. These risks are typically very hard to pin a cost number on, yet they are often one of the largest elements of total cost.

Consider the example of a typical SMI (Supplier Managed Inventory) program. Most of these contracts specify a min and max setting, assuming a stable forecast and coverage for the FGI (finished goods inventory) and the unique materials that go into the product. As long as the forecast is stable, the expectation is that service levels and inventory turns will be high. However, because the stable forecast assumption is not tested against actual historical fluctuations or forward-looking scenarios and exposures to sudden increases or decreases in demand, the prospect of significant shortages or liabilities is not factored into the total cost analysis.

Which Supplier Would You Choose?

Consider the following comparison. The first supplier offers SMI, but expects the buyer to cover raw material purchases in a supply chain that is 4 months long (to procure and convert raw materials). The second supplier expects purchase orders with a 2 month leadtime because that covers their 2 month long supply chain. Against a stable forecast, the SMI program offered by the first supplier clearly dominates by minimizing inventory. How unstable does the forecast have to be before the second supplier dominates by decreasing the risk of having excess material in the pipeline? A well-implemented framework for quantifying and measuring this work will answer these questions.

How Risk-adjusted Cost is Calculated

As with quality and other metrics included in the total cost calculation, risk metrics are also easier to visualize than to realize. Fortunately, there are tools available today that can help with these calculations. The basic measures in the risk-adjusted Total Sourcing Cost calculation are “fully-loaded” price, inventory/liability costs, and shortage related costs. The fully-loaded cost should reflect most of the terms in the total landed cost model, such as freight and taxes, as well as volume discounts, price floors and caps, restocking fees, etc. This part seems pretty straightforward.

The key distinction in a risk-adjusted calculation is that fully loaded cost and additional metrics are evaluated over hundreds of forward-looking scenarios, so that metrics such as the average inventory level, and average percentage short can be computed across a large number of scenarios. Furthermore, true risk metrics such as the probability that inventory will exceed, for example 90 days, or that shortages will exceed 10%, or that the backlog will exceed 1 month, can also be computed. Price risks can also be projected. A buyer may want to evaluate the exposure to expedite fees on production at the suppliers, or on freight, or the exposure to price increases in a capacity constrained supply market. In companies where these metrics have been successfully introduced, management is specifying targets on both the average performance of the contract, as well as performance against different risk metrics across a range of scenarios.

A CPG Example

Figure 2 shows some of these risk metrics in action for a CPG buyer. The report compares two launch plans and compares them over hundreds of demand scenarios. The report groups results into the lowest 25%, the middle 50%, and the highest 25% demand scenarios (note this grouping of hundreds of scenarios is distinctly different from running three scenarios). The top of the report shows a pro-forma cost statement for each scenario group, and the bottom provides explicit measures for service level and inventory performance.

CLICK ON IMAGE FOR LARGER VIEW

There are several key insights from this report. First, some alternatives may be most attractive due to their performance in the low or high cases. For example, even if the new approach showed a 1% increase in Totals Sourcing Cost in the mid range (currently 2.4% better), the fact that it was 11.8% better in the low range (because it reduced the inventory write-offs from $346k to $234k) and 4% better in the high range (because it reduced shortages from 2.8% to 0.5%) may make the second approach preferable. Second, while performance along the mid-range may look pretty reasonable on inventory and shortage metrics, the exposures to inventory in the low case and shortages in the high case may be completely unacceptable.

Figure 3 shows even greater detail on the inventory story. The following graph shows the distribution of outcomes over all of the demand scenarios considered. The colors, as indicated by the legend, correspond to percentiles. The top of the gray represents the 90th percentile; 90% of the outcomes were below the top of the gray bar. The top of the dark blue bar corresponds to the 75th percentile; 75% of the outcomes were below the top of the gray bar. Revisiting the SMI example discussed throughout this article, this chart might only show a gray bar, suggesting that the 75th percentile of inventory was zero, but the top of the gray bar may show an exposure much greater than zero, just as in the last downturn, when liabilities ballooned to 180 days, and then to 360 days of supply.

On the chart at the left side of Figure 3, we see a typical fashion goods launch strategy; positioning a large supply of FGI (in this case the black line shows that prior to demand a large buffer was installed) to fill the channel and capture the benefits of a successful product. In the remainder of the product life cycle, the legacy of this risky positioning translates into substantial inventory levels. In contrast, the chart at the right side of Figure 3 shows that the suggested alternative substantially reduces the inventory levels, on average and at each percentile.

Building a homegrown model that can do a decent job of evaluating all these variables at once (price, demand, lead times, safety stocks, cancellation fees, storage costs, shortage costs, late fees, buyout options … the list goes on) across numerous scenarios is expensive, time-consuming and usually requires hiring several PhDs. Thankfully, tools for doing this type of analysis are already available from a company called Vivecon (who hired their own PhDs). This kind of system is a big step forward over using gut feel or spreadsheets in quantifying the impact of risks in order to make smarter decisions for sourcing, contractual clauses, launch strategies, and inventory strategies.

Conclusion

Most procurement people know they can no longer afford to be dumb low-price chasers and that they have to become smart lowest-total-cost buyers and relationship managers. While the ideal dream of a system that magically includes every single facet into a total cost calculation may be pie-in-the-sky, there's no longer any excuse for continuing to evaluate only material and transportation costs. Total landed cost tools that can include duties, tariffs, and a variety of other costs are available from many sources. And now there are tools that can calculate risk-adjusted total cost as well. By calculating a total cost that adjusts for the cost of various risks (price fluctuations, shortages, various demand scenarios, expedite costs, etc.) sourcing personnel can be much smarter in evaluating their alternatives. In the end, the organization with the lowest total cost will have a major advantage over their competitors.

[1] A number of companies offer total landed cost calculators, such as PeopleSoft, G-Log, TradeBeam, Blinco Systems, Xporta, Arzoon, NextLinx, Manugistics, i2, Moai, as well as many 3PLs.

©2004 ChainLink Research, Inc.