APICS —The Performance Advantage

November/December 2002
Volume 12 No. 10

Letters of Demand
The 3 W's and the 4 R's are key to improving demand responsiveness.
By Jeff Kavanaugh

Being able to offer the right product in the right place at the right time for the right price—getting these four R's right can be an elusive goal. Investment in information technology has been poured into addressing this problem, first with the wave of enterprise resources planning (ERP) systems and then with advanced planning and scheduling (APS) and supply chain execution systems. However, inventory and product availability are still major issues for products companies. In 2001, U.S. private inventories totaled nearly $1.4 trillion. Even a slight increase in the effectiveness of that inventory investment would free up substantial funds to drive higher product availability and lower working capital.

A slow demand environment certainly plays a role in increasing inventory levels, but frequent stories of stockouts and persistent high costs to maintain availability suggest that companies are still struggling to become more demand-responsive. For example, a recent study of convenience stores shows that stockouts are a significant problem for convenience store operators. During an average day in a convenience store, 9.2 percent of the items in the 15 categories studied are not available to the consumer. Even staples like bread and milk exceeded a 10 percent stockout rate.

Demand management (DCM) is a practical philosophy of forecasting, capturing, stimulating and responding to demand in an integrated, enterprise-wide manner. For products companies, managing by demand means quickly communicating demand changes throughout the firm and finding the best combination of supply chain efficiency and responsiveness. The approach a company takes to answering these three questions—where, why, and when—determines appropriate target inventory positions and directly influences the company's ultimate success. We'll examine the challenge of inventory optimization from a fresh perspective and show how the three W's can improve demand responsiveness.

At first glance, getting the 4 R's right appears to require a crystal ball and absolute control over supply networks. Forecast accuracy is certainly a major driver of inventory. However, many companies have come to acknowledge the existence of uncertainty in demand and lead times, and that there is a practical limit to how much uncertainty can be reduced. They have also accepted the idea that responsiveness is as important as—and more controllable than—forecast accuracy as companies move toward demand-driven replenishment. For products companies, inventory policies play a critical role in designing and maintaining responsive supply chains.

Analyzing uncertainty and inventory components—form, purpose, and time—is required to achieve demand responsiveness, but companies first need to determine the correct supply chain design for each of their products. Based on research in this area, Dr. Marshall Fisher developed a tool to align product strategy with the makeup of the physical network.

According to Fisher, there are two types of products, functional and innovative. Functional products have predictable demand, long product lifecycles, and typically low margins. Innovative products have relatively unpredictable demand and short product lifecycles but carry higher product margins and initially appear more attractive than functional products. Functional products require efficient supply chains, which emphasize low cost and minimal inventories. Innovative products require buffered, quick response supply chains. This cost tradeoff needs to be considered during the development of product positioning.

The functional vs. innovative product distinction is also important for inventory considerations. Strategic decisions about inventory budgeting, buffer placement, and deciding "how low can you go" starts with supply chain design and aligning product characteristics with network parameters. In the case of toothpaste, Procter & Gamble found that consumers did not really want 28 different types of toothpaste; at least they were not willing to pay a premium for that level of variety. The company recast its toothpaste as a functional product, reduced the number of stockkeeping units (SKUs), and focused on supply chain efficiency rather than innovation to drive margin.

Once strategic supply chain decisions are made, then the 3 W's-where, why, and when-must be considered:

• Where—inventory may exist in multiple echelons or stages throughout a physical network
• Why—there are several purposes for inventory, and each one is affected by different drivers and mathematical relationships
• When—inventory positions must be determined for specific points in time, especially when considering seasonality.

Together these three dimensions form a "meta-SKU," a term coined by Dr. Sridhar Tayur and resulting from his research in the areas of inventory and supply chain management. Dr. Tayur, a professor at Carnegie Mellon University and founder of SmartOps, is an acknowledged thought leader in these areas and has spent more than 10 years studying inventory and supply chain problems. In 1999 he helped produce a book on the subject, Quantitative Models for Supply Chain Management. This research has been validated with several Fortune 500 companies and has recently been captured in algorithms that model the 3 W's we will now examine.

Inventory levels are the result of business policies, and proper analysis requires representing inventory at each stage and in each location throughout the physical network. The most common forms of inventory include raw materials, work-in-process (WIP), postponement, and distribution. Distribution inventory could include vendor-managed inventory and physical or virtual pools.

The physical networks of large companies have many combinations of suppliers, manufacturing plants, distribution facilities, and retailers or dealers. The demand-supply network depicted in Figure 1 illustrates how inventory is held at multiple points throughout the conversion of raw to finished and delivered goods. Also note the lead times for each major stage in the network, as inventory must be held to avoid starving successive resources. Customer segmentation and numerous selling channels also influence inventory location and quantity through their differing lead-time requirements. Companies do not have comprehensive visibility into inventory due to the different forms, multiple echelons, and locations within each echelon. Each form of inventory is driven differently from the others and should be analyzed using different techniques.

Figure 1

The other key factor in analyzing inventory forms is simultaneous consideration of multiple stages or echelons. Companies need to move from optimizing a single location to address their entire network.

• Why: Purpose. As Tayur's research indicated, total inventory at a given location is driven by specific purposes, each of which influences buffer requirements in a different way. These purposes include:
• Cycle: stock held to address the period of time between replenishment orders; this may also be driven by the purchase of raw materials in larger quantities than needed to achieve volume discounts

• Demand safety: finished goods held above the level of demand as insurance against the risk of stocking out, protecting the firm from unexpected changes in customer demand
• Supply safety: stock held as buffer against variability in capacity or materials
• Incubation: : stock held until meeting quality or regulatory requirements for waiting periods, such as post-production seal testing for food containers

• Pipeline:stock in transit between holding points in the physical network or due to economies of scale offered by transportation companies

• Pre-build: building up finished goods for seasonal demand in limited capacity environments to enable consistent production rates

• Merchandising: assortment of items displayed at or near point of purchase in order to increase sell-through.

These examples illustrate that all inventory is not created equal, and that the considerations involved for each type of inventory may vary significantly. For each purpose, tradeoffs must be considered between the amount of inventory held and other factors relating to demand and supply-related costs. Making well-informed decisions about these tradeoffs requires sound policies, procedures, and measurement systems.

Time. There is also a time element to inventory, as the profile for each meta-SKU varies over time. Time is a distinct attribute for each item carried in inventory. All else being equal, greater amounts of demand and supply uncertainty will exist as meta-SKU analysis moves from the current time to consider points in the future. Other time-based considerations affect inventory as well, for such situations as seasonality, new product introductions, and intermittent demand (like low-volume service parts).

Inventory analysis tools Most major products companies have implemented ERP systems, and in most sectors they have also widely adopted APS systems. However, many do not create and maintain effective inventory policies. As a result, these companies do not know their optimal inventory profile or the impact of having too much or too little inventory at each meta-SKU.

To properly set targets, companies need to perform inventory analysis that takes into account uncertainty and the meta-SKU attributes previously discussed. Inventory and its drivers can be analyzed to show each inventory form and purpose across relevant time periods. The output of this analysis establishes a baseline condition of inventory, availability, and cost parameters.

In Figure 2, the baseline condition shows total cost and service level tradeoffs for a company's current situation. This can assist a company in determining where it wants to be along the total cost vs. product availability frontier-the most appropriate availability for a given product, based on total costs (a). The analysis can also show how improvements in each driver can move the cost vs. availability frontier to the left (b), reducing cost for each level.

Figure 2

A common inventory measurement, inventory turnover, fails to capture the effect of product availability, as it is driven by inventory in relation to cost of goods sold. A better metric is return on inventory, defined as profit for a period divided by average inventory value during that period. By using profit as the numerator, both inventory position and service levels are considered, and the result is more meaningful. Profit also captures costs for promotions and other stimuli when used to stimulate demand and draw down inventory.

The basic analysis components just described are standard operations management computations and can quickly establish baseline conditions and sensitivity to changes in each driver. However, it is also useful to consider additional measures of demand responsiveness. This includes performing ROI analysis at a divisional or product family level to account for differences in business environment and strategy. It is also useful to analyze profit sensitivity to changes in operating costs and revenue.

Taking action. What should you do to improve your company's demand responsiveness through inventory optimization? First, understand how competitive product availability is defined for each of your products. Then match that availability requirement to your business strategy. Depending on the situation, your approach to demand responsiveness and inventory optimization will vary:

• Pursuers of market share need to avoid the temptation of setting service levels so high that profitability is sacrificed on the altar of fill rates. Analysis can provide insight to find the appropriate point on the availability vs. total cost frontier and potentially move the frontier to achieve lower total costs at each service level.
• Profit seekers need to avoid the trap of arbitrarily setting inventory positions and fill rates to meet short-term profitability objectives. Service-level objectives and target inventory positions need to be determined by understanding customer requirements and the firm's response capabilities.
• The "inventory is bad" camp needs to take a fresh look at inventory and view it for what it really is-a current asset that buffers supply processes from the lead time and variability of demand.