The problem with
Over the past several decades, numerous industry programs have been developed to improve the flow of goods across the supply chain. Most of these programs have focused on the final link of the chain, between retailers and their immediate suppliers. This link is the first one to feel the impact of changing consumer preferences, and managing it well requires sophisticated forecasting and planning.
Historically, this link has been controlled by independent retailers, who often lacked the tools and information to forecast and plan systematically. Retail replenishment programs are designed to counter these problems and bring order to the final link of the supply chain.
The first generation of programs, which are still widely used today, simply shifted the control of inventories from retailers to suppliers, allowing suppliers to use their superior market data and analytical tools to regulate the flow of goods (see Figure 1). In a consignment program, suppliers decide when to ship the goods, and they retain ownership of those goods until they are sold by the retailer. Although consignment gives suppliers the control they desire, it also makes them wait longer to get paid.
A vendor-managed inventory (VMI) program separates control from ownership: suppliers decide when goods are shipped, but retailers take ownership upon delivery. Although VMI obviously benefits suppliers, retailers also like it because it simplifies their inventory management and usually reduces the amount of inventory they need to hold.
The success of VMI across many industries led to several industry-specific programs for managing retail replenishment (see Figure 2). One of the earliest was the quick response program adopted in the apparel industry in the 1980s, in which electronic point-of-sale systems captured data about clothing sales and transmitted that information to suppliers over electronic data interchange (EDI) connections. Suppliers responded with daily shipments of pre-tagged items that could move directly from their trucks to the selling floor.
In the late 1980s, the apparel industry rolled out an extension of the quick response program known as continuous replenishment. As shown in Figure 2, this program incorporated VMI for tighter inventory control, and it introduced joint forecasting so that suppliers and retailers could pool their understanding of consumer demand to better predict the sales of each product. Because a replenishment agreement acted as a standing purchase commitment, members of the program could eliminate individual purchase orders altogether, saving time and effort all around.
In 1993, the grocery industry launched its own version of continuous replenishment — the efficient consumer response (ECR) program. ECR added category management, which organized replenishment according to groups of products that consumers view as equivalent, and it used activity-based costing to monitor the profitability of each product category.
All of these programs were introduced with great fanfare, and there are countless case studies showing that each has succeeded in reducing inventories and accelerating the flow of goods across the chain. These glowing accounts are bolstered by reports in the business press about the remarkable economies produced over the past two decades through the relentless reduction of inventory. But a closer look at actual inventory levels suggests that there may be a problem with these programs.
Two years ago, James Ginter and Bernard LaLonde of Ohio State University conducted a comprehensive analysis of 8,000 U.S. corporations, looking for changes in inventory levels from 1979 through 1999. The two university professors took as their measure the value of inventory divided by the cost of goods sold (COGS), using COGS rather than sales to eliminate the effects of changing gross margins over the years.
After culling out companies with sales of less than $100 million and eliminating companies that didn’t survive the full 21 years, Ginter and LaLonde performed a trend analysis to determine whether the companies in each of 14 industries experienced a significant change in inventory levels over the period of the study. In addition to looking at overall inventory, the researchers performed separate analyses for raw materials, work in process and finished goods, providing a detailed picture of where inventory changes occurred.
The results of the study are shown in Figure 3, which groups industries according to their overall success in reducing inventory. The five industries in the first group all achieved significant reductions across the board. The next four industries reduced total inventory, but none was able to bring down finished goods. The next three industries failed to reduce total inventory, despite significant reductions in raw materials and work in process (WIP), and the last two showed an overall increase in inventory levels.
What do these results say about the retail replenishment programs of the 1980s and ’90s? They certainly don’t support the common claims for success. The data shown in Figure 3 were collected over precisely the time frame covered by these programs; if the programs were working, then the sponsoring industries ought to be near the top of the chart. Yet just the opposite is true.
The apparel industry, which sponsored both the quick response and continuous replenishment programs, had the second-worst performance, increasing both total inventory and finished-goods inventory. The food products industry, which sponsored the efficient consumer response program, came in dead last, with significant increases in every form of inventory over the period studied.
Further examination of Figure 3 provides some insight into the cause of that failure. Looking across the groups, it’s clear that the major problem lies in managing finished goods; while all but one or two industries achieved significant reductions in raw materials and WIP inventory, only five of the 14 managed to do so for finished goods. Indeed, half the industries experienced significant increases in finished goods, despite attempts to achieve reductions.
Notice also that the industries in the top ranks of the chart sell primarily to other manufacturers rather than to retailers, suggesting that the problem lies in the retail replenishment link. Yet retailers themselves are doing very well, having held the line on finished goods and reduced all other forms of inventory. The conclusion: Companies that supply retail outlets appear to be holding more finished goods than ever before, despite systematic attempts to reduce those inventories.
In discussing their data, Ginter and LaLonde suggest that this surprising result can be attributed to the changing power relationship between retailers and suppliers. When retailers were small and highly fragmented, they purchased goods the way manufacturers wanted to sell them, in quantities that favored economies of scale in production and distribution. Over the past 20 years, the rise of mega-retailers such as Home Depot and Wal-Mart has shifted the balance of power dramatically, and even the most powerful manufacturers now agree to very demanding terms in order to maintain access to these vital outlets. Those terms invariably include frequent, small shipments with short lead times, often directly to individual stores. As retailers have gained power, they have pushed the burden of inventory upstream onto their suppliers.
Even VMI, which suppliers once favored, is being used to push inventory upstream. Wal-Mart is now tying payment for VMI goods directly to its point-of-sale systems, paying suppliers for their products only when those products are actually sold. In effect, this tactic turns VMI programs back into consignment relationships (again, see Figure 1), allowing Wal-Mart to take the affected inventory off its books entirely. This is great for Wal-Mart, but it does nothing to reduce the total inventory in the chain. It merely shifts the burden of ownership one link up the chain.
Of course, suppliers can and often do push part of that burden further upstream by demanding smaller, more frequent shipments from their own suppliers. In fact, that’s a key reason why, as Figure 3 clearly shows, those suppliers have been successful in reducing their inventories of raw materials even as finished goods increased. But here again, the problem of holding inventory isn’t really being solved; it’s just being pushed further upstream.
And there, in a nutshell, is the problem with these programs. In effect, they are playing a shell game with inventory, moving it from place to place without actually getting rid of it. So be cautious about the promises of win-win relationships under these programs. As long as costs are being shifted rather than eliminated, someone has to lose. LT
David A. Taylor, PhD, is the author of Supply Chains: A Manager’s Guide (2004, Addison-Wesley) and the principal of SupplyChainGuide.com.
For more details on the study cited in this article, see:
James Ginter and Bernard LaLonde, An Historical Analysis of Inventory Levels: An Exploratory Study, Ohio State University, Nov. 2001 (www.fisher.osu.edu/supplychain/home.html).