"I will not tell you the obvious," says MIT professor David Simchi-Levi, and he means it. Take the issue of supply-chain flexibility. According to conventional wisdom, the more of it you have, the better off you are. But Simchi-Levi believes that companies can spend too much time and money on achieving total flexibility in their sourcing and fulfillment strategies. It's a question of diminishing returns.

He defines flexibility as the ability to react to unanticipated change in areas such as consumer demand, commodity prices, labor costs and exchange rates. Simchi-Levi doesn't discount the importance of possessing that trait, calling it "one of the most important capabilities a company can invest in." It's vital that a company be capable of reacting to changing conditions without incurring higher costs or slower delivery times, he said at a recent conference in San Francisco on the topic of smart supply chains, sponsored by IBM. (Simchi-Levi is also ILOG chief scientist within the IBM Software Group, http://www-01.ibm.com/software/websphere/ilog_migration.html.)

The question is, how flexible should a supply chain be? Should every plant make every product, to minimize the impact of a shutdown or line failure at any given facility? The cost of such a scheme could be prohibitive, and virtually the same benefits can be derived from a less drastic solution.

Take the case of a food and beverage manufacturer with five plants, each dedicated to making one of five product families. Simchi-Levi looked at five possible scenarios for reallocating production, where a single plant might make anywhere from two to five different products.

The company in question was having trouble devising accurate demand forecasts, so it needed a healthy measure of flexibility to adjust for real-world conditions. At the same time, its existing production strategy was resulting in unacceptably high transportation costs. Clearly, the answer lay in achieving a balance between operational and supply-chain considerations.

Not that much of a balance, it turned out. Simchi-Levi and his fellow researchers discovered that the client could realize 80 percent of the benefits of full flexibility by adding minimal flexibility to the line - that is, having each plant make just two different product families. The company could then juggle the production levels of each product according to need. Having what amounts to a long supply chain "essentially connects each plant with all the others, even though each plant doesn't make all of the products," Simchi-Levi explained. What's more, no single plant is isolated from the others because there isn't an exact match between any of them.

He cited another example of a bottling company that added a relatively small measure of sourcing flexibility, making available an additional 12.3 million cases of product thanks to a reduction in out-of-stocks at the warehouse. In essence, the company added one and a half production lines worth of capacity to its supply chain without any additional capital expenditure.

The same goes for determining the size and character of a supply network. A U.S.-based manufacturer of household goods needed to cut costs by closing some of its 40 production plants around the world. At the same time, it was acutely aware of the risk of having too few sources of finished product. So what was the right tradeoff between slashing overhead and managing supply-chain risk?

At the outset, the company saw that it could cut $40m out of total costs by closing 17 plants. Not a bad day in the executive suite: "Think of how much revenue you'd have to generate to achieve the same effect," said Simchi-Levi.

Think again. With so many plants out of commission, the company was facing longer lead times and a greater possibility of supply-chain disruptions. Furthermore, because nearly all the surviving plants were in Asia, it was likely to encounter problems in serving customers in North America and Europe.

Simchi-Levi proposed a more conservative plan. By closing just 10 plants, and maintaining a redundancy of seven facilities, the company would incur costs that were $2.4m higher than those of the fully optimized strategy. That compared with a $50m penalty for doing nothing. At the same time, it would greatly reduce the risk of not meeting unexpected demand. Simchi-Levi's graph shows a sharp drop in costs with an initial reduction in plant locations; the line levels off and the decline becomes more gradual with each additional closure. The lesson, he said: "A small investment in flexibility can get you almost all the benefits."

Rising supply-chain costs are a constant concern of businesses, especially in today's poor economy. Total U.S. logistics costs as a percentage of gross domestic product have risen 52 percent over the last five years, Simchi-Levi noted. Factors include higher energy prices, pressures on rail capacity, a shortage of truck drivers, and new security requirements in the post 9/11 era.

The trend has prompted many executives to seek out "best practices" for streamlining their supply chains. One such maxim dictates that a company have just one supply chain for the entire organization, a strategy that Simchi-Levi believes isn't always viable for industries such as high-tech and retail. Companies need to look beyond obvious factors such as labor rates to determine the true cost of supporting a product throughout its natural lifecycle. Conventional thinking doesn't always embrace areas such as transportation, packaging and product design. Said Simchi-Levi: "There are times when companies are looking for more than best practices."

Comment on This Article