Restructuring the Supply Chain

In late 2003, Qualcomm’s CIO Norm Fjeldheim started seeing signs that the demand for cellular phone chips was rising much faster than expected. Consumers around the world were snatching up cell phones at an unprecedented rate, and manufacturers of the popular handsets were scrambling for more chips. For Qualcomm, which sells chips to the cellular industry, the unexpected cell phone frenzy hit its supply chain where it was most vulnerable. “Our customers were not getting everything they asked for,” says Fjeldheim, noting that demand for the chips rose 37 percent in one year. “We could not increase our supply, and some deliveries just weren’t possible.”

Out of stock. Whether referring to semiconductor chips or potato chips, these are dreaded words for those in charge of far-flung and increasingly complex supply chains. In Qualcomm’s case, the company not only missed an opportunity to significantly boost revenue but also was forced to reevaluate the way it did business.

Since that period of panic (which lasted for most of 2004), Fjeldheim and his colleagues at Qualcomm have reorganized the supply chain so that they won’t get caught by surprise again. Company officials are bringing supply chain, finance, IT, and sales and marketing together for regular demand-planning sessions. And the company is trying to increase the flexibility of its supply chain by working with multiple suppliers, rather than single suppliers, to build a set of chips. It has also started sharing more information (via Web connections and file downloads) with its 10 chip makers. Should there be another unforeseen spike in demand, Qualcomm will be better able to shift production back and forth between suppliers, if necessary. So far, Qualcomm, which ships 140 million chips per year to cell phone manufacturers such as Samsung and Motorola, has improved its on-time product delivery rate—which had fallen below the 90 percent level during the chip shortage—to 96 percent, a high rate in an industry grappling with shorter lead times for its products than many other manufacturers.

AMR Research says the payoff for companies with high rates of “perfect orders”—those that are complete, in the right place, undamaged and on-time—can be substantial. AMR recently ranked the world’s top supply chains (see “Top 10 Supply Chains,” this page). Companies that ranked high—such as Dell, Nokia and Procter & Gamble—carry less inventory, have shorter cash-to-cash cycle times and are more profitable. A 3 percent improvement in perfect order fulfillment translates to a 1 percent increase in profits, AMR says, while a 10 percent increase means an additional 50 cents in earnings per share. (Qualcomm, with revenue of $4.9 billion in 2004, is too small to make it onto AMR’s list, but the company qualifies as a top supply chain because of its ability to effectively collaborate with handset makers and cell phone service providers, says Kevin O’Marah, a supply chain analyst at AMR.)

To get to this Holy Grail of order fulfillment, however, takes more than plugging data into software programs. In fact, companies that want to achieve and then maintain high perfect order rates may have to restructure their supply chain processes from end to end. They have to build systems that connect them in real-time to both suppliers and customers, thus enabling the development of tighter relationships with both parties. They have to feed information from customers back to suppliers and get a clear forecast of customer demand. And internally, they have to improve long-term demand visibility through constant collaboration between supply chain leaders and the sales and marketing departments.

Striving for perfect orders can be costly. In order to get the most from investments in supply chain technology and processes, companies need to take a targeted and gradual approach. “If you don’t first determine where the gaps are in your supply chain, you might be shooting blindly,” says AMR’s Debra Hofman. In fact, for some companies—for example, those that make commodity parts such as ball bearings or other items that can be stored cheaply—the cost of striving for perfect orders may not be worth it.

For most industries, however, including automotive, electronics and retail, the payoff for high order-fulfillment rates can be substantial. And CIOs who want to be key players in their companies’ strategic planning need to play an important role in overseeing the new supply chain strategy. Since they understand what information technology is capable of, they are in a good position to argue the business case for greater investments in the supply chain. “Supply chain is no longer a back-office activity,” says Kevin O’Connell, director of manufacturing and procurement processes at IBM’s integrated supply chain division. “It has become a potential competitive weapon in the boardroom, and IT is extremely important because technology ultimately enables the various processes.”

O’Connell and others stress that companies must first revamp their business processes and then choose the technologies that support those new processes. “Getting our supply chain in order is 75 percent process and 25 percent tools and technologies,” Fjeldheim says.

In Pursuit of Perfection

Semiconductor manufacturers in Silicon Valley were among the first to use the term “accurate order” in the early 1980s, according to Hau Lee, the Thoma professor of operations, information and technology at Stanford University’s Graduate School of Business. By the 1990s, this focus on measuring order delivery rates had spread to other industries, including food service, which was losing money because food distributors (the companies that deliver food to restaurants and other outlets) would withhold payment for what they considered imperfect orders. Lee says that while manufacturers understand the importance of aiming for accurate orders, many have underestimated the cost consequences of imperfect orders. At Seven-Eleven Japan, says Lee, the company noticed that highly paid truck drivers were often idle because of an imperfect order. As a result, the company started measuring perfect order rates and noticed that when an order is perfect, the driver can leave right away.

Companies trying to boost their rate of perfect orders need to first identify what makes an order imperfect. According to AMR’s Hofman, orders can be imperfect for any of the following reasons: out of stocks, manufacturing and transit delays, late and inaccurate shipments, poor quality of finished goods or damage to finished goods in transit. Companies striving for perfect orders can end up increasing their supply chain costs, Hofman says, especially if they are trying to deliver goods more quickly. Pierre Mitchell, supply chain analyst at The Hackett Group, says that companies should benchmark their supply chain costs against others in their industry and balance that against how important perfect orders are for their customers. Companies that manufacture parts for automobiles, for example, may have a very short window of time for delivery of their product. “If you’re not on time, the vehicle assembly line may shut down,” Mitchell says. If a manufacturer is not serving a customer well in this type of scenario, it could face steep financial penalties. However, if a company is making a commodity item, such as packaging material that can be stored at little cost, achieving perfect order is less important.

For those companies that decide perfect order fulfillment is worth pursuing, the process starts with demand forecasting, a famously unscientific approach that some have compared to gazing into a crystal ball. Most experts agree that demand-forecasting software alone can’t guarantee an accurate forecast. As a number of well-publicized mishaps at companies such as Nike have shown, ensuring the accuracy of the data going into the forecasting algorithms is key to avoiding expensive miscalculations.

So companies are changing their supply chain processes to get a better view. At Procter & Gamble, for example, the company started to revamp its supply chain four years ago to increase efficiencies and focus on getting a clearer picture of demand from actual consumers to supplement the traditional use of historical sales data. The idea is to provide information to suppliers on what is sold today through aggregate point-of-sale (POS) data. P&G, of course, has been exchanging POS data with some of its top retail customers, such as Wal-Mart, for the past 20 years through its automated replenishment system. When a cashier at Wal-Mart sells a box of Tide, that data flows back to a P&G distribution center. Once a certain number of boxes has been sold, a trigger goes off to send another shipment of Tide to the appropriate Wal-Mart distribution center so that it can be shipped to the store. But today, P&G is working to provide its suppliers (the companies that make ingredients for P&G products) better access to production data. The company is doing this by collecting POS data from its customers and then transmitting production requirements to suppliers on its supplier portal. “This allows [the suppliers] to better plan their production, and deliver raw and packing materials on a just-in-time basis,” says Patrick Arlequeeuw, vice president of the global supply network at P&G.

Collaborating End to End

In the past, when a retail customer saw that a product was missing from the shelf, the retailer would check his back-office inventory and the distribution center. If necessary, he would then place an order with P&G, where someone would look at inventory levels to see how quickly she could get a shipment of Tide to the retailer’s distribution center. Then, based on the historical sales data, P&G would contact suppliers for needed ingredients and packaging materials to avoid a repeat out-of-stock experience.

Now, for an increasing number of products it sells, P&G aggregates the POS information from the store and provides its suppliers with this information through a supplier portal linked to P&G’s proprietary SAP system and a customer connect portal. “Everyone in the network works from the same data,” says Arlequeeuw. The result is that P&G’s average out-of-stock rates have fallen from 20 percent of brand categories with a 10 percent out-of-stock rate, to just 7 percent of categories experiencing that rate. Over the same period, says Arlequeeuw, the percentage of categories that are achieving an out-of-stock rate of less than 5 percent has increased from 43 percent to 60 percent.

The problem with P&G’s old supply chain model was that replenishment of empty shelves could take several weeks, or even months at times. “Now, because we provide information to our suppliers on a daily basis, they understand what is happening and can replenish shelves right away,” Arlequeeuw says.

P&G’s demand forecast, while still not 100 percent perfect, comes closer to delivering accurate orders because it contains information about what is happening on store shelves on a daily basis—information that is ultimately transferred to suppliers. The goal, says Arlequeeuw, is to have enough flexibility in the supply chain so that plants can produce for tomorrow what was sold the week before.

This approach should give P&G and other companies a better chance at nailing accurate customer demand. And companies that have better demand forecast accuracy will have less inventory, stronger order fulfillment and higher profit margins in the end. Hofman says P&G is moving toward having its entire supply chain capability driven by demand. P&G is aggregating the POS data and sending it to their production plants. “This is not yet widely done,” she adds.

It was at a companywide strategy session in August of 2004 that Qualcomm executives—still smarting from the surprise jump in chip demand—decided it was time to shake up the company’s supply chain. To get a better handle on supply and demand, the company decided to merge its supply-planning group with its demand-planning group to create one integrated supply chain organization. Where in the past the separate planning groups worked on their reports in relative isolation, each region now collects forecasts from customers and reviews them with Qualcomm’s marketing forecast.

Qualcomm’s top management, including its CIO, recognized that if the company wanted to boost its order-fulfillment rates, it needed a longer-term demand forecast that took into account suppliers’ capacity capabilities. And to this end, the company would need to encourage collaboration not only in the supply chain division but across the organization. Qualcomm would have to produce a more detailed and frequently updated demand plan to serve as a road map for the entire company. First, the company had to beef up its sales and operations planning (SOP), a process that requires supply chain, finance, product marketing and sales executives to work together to devise a more accurate demand plan. That involves plugging all forecasting and demand data into Oracle demand-planning software, which has reduced errors and now allows for more analysis of the forecasts than the previous Excel-based process. “Ultimately, a company needs one system that represents the truth,” says Mitchell of The Hackett Group. “Through sales and operation planning, the company can take its forecasts and turn them into a demand plan that everyone in the company will march to.”

Bringing Suppliers into the Loop

At Qualcomm, supply chain executives meet on a weekly basis with sales and marketing to review the forecasts and prepare for a larger, monthly gathering. The monthly planning meeting brings together the company’s top management with finance, sales, marketing and supply chain to discuss and analyze an overall demand forecast. “The end goal is to get a longer view of capacity requirements,” says Vinay Asgekar, Qualcomm’s senior director of supply chain management. Through this change in process, Qualcomm now creates 12-month SOP forecasts, and plans to extend that to 18 months in the near future. “The mind-set has changed completely when it comes to supply chain,” says Asgekar. “Now we know we have to get a much longer view of demand in order to make these perfect orders.”

At P&G, sales and operations planning is even more important now that the industry is moving toward 20 percent to 30 percent of goods on promotion or sale, says Arlequeeuw. Promotions can easily boost demand and lead to an out-of-stock at the store. That means that sales and marketing people, who are in charge of promotions, need to share this information with logistics and supply chain in order to put together the best possible demand plan.

“Sales, logistics and general managers need to work together to get an aligned picture of demand,” says Arlequeeuw. In general, key leaders within each product category at P&G meet at least once a month to go over the demand plan. But some units, especially those facing pressure from consumers to frequently offer new and innovative products, meet more often. For example, the cosmetics division, under consumer pressure to produce new items on a constant basis, gets together on Mondays to exchange demand data.

In addition to improving collaboration within the company, organizations seeking higher perfect order rates need to share information on promotions and demand with suppliers. And that goes for more than just manufacturers. At West Marine, the largest boating supply retail chain in the United States, company executives learned this lesson the hard way. Trouble at the California boating giant’s supply chain started as early as 1996, after the company acquired a competitor, E&B Marine. New distribution centers were poorly connected, and out-of-stock levels during peak season in 1996 climbed to more than 12 percent. At the same time, sales dropped by almost 8 percent. By 1998, with earnings sagging, the company’s board of directors brought on a new CEO who quickly recognized that the supply chain was broken and moved to turn the boating company around by putting together a team to fix it.

Larry Smith, West Marine’s senior vice president of planning and replenishment, says the company was able to right itself through a multiyear effort to link stores with warehouses and provide demand forecasts to its primary suppliers. Although he does not measure perfect orders in the same way that Qualcomm and other manufacturers might, he says that efforts to link sales data to suppliers have dramatically increased on-time replenishment rates. Over a period of three years, the boating supply company integrated its distribution center with store-level replenishment systems, linked to suppliers with EDI, and invested in a range of vendor and customer-forecasting and replenishment systems that connect stores to distribution centers and suppliers. Smith says West Marine is now collaborating with 200 suppliers, 70 of which are loading its forecasts into their ERP systems. As a result, the retailer has been maintaining peak-season in-stock levels at its stores at 96 percent or better.

Five years ago, the company’s on-time fill for vendor orders at its distribution centers had sunk to between 30 percent and 40 percent. Now, they are at 80 percent, and Smith attributes the increase to the company’s ability to quickly share sales information with suppliers. Every night, the company electronically polls all of its 390 stores for sales, updates the in-store replenishment systems and, with that data, projects sales and orders. West Marine sends the aggregated forecast data to its vendors via automated e-mail. Every Monday, the company sends in-stock reports to 200 suppliers. Then, with a year’s worth of data, it creates an annual forecast, runs order calculations for distribution centers and sends its order forecast to suppliers. On a quarterly basis, company officials measure the accuracy of vendor orders.

According to Stanford’s Hau Lee, West Marine’s internal aggregation of the data helps suppliers understand demand according to specific regions and seasons. For example, boating supply sales in Boston in the summer will go up, whereas demand might be spread across the year in Florida. “They are helping their suppliers make sense of their POS data, and their on-time delivery performance has risen because of that,” Lee says.

Qualcomm is also working to improve the way it shares information along the supply chain in its effort to boost its perfect order rate. Right now, Qualcomm shares information with suppliers via Web connections, or by sending them downloaded files. Over the coming year, the company plans to build an “e-hub” to connect to its growing stable of suppliers, although they have not yet decided whether to go with an EDI-type model, build their own system or use some sort of a mix. Fjeldheim and Asgekar say that while they are better prepared today to react to a sudden spike in demand such as the one that shook the company in 2004, Qualcomm is “not fully there,” and has more work to do to better understand its suppliers.

The goal? More perfect orders and better profits.

“As you improve your perfect order rate, you squeeze losses—such as extra inventory and wrong quantities shipped—out of the system,” says P&G’s Arlequeeuw. “What we’re seeing is that perfect order can lower your costs if you do it right.”

Senior Writer Susannah Patton can be reached at spatton@cio.com.