Management Accounting

Your Supply Chain Has a Weak Link

A strong SCM system is essential to a strong bottom line. Here are ten signs that some part of your supply chain isn't pulling its weight.
Jennifer CaplanMay 21, 2003

Have any “sure signs” of your own? Send them to [email protected].

1. Departmental “Silos” Create Conflicting Goals

When a company’s director of inventory keeps tabs on safety stock, the head of production maximizes uptime at the plant, and the procurement manager drives down costs with a key supplier, you’d think that company would be in pretty good shape. But when a company monitors supply-chain performance solely by using departmental metrics, it may end up driving locally minded “silo” behavior at the expense of the overall smooth functioning of the chain.

Drive Business Strategy and Growth

Drive Business Strategy and Growth

Learn how NetSuite Financial Management allows you to quickly and easily model what-if scenarios and generate reports.

The director of inventory, for example, might pull a day of safety stock out of the system to reduce carrying costs. That might go a long way to help him meet his departmental targets — but it could also jeopardize the company’s ability to respond to an unforeseen spike in customer demand. (Some companies, such as Sun Microsystems, go so far as to align their incentive programs with broader operational metrics.)

Even when metrics are in synch across the company, departments may fail to communicate properly, or fail to exchange critical information altogether. Sales and marketing, for instance, may offer discounts on certain products to drive demand, but without informing the manufacturing department. When manufacturing notes the increase in sales, it might ramp up production in the belief that demand has risen. (Demand has risen, of course, but only temporarily, and not at a price that the company plans to maintain.)

“You’ve got different people, a bunch of complex business processes, each with different objectives and tons of data, explains Mike Mansbach, a senior vice president at SeeCommerce, a provider of supply-chain applications. Without aligned metrics and good communication, “you start to have a lot of dysfunction in the supply chain.”

2. Dirty Data Obscures Your View

Ultimately communication can be no better than the information that’s being communicated. So when databases are poorly connected, or when they’re updated on different schedules, it can be nearly impossible to get a comprehensive view of supply-chain performance.

“If you make a decision before getting a full view from all your systems, it could be turn out to be disastrous, ” says Mansbach. Many companies, he adds, have gone to the opposite — and costly — extreme of “maintaining legacy systems simply because they don’t know whether they are valuable. They’re collecting data at very high costs in maintenance and licensing fees and labor that has little value.” (To combat the problem of “too much data,” some companies have turned to new methods of enforcing data-growth and data-retention policies.)

That dirty data — even seemingly minor inaccuracies like misspellings, typos, and missing information in database fields — can actually subtract value by throwing forecasts and calculations well off the mark. That’s especially true when dirty data is duplicated or fed into further calculations, creating the “bullwhip effect,” in which the amplitude of the error increases as you move further down the supply chain.

“Perhaps the biggest problem with forecasting accurately is data cleansing,” says Katrina Roche, COO at Evant, a planning and replenishment software provider. “Data integrity and validation,” agrees Kim Perdikou, CIO of Juniper Networks, a provider of Internet systems, is an inescapable “part of the maintenance.”

In the early ’90s, when companies began to automate business processes, much data cleansing was done manually. Some still do, notes Roche, describing one case in which “it took one customer six months to cleanse all their data.” Many software providers, such as Evant, have developed tools using mathematical algorithms that help cleanse data automatically, by cross-referencing disparate data sets, synchronizing overlapping information, correcting inconsistencies, and removing spikes and “outlier” information that might throw off forecasts. Adds Roche, “We ran the same data through our system in a matter of hours.”

3. Your Inventories Are Out of Whack

“Many of the companies we work with have some level of inventory problem,” says Mansbach of SeeCommerce, “but they’re usually not sure exactly where the problem is.”

Poor visibility and poor communication are obvious candidates. Managers who don’t have good visibility into customer demand, notes Jim Fuller, controller of global product services at Cisco Systems, often obey the natural “cover your back” instinct, especially when they can’t afford having a part out of stock. The lower the visibility, the higher the level of safety stock — and the higher the cost of carrying it.

If demand for a given item is particularly volatile, of course, that item might require a higher level of safety stock. And if it can be obtained at a huge discount (with reasonable assurance that it won’t become obsolete) then it may pay to stock up.

But you can have “too much of a good thing,” according to analysts at Forrester Research, who note that up to half of most companies’ inventory is safety stock. Part of the problem is that managers don’t reassess optimal stock levels frequently enough. In many cases, companies will calculate safety stock levels once — usually during ERP implementation — and stick relatively close to those levels thereafter. (No one wants to face anything like “The Great Inventory Correction” ever again.)

Recalculating safety stock dynamically — taking into account changing demand patterns and lead times — can allow companies to free up cash and reallocate resources. What’s more, by maximizing inventory turns, pruning obsolete inventory, and maintaining more efficient stock levels, companies can reap more generous returns on their assets.

Fuller has seen the benefits of a quick-response system first-hand. Managers at Cisco create statistical models of when parts may fail so the company can have replacements available to ship to customers from a specific depot — often within a two-hour timeframe — and hold on to key business.

To help it stay within that narrow window for shipment, Cisco implemented software from MCA Solutions, which can alert inventory managers of potential fluctuations in demand for routers throughout the country. Since the system has been in place, Cisco’s inventory investments have decreased between 10 percent and 15 percent, and service orders have risen by the same amount, explains Fuller, who adds, “We’re making better decisions now.”

4. Your Demand Forecasts Miss the Mark

Traditionally, forecasts assume that demand is relatively fixed. If a thousand units of any given product were sold last year, the assumption is that a similar amount will be sold the following year, perhaps with a small percentage increase based on subjective rules of thumb.

It’s also true, traditionally, that demand tends to have a tenuous correlation with historical figures, and that it’s often subject to significant and unpredictable fluctuations. What’s more, demand forecasts are often made on an aggregated basis. The key to more-accurate forecasts, explains Forrester Research analyst Noha Tohmay, is not just making sure that shelves are full, but gaining a more nuanced understanding of customer segmentations and measuring the profitability of demand for different items.

Demand management software providers such as Evant help managers track demand patterns for individual products at different stores, by helping monitor the flow of goods from warehouses and distribution centers through retail stores and customers. Evant’s software reforecasts the demand for every item in each store every week, and recommends inventory levels and other guidelines for individual products to ensure that supply matches demand as closely as possible.

Armed with more detailed point-of-sale data, suppliers can gain a more accurate picture of what retailers will order in the future. Store data can be aggregated if need be — say, to match up retailer and supplier distribution centers. Demand management software can also smooth out forecasts by allowing for irregular events like product promotions and seasonal changes. (Many companies are looking at all sorts of ways to use all the available data.)

Since going public nearly 10 years ago, O’Reilly Automotive has seen nearly a fourfold increase in the number of its auto-parts stores. To help maintain control of its expanding supply chain, O’Reilly turned to Evant. Since doing so, it has been able to increase, and better track, the number of stock keeping units (SKUs) handled by each of its buyers from 40,000 to 100,000.

Evant’s software also allows O’Reilly to take in data from outside sources to make its projections more accurate. For example, the auto company now integrates registration data — such as location, model, and age — from R.L. Polk, a provider of consumer marketing intelligence for the automotive industry. O’Reilly has been able free up $66 million in capital over the last few years through inventory reductions and other efficiencies.

5. Your Time-to-Customer Is Too Long

In the 1960s and ’70s, manufacturers competed on the basis of cost efficiency. In the 1980s, quality was the rage. Today, much of the focus is on speed. Companies that outsource manufacturing and stretch their supply chains around the globe, however, are finding that time-to-customer is becoming more and more difficult to trim.

There’s also a fine balance between getting products out the door more quickly and delivering high quality. And as they face increasing pressure to satisfy customers, managers must also be able to make realistic promises; failure to deliver against expectations can be one of the most deadly blows to customer loyalty.

At Internet systems provider Juniper Networks, determining a delivery date for customers once required four days, while managers tried to understand all the requirements and run calculations manually in a spreadsheet. They’ve been able to reduce the time in half with software from Valdero, whose products help companies respond more quickly to fluctuations in demand and supply.

For its part, ColorCon, a producer of pharmaceutical products, turned to Oracle to help it reduce time-to-customer, which once ranged from eight to ten days. ColorCon’s current range is five to seven days, attained in large part because it can provide suppliers with more-accurate demand plans, says CIO Perry Cozzone. “Out goal is to reach five days worldwide,” he adds. (What’s the effect on working capital? See for yourself with the CFO PeerMetrix interactive scorecards.)

It’s essential to strike a balance, however, between satisfying customers quickly and creating realistic expectations. Many expedited orders “might indicate that you need to look again at the time you’re allotting in the first place,” maintains Scott Wilson, project manager at Strategic Systems International, a supply-chain-management software provider.

Too much emphasis on faster delivery at the expense of careful planning can be a recipe for disaster, says Juniper’s Perdikou, especially as companies move to automate more facets of the supply chain. “If you have any problem,” she says, “automation just makes it a faster problem. The challenge is finding a healthy balance.”

6. You Don’t Keep Close Tabs on Your Suppliers

Many companies today don’t measure supplier performance frequently enough, says Mansbach of SeeCommerce, and most still use manual processes to do so. Managers may want to rely wholeheartedly on their suppliers, true — but it’s also true that it can be easier to let long-term, entrenched relationships continue even if suppliers are not performing optimally, since the pain of switching can be greater than dealing with a couple inconveniences here and there.

Companies that produce supplier report cards once a year, says AMR Research analyst Larry Lapide, aren’t watching close enough. “A lot of them are old by the time they’re completed” and the company discusses matters with the supplier, he notes. More-frequent supplier report cards allow managers to identify problems before they fester too long. (Exactly what might constitute a problem is often a matter of some dispute between IT and non-IT executives.)

Identifying problems quickly is especially important to ColorCon’s Cozzone, who concedes that “We’re not an organization that can switch suppliers quickly.” Using on-time delivery as a key metric, the pharmaceutical company measures supplier performance, on a monthly and quarterly basis, to determine whether negative trends are developing. If so, “we will do everything we can to work with the vendor to remedy the problem,” says Cozzone.

Mansbach adds that even companies that evaluate suppliers more often tend to do so using functional metrics — like on-time delivery — rather than ones that are more closely correlated to customer service. Avoid those “silo-based” metrics, he cautions, and focus on companywide performance objectives.

7. Your Suppliers Aren’t Responsive Enough to Changes

Even if you keep close tabs on your suppliers, their inability to respond quickly to operational changes can mean a missed market opportunity. “Companies need to have a strong collaborative relationship with their suppliers to ensure that inventory is going to be where it needs to be to meet future demand,” stresses Forrester’s Tohmay.

But developing closer relationships with suppliers also means streamlining the supplier base, speeding up the supply negotiating process, eliminating redundant purchasing processes, and above all, aligning internal performance metrics to those of your suppliers to ensure everyone is working toward similar goals. “If you have a metric that is shared with major suppliers, it will cost them less to provide the goods you need to fulfill your customers,” notes SeeCommerce’s Mansbach. (By contrast, companies that compel their suppliers to adopt certain business practices, such as joining an online trading exchange, could be damaging their supply chains.)

When Sun Microsystems began to outsource more of its procurement activities and depend more heavily on contract manufacturers to assemble its products, managers needed a system to help them track their relationships and communicate with partners more efficiently.

That’s when the company turned to collaborative planning applications from Manugistics. The software gave Sun a centralized system that allows managers to share information about more than 100 products — such as demand and supply, orders, and shipments — with its broad supplier network.

What’s more, Manugistics’ collaborative planning software provides real-time notifications of unplanned events and helps managers better anticipate supply shortages before they become a problem. Since it began using the software, Sun has been able to reduce its product cycle time from several weeks to days.

8. You Don’t Consider “Customer Service” a Key Metric

In recent years, many companies intent on improving their supply chain management have focused on capital intensive, upstream processes such as procurement, logistics, and manufacturing. Far fewer have begun downstream, with customer satisfaction levels and demand patterns, then worked their way back up through the supply chain to ensure that demand is met.

As customers become more selective, and maintaining a competitive edge increasingly means keeping customers happy, satisfaction rates are becoming the main drivers of supply chain management strategies. “Supplier management is important to us, says ColorCon’s Cozzone, but it’s “nowhere near as vital as understanding our customers and being proactive with them.”

One reason that customer satisfaction has been underemphasized is that it can be such an amorphous concept. Measuring satisfaction, concedes SeeCommerce’s Mansbach, is as much an art as it is a science. SeeCommerce has been helping companies take a demand-driven approach to their supply chains by beginning with customer satisfaction targets and translating those into departmental metrics throughout the supply chain.

A company that wanted to achieve a 96 percent customer satisfaction level, would first convert that figure into the requisite days of safety stock, as well as days for product development (if any), manufacturing, and delivery.

If five days of safety stock was the optimum level to help achieve 96 percent customer satisfaction, the executive in charge of inventory would need to be able to quickly identify if stock levels had fallen to four days (which might cause the company to miss a delivery date) or risen to six (which might mean that working capital would be tied up in inventory and thus unavailable for use elsewhere). After identifying an exception, the executive would still need to hunt down the cause of the problem and return safety stock to the appropriate level. (Some companies have begun segmenting their customer base so they can devote more resources to their more valuable customers.)

9. You Ignore the “Custom” in “Customer”

As customers increasingly demand tailor-made products, customization is starting to make a stronger mark on the demand side of the supply chain. The better to meet customer demand, more companies are adopting “postponement” strategies — holding off production until after an order has been placed. Think Dell. The computer company will stock software, CPUs, and other components but won’t assemble the final product until customers submit their specifications.

Holding “components” in inventory, and assembling them only after the customer has place an order, is a postponement strategy that is more appropriate for Dell than for Benetton. Nonetheless, the apparel maker does very well by manufacturing thousands of undyed garments, then dying them only when it receives an order for a particular color. This same strategy is used by any number of companies that offer personalized products, from coffee mugs to cardigans. “It’s just not very efficient to keep inventory for every single item customers might want,” says AMR’s Lapide.

For its part, Spanish clothing retailer Zara focuses on satisfying an important group of internal customers — the managers of its more than 1,000 stores worldwide. Zara makes it a priority from stores to designers and manufacturers. Since Zara designs and manufactures most clothing in its own facilities rather than turning to outsourcers, the company can deliver new fashions from sketch to store in 10 to 15 days. That quick turnaround provides better customer service while minimizing inventory, says Forrester’s Tohmay. (What’s being called PLM software — for “product life-cycle management” — is helping many companies manage products from conception to manufacturing to service and retirement.)

10. You Don’t Have a Unified Monitoring System for Your Supply Chain as a Whole

Even if you’ve overcome many of the obstacles discussed above, it’s critical to monitor your supply chain as a whole.

“Many of the monitoring tools in the market right now are all about alerts,” says Forrester’s Tohmay. These systems allow managers set up business rules, and if something falls outside those parameters, an E-mail alert will automatically be sent to the appropriate people.

“That’s helpful,” contends Tohmay, but “the user still needs to resolve a problem.” Strong monitoring tools should not only have the capacity to synthesize available data, they must also have the analytical intelligence to come up with recommendations, if not solutions. Companies have gotten pretty good at collecting vast amounts of information, she adds, but they’re still weak when it comes to making sense of the data, finding the significant drivers and separating the noise from what’s significant, says Tohmay.

That’s precisely what SeeCommerce — and a growing number of software companies including i2 Technologies, SAP, and Manugistics — help managers to do. Using Web-enabled portal technology, the systems help managers track key indicators, identify exceptions, and notify key stakeholders who might be able to resolve a problem.

Managers at DaimlerChrysler’s Mopar unit, for example, use Web-based portals to monitor the supply chain from their desktops. They can select a specific country, for instance, then drill down to see all the production plants in that country, the inventory levels at each plant, and the SKUs for that inventory.

“They don’t want to see everything that goes wrong,” however, says Mansbach, only those items that are “material and potentially costly.” Rather than look at 300,000 order lines and sift through oceans of information, for example, Mopar’s managers can use monitoring software to get a bird’s-eye view of the entire supply chain, so they can determine where their resources can be used more efficiently. (Find out more about how companies are “Working on the Chain” to make their supply chains as cost-effective, transparent, and responsive as possible.)

Have any “sure signs” of your own? Send them to [email protected].

Jennifer Caplan is an assistant editor at Additional reporting by assistant editor Craig Schneider.

4 Powerful Communication Strategies for Your Next Board Meeting