How Accurate Is Your Forecasting?

Leave the guesswork to fortunetellers  learn how to properly use forecasting to improve your distribution processes.

Forecast accuracy has always been a problem in the electronics manufacturing supply chain, but it has become even more so since the outsourcing trend became popular. In fact, many industry observers have blamed forecast inaccuracy, especially in the communications electronics industry, for the steepest decline in electronic component sales in history, which occurred during the recent economic downturn.

As these last few quarters have demonstrated, when forecasts are inaccurate, all the fancy supply chain collaborative tools and processes in the world aren't sophisticated enough to prevent a meltdown. In fact, industry supply chain practitioners may be desperate enough to consider crystal ball gazing a more-effective forecasting tool.

But there may be some help available. One explanation for the high-tech supply chain meltdown is the confusion over the fact that there are several types of forecasts that are generated under very different circumstances and for very different purposes. These various forecasts often are misapplied, causing all sorts of problems with inventory misalignment and inefficiency.

Forecasting 101

One type of forecasting is generated as part of the production cycle. It is usually connected to an enterprise resource planning (ERP) system and acts as a supply chain integration tool, tying the OEM to its distributors and electronic manufacturing services providers. This automatic, electronic signal keeps inventory moving from suppliers to the production line, allowing for smooth manufacturing cycles and ensuring that the original equipment manufacturer (OEM) always has enough products to sell to customers. However, in order to be truly effective, this type of forecast must be tied to real-time demand information. And, for most companies operating in early 2002, this capability is not yet in place. Consequently, this forecast most often operates in somewhat of a vacuum.

Another type of forecast comes from the marketing department. This forecast, not surprisingly, is based mainly on marketing activities. Promotions, competitive analyses, consumer surveys and the like are used to generate marketing forecasts that drive production. In the past, (before outsourcing), when the marketing department was part of the same organization as manufacturing, there were informal controls that came into play to balance the sometimes overzealous forecasts that came from the marketing department. Now, with production being done by contract manufacturers, those informal controls are gone and the marketing forecasts are passed unchallenged to manufacturing. During the communications industry boom period in 2000, cell phone OEMs were issuing announcements about gains in marketshare, that, when aggregated, were inflated to the point of absurdity. 

In another example, Dr. Hau Lee, director of the Stanford Global Supply Chain Forum, tells a story about Volvo that perfectly illustrates what happens when communication breaks down between marketing and production. In the mid-1990's, Volvo was faced with an excessive inventory of green cars in the middle of the year. Eager to get rid of this inventory, the sales and marketing group started aggressively offering special deals, discounts and rebates on green cars to their distributors. According to Dr. Lee, the good news was that the scheme worked, and green cars started to sell. But the bad news was that the supply chain planning group, not knowing about the promotions, erroneously thought that demand for green cars had risen and decided to produce even more green cars for the coming model year. The end result? Volvo had an overstocked inventory of green cars at the end of the year.

Wanted: Supply Chain Manager

The outsourced electronics manufacturing business model requires some tweaking to successfully weather the next storm. In the mid-1990s, supply chain management was driven largely by the escalation in the use of electronic data interchange (EDI). This tool allowed certain types of information to be exchanged electronically, improving the speed and accuracy of the data. The chain was seen as having three basic flows:

·        Information  sharing forecasts orders and change orders

·        Material  moving to just in time

·        Money  using electronic funds transfer to streamline transaction-heavy processes.

These innovations served the electronics industry well, and early adapters were able to gain competitive advantage by significantly reducing the direct and indirect costs associated with material management and ownership. Many used the SCOR model, which focuses on plan, source, make and deliver to improve processes and align with industry standards.

But by the close of the millennium, most of the advantages had been wrung out of the early versions of the extended enterprise. However, the cost of electronic products continued to be driven by materials  upward of 85 percent of the total cost. Consequently, visionaries started to look at new ways to impact this cost. This investigation led to the production environment as a whole and to the extended enterprise. Areas such as design showed promise and, while the concept of design for manufacturing was not new, renewed attention was focused on designing products with components that were optimized in terms of cost and availability during the volume production cycle, to reduce total costs, mitigate liability and avoid product obsolescence. Further, as world markets emerged, it became critical to optimize material and distribution around the globe.

Now there is an increasing divide between the theory and practice of supply chain management, due in some measure to a lack of trust among supply chain participants. Information critical to the efficiency of the entire supply chain resides in isolated pockets, hoarded and protected to the detriment of all players. Companies must begin to look at the supply chain, not as a way merely to cut costs, but as a source of continuous value creation.

The tools and processes are available to link the various types of forecasts and supply chain players together. The will to collaborate for greater efficiency is emerging, but what is needed is one supply chain entity that is in a position to see the big picture and aggregate all the information flows. So now the problem becomes deciding which entity is in the best position to move the levers on the dashboard to most efficiently regulate and manage material and services for the extended enterprise.

Who's Driving?

Some would argue that EMSI providers are best positioned to manage inventory for their OEM customers. Yet, according to figures recently collected by iSuppli, an El Segundo, California-based supply chain analyst firm, the track record of EMS during this recent downturn has not been very good. During the fourth quarter of 2001, EMS providers held nearly 40 percent of the excess semiconductor inventory, while distributors held only 5 percent. The reason EMSI providers find themselves in this predicament is that they are under contract to follow their OEM customers' market-driven forecast. Even when the largest EMSI providers aggregate their customer bases, there is no way they have the ability to add accuracy to the OEM forecast. They simply don't have a wide enough spectrum of customer and supplier relationships  nor do they have the information technology tools in place  to have good forecast accuracy.

Industry consultants and analysts may have good theories and strategies, but they don't have the component usage data and linecard relationships or the broad customer base to have granular data about demand. Further, they don't have the supply chain information technology infrastructure in place to transfer real-time data among players. And each OEM is enmeshed in its own corner of the market and can't see the bigger picture about broader industry trends.

During the communications industry meltdown, many in the distribution industry saw that the OEM forecasts were inflated to the point of absurdity. However, when warnings were pushed to the customers, they went unheeded. At that point, it seemed clear that the supply chain model in place needed some tweaking.

The value that distributors bring to the supply chain arises from distribution's core competencies: in IT, logistics and warehousing, customer and supplier relationships, and the aggregated demand information coming from its suppliers and customers. Inventory assets and the ability to leverage them profitably are an integral part of distribution's core business model, which is, in a way, unique in the supply chain. The semiconductor supply chain is immensely complex, with long product lead times and short product life cycles. With time-to-market constraints tightening, the consequences of excess inventory will become more and more dire. There is no room for failure. Companies must populate their supply chains with the relationships and business models that make the most strategic sense and provide the most value to the extended enterprise. Distribution has proven its value and stands ready to serve its constituents with new tools and processes to improve supply chain efficiency. Perhaps it's time to abandon crystal balls altogether and begin forecasting with a more enlightened approach.