Web browser brings product demand into focus

Eastman Chemical now disseminates real-time monthly sales forecasts over its corporate intranet, giving users worldwide a clearer picture of product demand.

By James Aaron Cooke -- Logistics Management, 5/1/1999

The corporate intranet has become a critical tool in Eastman Chemical's efforts to improve forecast accuracy.

THINK FORECASTING IS EASY? Just ask a weatherman. Businesses, too, find predicting the future to be a challenge. Replenishment forecasting, in particular, represents a dicey proposition. It gets even dicier when it involves predicting demand worldwide in dozens of overseas markets.

Two years ago, Eastman Chemical Corp. decided to meet the forecasting challenge head on. The Kingsport, Tenn.-based chemical giant put together a cross-functional team of managers to examine demand across its supply chain. The initial team included representatives from the sales, planning, purchasing, logistics, and business organization functions. "We were looking at a way to translate demand at the customer level into [information we could use to supply] product at the right place at the right time in the right quantity," recalls John A. Hewson, Eastman's manager of global forecasting.

With help from a software program, Eastman has succeeded in coming up with swift and more precise demand projections. It now disseminates those real-time monthly forecasts over its corporate intranet, the company's in-house computer network.

The wider distribution of information has enabled the chemical company to better synchronize production and inventory with customer demand. "We're generating monthly forecasts of customer demand in a defined, controlled way that provides us with the basic [mechanism] to plan manufacturing, procurement, warehousing, shipments, and logistics--the whole supply chain," Hewson says.

Change in Focus

Eastman Chemical Corp., founded in 1920 as part of Eastman Kodak Co., was spun off in 1994. Last year, the company recorded $4.48 billion in sales, about 40 percent of which came from overseas markets. The chemical company employs some 16,000 people in more than 30 countries.

Eastman produces more than 400 chemicals, fibers, and plastics. Its chemicals are used to manufacture a host of products, including paints, inks, automobiles, tapes, and eyeglasses. One of its best-known products is polyethylene terephthalate (PET) polyester, which is used to make soft-drink bottles. "We're the world's biggest producer of PET plastic," says Hewson. PET plastic, he adds, is subject to seasonal demand because more soft drinks are consumed in the hot months. By predicting that demand effectively, he adds, Eastman can optimize production capacity and inventory.

In order to coordinate its global supply chain more efficiently, Eastman decided it needed better information on customer demand. "We recognized that in order to improve planning across the enterprise, the first thing you need to do is quantify demand at the customer level," says Hewson.

In the past, Hewson reports, forecasts were created on an ad hoc basis by different people at different times. "We had islands of analysis," he says. "People were using different numbers."

That method of forecasting was fine when companies in the chemicals business ran a "push" type of operation. As Hewson explains it, chemical companies would run their plants at full capacity and then find markets for their wares. "Our goal was to run the plants at the lowest possible cost," he explains.

But that production method doesn't work now that companies have switched to a different philosophy--basically, one of global replenishment that's based on customer "pull." Instead of the chemical producers "pushing" their products into markets, Eastman's customers determine production and inventory flows by "pulling" the products they want. "The transition is from manufacturing push to demand pull," says Hewson. "The idea is that you manufacture just enough to meet customer demand."

Complicating the forecasting task is Eastman's far-flung international production network. The company has U.S. plants in Arkansas, South Carolina, New York, Tennessee, Texas, and West Virginia. It also has factories in Argentina, Canada, England, Mexico, Hong Kong, Malaysia, the Netherlands, Singapore, Spain, and Wales.

As the first step toward improving control over its global forecast, Eastman formed a small group inside the company to develop the requisite data to gauge future production. Hewson was chosen as the company's forecasting "champion." "As a global company with ambitions to become the world's [top] chemical company," Hewson reports, "we had to come up with a supply chain management tool for forecasting."

Forecasts on Demand

Hewson and his team first turned their attention to the company's forecasting method. Although Eastman had installed SAP's R/2 enterprise resource planning (ERP) system, it was doing its forecasting on an older system.

It was clear that additional forecasting support software was necessary for the job and that the new software would have to be linked to the ERP system. After a review, Eastman chose the "Demand Planning" (DP) application from Atlanta-based Logility. The DP software runs on a Windows NT platform with an interface to the SAP system. That package also contains a module called "Demand Chain Voyager," which allows the company to exchange forecasting data over the corporate intranet. Voyager works with such common Web browsers as Netscape and Internet Explorer.

The Logility software, which Eastman installed earlier this year, enables the company to shift responsibility for demand planning to the employees who are closest to its customers--the worldwide sales force. "The sales force is the primary group responsible for demand," says Hewson, "but there are checks built in to counter an optimistic salesman. We've made forecast accuracy a significant measure in the enterprise."

The Voyager module is loaded onto the laptop computers used by Eastman's 250 salespeople, who are scattered throughout the world. Because the software resides on the laptop computers, the sales reps can generate numbers without being hooked to the company's network. "It's a dial-in connect where you can download data and work offline on the laptop," Hewson reports. "Then when you finish, you upload the data."

Every month, Eastman's sales reps worldwide record their individual sales forecasts on the company's intranet. "They will dial in to get onto the intranet and download the latest forecast generated based on historical statistics," says Hewson. "They view it on their laptops using the Voyager software and we have them type in any exceptions."

All of those sales forecasts then are sent to the company's head office in Kingsport, Tenn., where they are reviewed. (Hewson notes that access rights to the forecast are restricted to the reps, their sales managers, and a few other select managers.) Product managers then double check the forecasts submitted by the reps.

Hewson says his company views the forecast information as the primary driver behind its supply chain replenishment strategy. "We're in the game of huge plants and huge capital investment and [we have] to run those plants efficiently to keep costs down," says Hewson. "Our forecast is critical to our knowing further into the future what the demand will be."

But Eastman's global supply chain plan is not yet fully in place. In fact, the installation of software for real-time monthly forecasts is just the beginning of the company's plan to optimize its supply chain. "This is an ongoing project," says Hewson. "We're only through the initial phase to improve the quality of the forecast."

The demand forecasts generated by the software now are shared with such functions as logistics and purchasing to corroborate the projections. "We're starting to look at the data and do some reality checks," he says. "We finally have one data source for the forecast that everybody has access to. Everybody looks at the data to see if it looks wrong or right. And it's engendering dialogue."

The Right Direction

The software-generated forecasts already have proved more accurate than previous ones. Hewson says that in the past, demand forecasts had a 20-percent error rate compared to actual sales revenue. The first software-generated forecast had an error rate that was below 15 percent. "The very first time we used the new system's number, boom, we eliminated 5 percent error," says Hewson. "We're still not satisfied, but it's headed in the right direction."

In addition, Hewson says Eastman has begun a program where it's using the new forecasts to adjust warehouse inventory levels. "We've targeted a specific warehouse in which we'll determine inventory levels based on the new forecast," he says. "We're confident we can reduce inventory based on this approach."

In the future, the company plans to use its monthly sales forecast to optimize production and distribution. It will use those data to work with another piece of best-of-breed software, an optimization tool for analyzing the supply chain's makeup. "We'll take the Voyager forecast and feed it into a supply chain optimization tool that will allow us to determine the fastest routing for a shipment or the safest or lowest cost," says Hewson.

Precise forecasts will allow Eastman to develop a more efficient global supply chain. "By having more accurate forecast data," says Hewson, "we can [realize] savings in delivered cost, raw-materials cost, manufacturing cost, and inventory and warehousing costs."

Eastman's products provide the raw materials for a host of industrial and consumer goods.