By Mel Duvall  |  Posted 2005-11-08 Print this article Print

Midway through a $350 million plan to improve forecasts and lower inventories, Air Products found it needed more—an additional system to improve forecasts and lower inventories.

Plugging Holes

Air Products was founded in 1940 in Detroit on a simple premise. Instead of delivering gas products, such as oxygen and propane, in compressed-gas cylinders to manufacturers, founder Leonard Pool came up with another idea: build and install gas-generating facilities at the site of manufacturers that consumed large quantities of gas, like Weirton Steel.

Since then, Air Products has expanded into a variety of markets—providing oxygen, for example, for home health care, and specialty chemicals and coatings for large manufacturers such as Intel, which uses them to make semiconductors. The company's hydrogen is used by major oil companies like ExxonMobil in refining gasoline and by NASA to send astronauts into space, and its helium is found in half the world's magnetic resonance imaging (MRI) machines.

Unlike other manufacturers, Air Products cannot simply add another shift to increase output. Its investment in gas production plants requires years of planning to bring new facilities onstream. Its ability to produce hydrogen, for example, is limited by plant capacities, just as many oil and gas refiners are currently limited in how much gasoline they can produce due to outages caused by hurricanes Katrina and Rita.

Under normal circumstances, the company aims to balance supply against demand in such a way that its plants are operating at or near capacity, while getting top dollar for its products. If Air Products has excess production capabilities for hydrogen, for example, it may want to increase demand by lowering prices or bringing on new customers. When demand exceeds production, it may raise prices and allot more supply to higher-profit-margin customers.

And the pressure was on to pump up this balancing act. When chief executive John Jones assumed the company's helm in December 2000, he laid out a target of achieving a 13% operating return on net assets. That metric measures how well a company utilizes its investment in plants and materials.

When Jones set that goal, the company's operating return on net assets was 11.1%. A recession in the electronics industry and substantial increases in raw-material costs contributed to a drop to 10.2% in 2002 and 7.2% in 2003. It regained some ground the next two years. Now Jones has directed the company to come up with a plan to reach the 13% target by 2007

Putting a better tool in the hands of planners to match forecasted sales to production is a key plank in that strategy.

Reekie initially considered offerings from large-scale supply chain vendors such as i2 Technologies and Manugistics, but after spending $350 million on SAP, he knew spending several more million on i2 or Manugistics wouldn't fly. After evaluating several other options in the $1 million-or-less price range, Air Products selected software from Steelwedge of Pleasanton, Calif. A key feature of the software is that it offers point-and-click graphical abilities, allowing planners to dig down deeper into figures extracted from SAP to find answers to questions faster.

Steelwedge software now operates in the company's polymers, packaged gases (such as oxygen delivered in cylinders) and electronics specialty materials units. The goal is to roll it out to the rest of Air Products' operations by September 2006. Although it's too early to get hard numbers on the Steelwedge implementation, executives say the early results are encouraging.

Most of Air Products' sales and operations planning takes place at the company's 600-acre Allentown corporate campus in Pennsylvania's Lehigh Valley. The region is better known as the home to steel producers, and Air Products owes its location here to decades-old relationships with steelmakers that once dominated such centers as Allentown and Bethlehem. There are no heavy-industry facilities on site, however, and the headquarters has more of a university-like atmosphere.

Within each of the company's 17 business units, planning for sales and operations typically starts with an analysis of demand—about four to six months before production work actually begins. Information such as sales history, increases or decreases to product volume as a result of sales in the pipeline, significant new or lost customers, and changes in product mixes are pulled out of SAP and run through forecasting models.

After the demand analysis is completed, key managers from sales and marketing meet to build a consensus around a demand forecast. Sales managers, for example, might be aware of several big deals in the pipeline and want to adjust predicted volumes. Marketing may be aware of an advertising push or the fact that a product is being phased out, and will want to fine-tune demand projections.

Before installing Steelwedge, this type of on-the-fly adjustment of the demand cycle was difficult, Reekie says. You couldn't, for instance, increase predicted volumes to a major Singapore electronics customer of a chemical like nitrogen trifluoride, which is used to clean the chambers in which semiconductor chips are made, and see in a graphical dashboard the impact to the overall demand plan and prices.

When a plan is approved, a similar process takes place on the supply side. Based on predicted demand for individual products, operations planners attempt to map out production to meet demand. Supply issues—say, the loss of a raw-material supplier—are worked out, and inputs such as changes to prices of raw materials are factored in.

The next phase is to bring all sides together for a "partnership" meeting that also taps key representatives from manufacturing, purchasing and customer service. The purpose is to bridge the gaps between the demand forecast and manufacturing constraints. It may be, for example, that demand for nitrogen trifluoride in Asia far exceeds the company's abilities to supply.

This, again, is an area where the new planning software earns its stripes, Reekie says. By performing a variety of what-if scenarios with the software, planners can attempt to rectify production problems. What if the price of nitrogen trifluoride is increased by 10%? Will that lower demand closer to supply? What if Air Products contracts with another supplier in Taiwan to meet demand? What will be the impact on profits?

In the past, most of this analysis was done using spreadsheets, a slow slog. Planners are now making the adjustments within Steelwedge and acting on the results, Reekie says.

The final stage of the sales and operations planning process involves getting executive buy-in to the demand forecast and operations plan. Senior executives review the proposals, keeping the company's financial objectives in mind. Here again, the new software plays an important role, Reekie explains. If demand for a product is predicted to fall by 15% and as a result drive down profits, executives can use the software's point-and-click graphical capabilities to determine which customer is responsible for the shortfall.

Contributing Editor
Mel Duvall is a veteran business and technology journalist, having written for a variety of daily newspapers and magazines for 17 years. Most recently he was the Business Commerce Editor for Interactive Week, and previously served as a senior business writer for The Financial Post.


Submit a Comment

Loading Comments...
eWeek eWeek

Have the latest technology news and resources emailed to you everyday.