The business of making paints for new and existing cars and trucks could scarcely be more complicated. Take DuPont Performance Coatings, which produces more than 150,000 colors, deals with some 15,000 customers and 4 million order lines per year on the aftermarket end in Europe alone, and has to cope with issues of seasonality, volatile demand, precise color matching and environmental considerations.
So it seems unlikely that the key to the company's push for more accurate demand planning and forecasting would lie in simplification. But that's exactly what Digital Tempus Inc., a vendor of planning software, recommended to the customer.
DuPont Performance Coatings is one of the major business platforms within DuPont de Nemours & Co., the $31.5bn chemical and industrial products giant. The unit consists of four segments: an original equipment manufacturing division, delivering paint to car manufacturers; an aftermarket segment, for refinishing and repair; an industrial division, servicing the heavy truck, marine and rail industries; and another for powdered paint products.
Planning within the aftermarket division is especially tricky, given the difficulty of exactly matching colors with older cars. On top of that, DuPont was faced with the need to comply with a new European directive which mandated a switch to water-based paints. The company was also looking to introduce private-label and economy lines to address new market opportunities.
The older planning software wasn't sufficient to address current challenges, let alone future needs. DuPont had been working with a Manugistics application which it had acquired back in 1995. The original system was heavily supply-oriented, designed to ensure that the company had inventory on hand at all distribution locations. But the calculations weren't based on actual market and product trends. Nor was it sensitive to a business where innovation is a must. Half of the division's revenues today come from products that have been on the market for less than five years, according to Filip Buytaert, demand and business development manager for Europe, the Middle East and Africa.
"We were more driven by warehousing, making sure that operations had a SKU forecast," Buytaert says. "We were not at all linked to market planning."
Assessing the Operation
In 2007, the company was coming out of a major restructuring which involved implementation of an enterprise resource planning system from SAP AG. To align its demand-planning and forecasting process with real market conditions, DuPont turned to Gaithersburg, Md.-based Digital Tempus. The vendor's first step was to "shadow" DuPont's planning team through multiple cycles and determine how processes might be improved.
Digital Tempus's answer was to get company operations "back to basics" - in other words, simplify what had become a needlessly complex exercise. Paul Strzelec, chief executive officer of Digital Tempus, says the client was "lost in the weeds," with tons of detailed data yet little connection to the big picture. There was no alignment between sales and marketing and operations. So Digital Tempus recommended that DuPont take a fresh approach.
The idea, says Strzelec, was to "structure the information in a way that you can see the major patterns first and work your way into the details, rather than the other way around."
The shift involved about 10 months of intense work, Buytaert says. Digital Tempus helped DuPont to reconfigure its Manugistics application, which is now under the corporate umbrella of JDA Software Group.
In the process, the company devised some more precise forecasting techniques. In certain instances, it deployed complex algorithms in place of simple moving averages. In particular, DuPont drew on Lewandowski modeling, which accounts for changing market dynamics and helps to identify long-term trends in consumption.
The new system allowed for the aggregation of products by territory, both in terms of volume and value. For the first time, DuPont could consider outside influences on the forecast, such as the new European regime mandating water-based paints. It could also plot demand for the company's new line of economy products. Buytaert says it's critical for DuPont "to stay really close to what's happening in the market." Better intelligence, he adds, breeds customer loyalty.
The Bottom Line
After about 10 months of deploying the new system, DuPont had improved its forecast accuracy at the item level from 50 percent to 75 percent, according to Buytaert. That led to an inventory reduction of nearly $3.6m. Non-productive inventories were reduced by 15 percent, freeing up cash flow. From the customer's standpoint - which, after all, is the most important metric - DuPont realized a 10-percent increase in on-time, full deliveries.
The company also scored some "soft" benefits, including the creation of a business-intelligence environment which draws on multiple sources of data, such as field sales and customer information, to create more accurate forecasts. DuPont can consider such factors as the impact of currency values, as well as both variable and fixed costs. The result is a system that offers all internal users a unified, high-level view of demand, and tears down the walls between functions.
DuPont just completed an upgrade to the latest version of the JDA/Manugistics software. Buytaert describes the changeover as "a great success." The company also recently extended its contract with Digital Tempus by another two years.
Strzelec says the customer is aiming at continuous improvement of its planning and forecasting efforts. He sees opportunities in all four business segments within DuPont Performance Coatings - "to make them as capable as the refinishing business has become."
DuPont Performance Coatings