Precision now, fulfillment later
Intertape Polymer, Godiva Chocolatier use forecasting suites to position themselves to reduce costs, better meet customer demand
By Jim Fulcher, contributing editor -- Manufacturing Business Technology, 12/1/2003 7:00:00 AM
The old forecasting process at Intertape Polymer Group was inefficient, requiring one forecaster to cut and paste data from a database to an Excel spreadsheet. Worse, however, was that the resulting single-forecast model couldn't account for seasonality and forecast at both sub-class and product levels.
"We couldn't fill our regional distribution centers accurately, so while we had excess inventory and could fill customers' orders, we couldn't necessarily ship them from the closest distribution center," says Dan O'Brien, corporate forecasting and inventory control manager at the Sarasota, Fla.-based manufacturer of Intertape brand sealing tape, masking tape, duct tape, and other specialty tapes; shrink film; and stretch wrap.
"As a result, we sometimes needed to fill an order and ship it from the east coast to the west coast," O'Brien adds. "We didn't have the right product at the right time in the right location, so costs associated with inventory levels—and shipping—were more than they should have been."
Today, use of the ForecastX Wizard and ForecastX Engine from John Galt Solutions offers Intertape the means to forecast more precisely, and subsequently fulfill that demand more efficiently. "Our forecast accuracy has improved by 13 percent, so we're now able to create a more accurate demand plan with production," O'Brien says. As a result, he adds, Intertape has been able to improve fill rates, cut both slow-moving and active inventory, reduce the number of production expedites, and pare down logistics costs.
Of course, the use of forecasting software is nothing new. For years, manufacturers have been turning to best-of-breed forecasting software vendors to augment or replace the forecasting modules found in ERP systems. What is new is the evolution of forecasting suites that can handle pricing and promotions, can adjust demand plans to events, or support sales & operations planning (S&OP) procedures. "The trend has been to not just offer a base forecast engine, but a suite of products that manufacturers may leverage to generate increasingly accurate forecasts," says Robert Carbone, president of forecasting vendor Futurion.
Fresh assumptions
Carbone says Futurion has added functionality for S&OP and other areas to complement its forecasting engine. In addition, he says, the company has developed an "assumption-based" forecasting capability aimed initially at pharmaceutical manufacturers. Carbone believes it will be a milestone in forecasting because at its core, the functionality combines forecasting with intelligence about a manufacturer's competitors.
The process begins by looking at a company's past performance as well as that of its competitors. Then, users factor in information about what they plan to do differently and what their competitors will do. The final step is to evaluate the impact of these assumptions.
"In the pharmaceutical market, companies such as Novartis [Basel, Switzerland] and Boehringer Ingelheim [Ingelheim am Rhein, Germany] have begun assumption-based forecasting because they have a critical need to know what other companies' work on competing drugs will mean to their business," Carbone says.
Assumptions, says Carbone, can be of different types. While some are market-driven assumptions, others are tied to changes in a product's life cycle. For instance, he says, a new scientific "indication" for a drug that opens it up for new uses will likely impact the market for that drug, and hence the forecast.
"Pharmaceutical manufacturers often are the software solution early adopters and trendsetters because they have so much at stake," Carbone says. "I believe the next industry to follow suit will be over-the-counter health-care product manufacturers, perhaps followed by those in the advanced electronics market."
Across other manufacturing sectors, the economic struggles of the last few years have focused more attention on demand management, says Larry Lapide, a VP with Boston-based analyst firm AMR Research. Manufacturers went from facing supply problems during the boom years, to demand problems. "Today, there's too much supply and not enough demand," Lapide says. "Having to chase demand is a different paradigm than manufacturers were following three or four years ago."
To get a better grip on the true demand picture, says Lapide, interest has grown in linking event management with forecasting processes, as well as on S&OP. "We're seeing a resurgence in S&OP because it centers on accurately matching supply and demand," Lapide says. "The bottom line is that manufacturers know they need to get their plan working well so they're in a good position to act accordingly to events or exceptions. Forecasting and demand management are central to these efforts."
Greater precision
Manufacturers can't respond well to events or exceptions if they haven't forecast well because they simply won't have the necessary materials on hand to act, says Charles Smart, president of forecasting vendor Smart Software. At the same time, inaccurate or inflated forecasts also can lead to excess inventory.
"Investigating just how much safety stock is needed may lead to discovering previously hidden and surprising costs," Smart says. "If manufacturers aren't using an automated process to project safety-stock levels, then they could be spending a lot more on inventory than they need to. Complicating the challenge is the fact that safety stock can't be a set level because it varies based on factors ranging from seasonality to component lead times."
In the case of Intertape, the need to project safety-stock levels wasn't as important as the need to consider seasonality. For example, demand is higher for HVAC tape—used for heating, ventilating, and air conditioning—in cooler months when furnace repairs are made. On the other hand, demand for high-performance automotive tapes is higher in warmer months when conditions are more conducive to driving.
"Seasonality is a huge factor in forecasting product sub-classes, but with our manual process, all of our SKUs and many types of tapes all followed the same forecast model—which used one year of historical data," Intertape's O'Brien says. "Today, with the John Galt Solutions software, I can look at two to three years of monthly sales for a particular product sub-class—including information about seasonality—to create a general forecast for square meters of a type of tape. If the forecast seems on track to sales and operations, and there aren't any capacity issues, we then forecast SKU by rolls, and then begin to forecast by sales region."
More dynamic
Even with broader capabilities in today's forecasting suites, it's important to remember that forecasting is a dynamic endeavor, says Anne Omrod, president and CEO of John Galt Solutions. To achieve ongoing improvements, a company must manage its forecasts.
"Companies need collaborative input, and they should review forecasts to make sure they seem reasonable," Omrod says. "However, it's also critical to analyze forecasts after the fact and compare them with actual shipments to determine how accurate the forecasts were, and look for unplanned trends. That data may then be used to make future forecasts even more accurate."
Karen Jensen, forecasting manager at Godiva Chocolatier, Reading, Pa., understands the importance of analyzing forecasts. The company uses a forecasting solution from Smart Software that considers three years of historical data along with input from the marketing department. Jensen regularly reviews forecasts for issues to discuss at forecasting meetings.
"I use the Smart solution to generate numbers each month for 10 different customer types, looking ahead 18 months. These numbers are then reviewed by our marketing department, which ultimately is responsible for the forecasts," Jensen says. "They see the statistically generated numbers, and may accept them or make adjustments. The forecast then is used to drive production."
The company holds monthly forecasting meetings to review the forecasts. Jensen also analyzes the forecasts to look for significant differences between the solution-generated numbers and the marketing group's numbers. "If I find anything, I question the appropriate marketing person," she says.
Jensen also uses a Godiva-developed Web-based tool to look for significant differences between the actual shipments and marketing numbers—and then determine why there was a difference. A number of factors may come into play. "For example, expensive chocolate is the first thing people cross off their shopping lists when the economy is slow," says Jensen. "The weather also may cause discrepancies. Warm weather isn't conducive to chocolate buying, but if there's too much snow or rain, people stay in and don't go to the mall."
This brand of more precise, dynamic forecasting may begin to catch on with more manufacturers. With the help of broader application suites from the forecasting vendors, greater precision and responsiveness to market conditions appears not only possible, but profitable.
Featured Company
Most Recent Resources
- FICO™ Xpress Optimization Suite Schedules Big Profits For Clients
- Strategic Pricing: Three Steps to Higher Profit Margins
- Driving Innovation Through Lean Product Development Practices
- Demand Planning Maturity Model Strategies for Demand-Driven...
- Simulation-Driven Product Development:Will Form Finally Follow...
