Good forecasting tools point the way to lower inventories
By Jean Thilmany, contributing editor -- Manufacturing Business Technology, 9/1/2007
For much of its existence, the Americas Beverage business unit of Cadbury Schweppes employed the common manufacturing industry practice of holding a certain amount of “safety stock” to avoid being unable to fill unexpected orders. Then the corporate office issued a directive that all business units would have to do a better job of managing cash flow.
To comply with that order, the Plano, Texas-based Americas Beverage unit would first have to do a better job of gauging demand for Dr. Pepper, Snapple, Squirt, and the other soft drinks it makes.
Two years later, the unit's safety stock levels have dropped significantly while its order fill rate is above 99 percent. Andrew Dennis, regional supply chain process director for Cadbury Schweppes, gives much of the credit for this change to the company's adoption of new inventory optimization software.
This software, developed by a company called ToolsGroup, allows Cadbury Schweppes to take a more granular approach to forecasting demand. For instance, it allows for forecasting at the stock keeping unit (SKU) level, as well for gauging how much of each SKU should be held at specific warehouses or distribution centers—areas that industry analysts say most manufacturers don't give enough attention.
“From a business perspective, it's critical to forecast not only at the SKU level, but for individual distribution centers as well,” says Julie Fraser, principal analyst with Cummaquid, Mass.-based Industry Directions. (See page 38 to learn about an Industry Directions study on forecasting.)
These needs have become more pronounced—and more difficult to meet—as supply chains have become more geographically dispersed, but Fraser says companies must step up to the challenge if they hope to maintain both high customer satisfaction levels and profit margins.
“It's always a trade-off,” she says, “between an ideal fill rate on the one hand and inventory levels that maximize profitability. You may want to service Wal-Mart, for example, at a 99 percent fill rate, while a smaller customer is satisfied with 95 percent.”
A layered approachThe magnitude of the problem caused Cadbury Schweppes to tackle it in stages, starting in its concentrated syrup division. This part of the business delivers syrup to restaurants, bars, and other commercial establishments that mix the syrup with soda water to make the fountain variety of various soft drinks.
Before implementing the ToolsGroup software, Cadbury Schweppes used spreadsheets to calculate safety stock levels, which generally were based on historical sales figures. Those forecasts were almost always inaccurate, however, because they didn't account for variables such as a sudden heat wave that might cause a spike in demand for soft drinks, or even Cadbury Schweppes' own sales promotions.
Dennis says that's changed with the new ToolsGroup software. The package holds the “Powered by SAP NetWeaver” designation, meaning it is tailored for easy integration with the SAP ERP suite.
At Cadbury Schweppes, the ToolsGroup application retrieves data on sales forecasts and customer orders from the APO (advanced planning and optimization) module of the SAP suite.
Running that information through its own algorithms, the ToolsGroup package designates safety stock levels for every SKU in the Cadbury Schweppes product line. It also determines the best locations for holding those items to satisfy the twin goals of meeting customer demand and minimizing operating costs.
These forecasts are updated weekly, taking into account the relevant variables that could not be plugged into the spreadsheets the company previously used for forecasting.
This new process has proved so successful that Cadbury Schweppes recently extended it to its finished goods division, which makes the bottled and canned versions of its soft drinks. That move netted a 12-percent inventory reduction—which translates to $7 million in savings—in the first three months. As important, says Dennis, “We are consistently meeting our customer service targets, with less manual intervention and extra effort than before.”
End the guessing gameA similar situation has unfolded at XRP, a South Gate, Calif.-based manufacturer of fuel-line parts for racing car and boat engines. XRP makes 5,000 types of parts, and fills an average of 300 orders per month from parts dealers around the world.
For years, XRP used a stand-alone database package to track customer orders and company finances. But XRP didn't have an inventory management system, so it could only guess about proper safety stock levels. Periodically, supply chain managers would manually comb through past sales data. By documenting which part numbers were the biggest sellers, they at least got an idea about which items it made sense to replenish. That information was then used to plan production.
When XRP decided to migrate to an ERP system, it chose a package from SYSPRO, a vendor that caters to midsize manufacturers. And when SYSPRO introduced a forecasting module, XRP was an early adopter.
The SYSPRO forecasting module allows companies to forecast at both the item and warehouse levels. Jarrod Barker, XRP's IT manager, says the ability to do item-level forecasting is exactly what XRP needed.
“With our former software package, our inventory [of certain items] often would go negative,” Barker says. “Now we use the forecasting software to tell us exactly what we need to have on hand.”
SYSPRO's forecasting package also examines past sales by part number, but it is better equipped to use that information as the basis for an accurate determination of safety stock levels. That's because the system can plug in intelligence—such as a particular customer's short-term buying plans—or other variables that a manual forecasting process can't consider.
Barker says XRP hasn't been using the forecasting module long enough for any hard numbers related to its impact on the business. Still, he says, “I know it works. It significantly improved our inventory management.”
| Author Information |
| Jean Thilmany can be reached at thilmanyj@yahoo.com |
|



















More results on MBT Research Library