Demand Planning's Suite Song
Forecasting and planning software helps companies match inventories to customers' future demands.
With the right crystal ball, you can manage inventory to perfection. You know what your customers will need next week, next month, and next year. You always have the right amount of every product ready to ship.
Fog-free fortune-telling paraphernalia being in short supply, though, companies turn instead to demand forecasting software. More than 2,000 of them have sought help from Demand Management Inc., a St. Louis-based firm with an expanding suite of forecasting and planning solutions.
Demand Management introduced the first version of its forecasting engine, Demand Solutions Forecast Management (DS FM), in 1985. DS FM looks at sales history and other factors, such as past and future pricing, to predict future demand for products.
Leanin' Tree, a vendor of western-themed greeting cards and gifts, purchased DS FM 10 years ago as part of an effort to reduce and better manage its inventory. Leanin' Tree sells its products to retailers and, through a catalog and web site, directly to consumers. It installed DS FM as a front end to another solution it implemented at the same time, the MACPAC manufacturing resource planning (MRP) system from Andersen Consulting. (It recently replaced MACPAC with software from SAP.) Together, the two systems helped Leanin' Tree consolidate three warehouse operations into one, says Lloyd Peirce, materials manager at the Boulder, Colo.-based firm.
Card by Card
The old forecasting system at Leanin' Tree lumped together all items in a product category. For example, it generated one demand figure for all greeting cards, not distinguishing between birthday, Christmas, or Mother's Day cards or between cards with different designs. "With Demand Solutions, you forecast at the actual SKU level," Peirce says. DS FM can also predict high and low demands for different seasons, rather than simply spreading forecasts evenly over a certain number of weeks.
To create a forecast, planners at Leanin' Tree download into DS FM sales figures for the coming month for each SKU from the company's sales system. Looking at this and at data for each SKU from the past 36 months, the system predicts demand for that item in the coming weeks.
Software filters help catch the 10 percent or so of situations that a computer can't get right on its own. For example, if demand for an item drops steadily over the years, the system will eventually project zero sales. "But it's still a good product," Peirce says. "You will always have some sales. If we see that trend happening, we put a different algorithm to it."
Along with cutting back from three warehouses to one, Leanin' Tree has improved customer service since implementing DS FM. Leanin' Tree's goal is to ship every order within 72 hours of its due date. In the old days, the company hit that mark only 60 to 70 percent of the time.
"Before, we didn't know what we were producing. The time to make our product was basically a month," Peirce says. "Now, we can turn cards around in two to three days, which has caused the customer service rate to go up. We're at 99 percent."
While Forecast Management remains the flagship product, developers at Demand Management realized early on that for many customers, basic forecasting is not enough. "As we got more into forecasting, we found out, as everybody knows, that the forecast is wrong," says Mike Campbell, Demand Management's president and chief executive officer. "We needed a tool to manage the variability of the demand."
That solution became the company's second product, Demand Solutions Requirements Planning (DS RP). Using current data imported from other enterprise systems, and from customers and suppliers, this module refines the forecast. It looks at factors such as inventory at the plant or distribution center and at the customer's site, what goods or materials suppliers are shipping, and what orders the customer has placed. The object is to tell planners what to do in the near future to keep inventory at the right levels.
"Demand varies from what the forecast says it should be," Campbell explains. "Taking into account customer orders, inventory, and other factors, we can determine what the requirements are."
Beyond providing a snapshot of inventory and demand, "the software also has numerous filters that let it find anomalies," Campbell says. For example, it might point out that the product a customer has ordered is out of stock, or it might indicate that certain incoming goods will arrive late. Taking all the variables into account, the system produces a plan of action.
Companies that implement Demand Solutions modules often use them to replace time-consuming, labor-intensive procedures.
"Many companies still do forecasting and replenishment planning using spreadsheets. That's okay, but it's a huge expense in terms of the ability to maintain that information," Campbell says.
That was certainly the case at Fender Musical Instruments Corp., where a proprietary system couldn't integrate the forecasting and planning process for its distribution centers in Corona, Calif., and Roermond, the Netherlands. "We had two, complete, independent processes that we tried to match up using spreadsheets," says Michael Gilreath, director of management support at Fender in Scottsdale, Ariz.
Beyond that particular problem, forecasting and planning with the old system involved "an enormous amount of data entry and looking up information for each SKU" requiring "days of aggravation and tedious work," he says.
Along with greater efficiency, Fender wanted a more powerful, versatile tool to help predict demand for its guitars, basses, and other musical gear. The old system used a single algorithm. "People were comfortable with it, but it was very limited in what it could do," Gilreath says.
For instance, the system predicted annual sales figures, then broke them down by month based on numbers the marketing manager provided. But as information on actual customer demand rolled in, it couldn't refine those monthly figures. "We couldn't make an adjustment other than off line, manually, or making a mental note of it," he says.
Fender implemented DS FM and DS RP in January. Now, once DS FM creates a forecast, the data flows into the planning module, which also looks at current inventory, purchase orders, and factory orders. Having considered all that information, it creates a production schedule.
With the new software, Fender can forecast sales separately for the two DCs but consider their inventory together in the planning process, Gilreath says. Because the work is automated, the employee responsible for forecasting in the company's European office doesn't have to tackle planning as well.
"Now that we're able to relieve him of the burden of planning, he can concentrate on the forecast, and we can integrate the planning into what we do in Scottsdale," Gilreath says.
Learning the ABCs
Fender is also starting to use the system to do more sophisticated planning exercises, such as "ABC analysis"—dividing its products into three tiers, according to how well they sell, and putting special emphasis on those products that contribute most to the bottom line.
"The ABC analysis in particular is one that we're looking at to change the way we manage inventory," Gilreath says.
Other modules in the Demand Solutions suite include:
- Rough Cut, a tool for comparing production or purchasing capabilities with the demand forecast.
- Sales & Operations Planning, which consolidates histories and forecasts so that senior managers can examine them together.
- Stores, a store-level replenishment tool.
- Feedback, which allows users and their customers to collaborate on forecasts.
- View, a reporting and analysis module.
- Electronic Demand Solutions Interface, an electronic communications tool.
The Feedback module is designed to answer the desire for collaboration in supply chain management. Campbell describes it as "the forecasting system in miniature." Loaded onto a laptop that a sales rep takes on a call, or presented to a customer over the web, it allows the customer to respond to the supplier's sales projections.
"You can put this in the hands of a salesman or customer and say, "Here's what we think the demand will be. Would you give us your input and tell us what you think the demand is going to be, so we can get as close as we can to what the numbers need to be?'" Campbell explains.
While it might not exude the same glamorous aura as a crystal ball, Demand Management's suite is working magic for companies that use it. At Leanin' Tree, Peirce says the biggest benefit is always having the inventory to fill customers' needs.
"Customers know when they place an order, they're going to receive it," he says.