Inventory Management: Grab the Bull by Both Horns

For direct merchandisers, the Internet brings opportunities for additional sales and another touchpoint to build relationships with customers. As Internet-based sales have risen to 50 percent of total sales or more for many companies, they’ve also introduced new challenges to demand forecasting and inventory management. Many traditional catalog merchants feel less in control of their inventory planning than in the past.

In many cases, demand forecasting and inventory management processes haven’t kept pace with ever-changing Internet marketing. Lower productivity is often the result, as more time is spent reconfirming marketing plans and checking and rechecking forecasts and purchases. More fundamentally, lost sales, back orders and overstocks increase because inaccurate or late forecasts result in purchasing the wrong inventory and late deliveries.

For many direct merchandisers, success lies in their ability to adapt demand forecasting and inventory management processes that handle both catalog and Internet business with equal confidence. The two most common approaches are offer-based forecasting and statistical weekly forecasting.

Offer-Based Forecasting
On the surface, it seems that forecasting for catalog and the Internet are incompatible. Traditional catalog merchandisers operate from predictable sales patterns generated through catalog-based forecasting methods (offer-based forecasting).

For example, the impact of releasing millions of catalogs on a single mail date creates a reliable sales trend. The use of space and position in the catalog allows the demand planner to adjust forecasts accordingly. The print production cycle also allows for a longer planning horizon. As a result, merchandisers drive their own trends by drop dates, layouts and promotions they determine.

Even outside factors, such as holidays and seasonal sales patterns, are predictable to catalogers. While there are still significant challenges for accurate item forecasting, the inherent catalog production process and planning cycle allows inventory planners to anticipate demand and manage inventory to meet that demand. It also instills a greater sense of control.

Statistical Weekly Forecasting
Internet sales planning has two fundamental differences from catalog. One, the annual sales trends are more consistent over time, as there aren’t single mailing dates where millions of catalogs are released. Two, the planning time horizon is much shorter because product changes can be made in a matter of hours. This requires an approach to demand forecasting that’s better suited to evaluating the factors that drive sales on the Web site using statistical weekly forecasting.

Statistical weekly forecasting incorporates a blend of high automation with targeted controls. Actual sales information is evaluated within statistical forecasting methods, such as exponential smoothing or dynamic regression. When a business has enough sales data, extremely reliable forecasts can be derived from historic trends.

The base forecast is computer-generated, freeing up staff time to adjust forecasts to recognize ever-changing Internet marketing events, such as e-mails, homepage features and temporary promotions. With statistical weekly forecasting in place, users spend less time forecasting and more time making purchasing decisions.

Changing Your Approach
There’s no single right answer to an individual company’s forecasting technique. For most catalog-driven businesses, the best strategy is to apply offer-based forecasting to time periods, either calendar quarters or months. Because catalog mailings are still the dominant factor in sales trends, this technique most easily and accurately forecasts the impact of catalog plans on Internet sales time periods. Conversely, if a company has extremely stable catalog mailings with few product changes from year to year, the statistical forecasting technique may be best.

Since forecasting is only half the game, it’s equally important to have an inventory master-scheduling tool that incorporates future demand and on-hand and on-order inventory by week, along with estimated returns by week to calculate inventory level by SKU by week into the future.

Successful multichannel merchandisers also understand that for either approach to work, it must be coupled with good top-down controls and targeted exception reports to identify less reliable forecasts so they’re able to respond to situations and resolve any problems more quickly.

Ray Goodman is senior vice president of Direct Tech, a provider of marketing and merchandising systems for multichannel marketers. You can reach him at (402) 895-2100 or rgoodman@direct-tech.com.