The Science of Forecasting

Inventory forecasting is tricky. IT bloggers speak of it as a "necessary evil" and lament the problems of balancing overstocks against out-ofstocks, while some industry experts warn against "phantom demand," "the forecasting tightrope," and "risk."

 

In the old days, forecasting was based on intuition and gut feelings. Today's solutions provide precise computer analysis of historical data. But what about the human element? A computer can't understand atypical information, such as slow store sales due to an unexpected blizzard.

 

Nevertheless, retailers -- especially those in segments sensitive to fashion trends, perishability and seasonality -- are turning to technology that add science to the traditional human art of forecasting.

 

ANCHOR BLUE SETS SAIL
Teenagers are notorious for wanting things now, so a retailer catering to the teen fashion market can't afford to run out of the latest glitter-accented tees or hip-hop jeans.

 

Anchor Blue, a specialty retailer of casual apparel and accessories for the teenage market, must deal with 30 million pieces of merchandise annually in 167 stores. Clearly this calls for a sophisticated and streamlined inventory forecasting system -- and plans for future growth include the addition of 230 stores in the next four years.

 

For help, Anchor Blue turned to Oracle. "As a starting point, we have invested in a suite of solutions from Oracle Retail with applications directed at merchandising, allocation, financial management, data warehouse and inventory management," reports Marty Kato, vice president of planning and allocation for Anchor Blue.

 

Anchor Blue has been using the system successfully since November 2005. The new system is designed to deliver supply-chain visibility, allowing merchants to better understand which sequined hoodies are moving faster than expected, and where the low stock positions are.

 

Michael Bush, Anchor Blue's CEO, is pleased with the new system: "Our new system meets our current business goals while allowing the flexibility to accommodate new demands as our company grows." Lyle Perrigo, vice president of information technology, shares the enthusiasm: "We now have the ability to react quickly to consumer demand. That way, we can keep a lot less inventory on hand, but we have more of the right inventory."

 

KEEPING IT FRESH
Lettuce wilts. Milk goes sour. And a plethora of products have "sell by" dates. In the grocery industry, it's critical to stock the right product at the right time. To do this, Redner's Markets implemented the DemandAnalytx (DAX) store forecasting and computer- assisted ordering (CAO) solution from Retalix, provider of integrated, enterprise-wide software solutions to the global food and retail industries.

 

Redner's operates more than 50 stores in Pennsylvania, New York and Maryland, where its Warehouse Markets include bakery and deli departments, and its Quick Shoppe locations offer convenience-store shopping.

 

"We are using Retalix to forecast sales and to order product based on a user-defined minimum inventory level," says John W. Sweigart, information technology director.

 

DAX uses sophisticated algorithms to interpret point-of-sale data, to forecast store-level demand and to optimize replenishment -- all the while, taking

 

into account the complex grocery environment. "Prior to this we used movement history reporting on certain items to provide an educated guess on replenish needs for our warehouse and our stores," Sweigart continues, "however, the order was not a computer-generated order."

 

Redner's hopes the new system will help increase profits in two ways -- by reducing spoilage and reducing inventory. Reduced spoilage will come from lowering inventory levels of perishable items, while reduced inventory will lead to increased turns and fewer dollars tied up in inventory. "We feel an increase in revenue will result from fewer out-of-stocks and increased sales," Sweigart says.

 

PREPARING THE DATA
The road to increased profits at Redner's is not without speed bumps. Asked about the challenges, Sweigart reports: "We had to develop category codes as our first priority. Secondly, we had to create files with the required information to feed the CAO program."

 

"We have our own warehouse facility," Sweigart continues. "All locations had to have a wireless network installed." Redner's is also dealing with the issue of persuading employees to have confidence in the system. "This has been a long-term process, but we think the tide is beginning to turn," says Sweigart."

 

Redner's began its pilot project with Retalix in early 2006 at two store locations, and now it has begun a roll out to the rest of the company. CG