Who's That Behind FosterGrant's Demand Planning?

Long lead times exacerbate the already difficult task of forecasting seasonal, fashion items like sunglasses. At AAi.FosterGrant, which sources much of its product overseas, the time between placing an order and having it available in the store was close to five months when Darrin Weigle, vice president of supply-chain management, joined the company two years ago. "We didn't chart supply-chain metrics then, but we knew that we rarely seemed to have had the right product in the right place at the right time," Weigle says. "As a result, we would often have to expedite inbound product to fight fires we were experiencing at that moment."

In addition to rising costs for expedited air service, inventory transfers also were eroding the company's profitability. Products arrived from overseas vendors specifically packaged and prepared for sale at different retail outlets. The very same product would carry a different SKU and different packaging depending on the destination store. When the mix wasn't right, the company had to use local contract services to re-package Customer A-destined products for presentation in Customer B's stores, or vice versa, a process "that can really can eat into your margin, when it typically costs between 14 and 20 cents per unit," Weigle says.

Part of the problem stemmed from the fact that the demand planning process the company then used failed to differentiate between orders that were placed in response to true demand, those that were in anticipation of a particular promotion, and those that were placed simply to use open-to-buy dollars.

"Basically, anything in the demand history, as far as customer orders and shipments were concerned, was considered viable demand that could be used to forecast on and project future sales," says Weigle. "That's obviously nonsense, if half of it is non-repeatable business. But if you can imagine a demand forecast based on everything that was shipped out of here over the past three years, you can envision how large our forecasts were. And there was really nobody even challenging whether the forecasts were anywhere near the realm of reality." Overflow warehouse facilities were needed to house incoming shipments, and handling and transport expenses were getting out of hand, he adds.



	"We had to put in place the proper foundational elements for a successful, best-in-class supply-chain strategy." - Darrin Weigle of AAi.FosterGrant

After installing a demand planning solution from Prescient, West Chester, Pa., and new business processes, Weigle says both short- and medium-term forecasts have improved significantly. "As a result, when we are forced to expedite inbound freight, we know we are doing it for a good reason," he says. "And when we transfer a product and change its identity by changing the trim, we know we are doing it to fill a hole that is true and real, not a perceived shortfall that may in fact not be real but merely the result of a forecasting error. Our supply-chain management metrics are in place and on the rise, our vendor scorecards show that our retail customers are happy, and our consumers are finding the products they want when they want them. The change has been dramatic across the board."

Headquartered in Smithfield, R.I., AAi.FosterGrant is North America's top supplier of sunglasses and assorted optical products priced below $30, a category representing 85 percent of all sunglass units sold domestically. The present company was formed in 1996 when AAi, exclusively a jewelry vendor for 35 years, purchased FosterGrant. The merged entity now supplies sunglasses, reading glasses and costume jewelry to hundreds of retail customers across a wide variety of domestic and international channels of distribution within the continental U.S., Canada, the United Kingdom, Mexico and Germany.

Optical products sourced from China and Taiwan constitute 65 percent of the company's business. Sunglasses are the dominant products in this sector, outpacing reading glasses by nearly a 3-1 margin. The remaining 35 percent of revenue comes from jewelry products sourced from China and Korea.

Wal-Mart and Target Stores, along with Rite-Aid Pharmacies' 3,800 outlets, account for the lion's share of AAi.FosterGrant's business, as the company maintains hundreds of SKUs of jewelry alone on the shelves of the chain-store giants.

Though AAi.FosterGrant owns all product design and development rights, the company contracts with offshore vendors to manufacture approximately 99 percent of its products. Vendors are responsible for sourcing all component parts, including product trim such as security tags, tickets, lens labels and store-specific packaging.

Manufacturing Outsourced
Inbound transportation arrangements primarily are governed by geography. Coordination between vendors in China and Taiwan and the company's designated freight forwarder is facilitated by AAi.FosterGrant's offices in Hong Kong, which function at the direction of the company's Supply Planning Department, based at the Smithfield headquarters. Shipments from other vendors are coordinated through APL Logistics. Inbound shipments move from the ports of Hong Kong, Shanghai and Quingdao in China, Busan in Korea, and Kaoshiung and Keelung in Taiwan, arriving at the Port of Long Beach. Containers then travel inbound via intermodal service to Boston, where they are cleared through U.S. Customs and routed on to the company's North American distribution center in Smithfield.

Approximately 90 percent of AAi.FosterGrant's outbound shipments move through the Smithfield distribution center. From there, the movement is almost exclusively via United Parcel Service to the company's 250 customers across the U.S. and Canada (except for Wal-Mart, which arranges to pick up its own freight directly). The majority of shipments are direct to stores, while a smaller share does move to distribution centers. Wal-Mart picks up its shipments directly from the Smithfield DC.

The company also takes advantage of contract logistics operations for special product rollouts and for other special events. This allows shipment consolidations to occur overseas close to the suppliers and then be routed to contract facilities on the West Coast for broadcast shipment to stores and distribution centers throughout North America.

Strategy Overhaul
Prior to 2001, inventory and procurement plans at AAi.FosterGrant were driven by rudimentary forecasts generated within the company's ERP system. These operational forecasts were updated monthly with little or no user intervention, says Weigle. He abandoned the ERP system's planning functionality shortly after his arrival at AAi.FosterGrant in the spring of 2000. His mission at that time was to develop and deploy a supply-chain strategy involving people, process and technology that would drive cash out of the company's inventory while simultaneously supporting corporate growth plans.

"At the time, significant gaps existed within AAi.FosterGrant's demand planning, supply planning, and order fulfillment operations, which together caused the company's overall supply-chain planning function to fall far short of 'best in class'," says Weigle. For example, there was no formal supply-chain management organization in place, nor was there any clearly defined organizational alignment, scope or integrated workflow procedure, he says. Because of poor order management procedures, the demand history residing within the ERP was corrupt; and the ERP system's "black box" forecasting functionality provided, at best, rudimentary forecasts with no user interaction. "We had no standard operating procedures to govern inventory availability or order fulfillment, we lacked a vendor delivery window policy, and we relied on improper lead-time assumptions and poor DRP [distribution resource planning] functionality," Weigle explains. "Add to that the fact that we had no reporting of supply-chain management metrics, and it was clear almost at first glance that a complete overhaul of AAi.FosterGrant's supply-chain processes was in order. We had to put in place the proper foundational elements for a successful, best-in-class supply-chain strategy, and we knew that the selection of the appropriate enabling technologies to support these foundational elements was going to be of paramount importance."

With respect to demand management, Weigle's goal was to achieve a highly flexible, responsive and interactive process that could serve as a cornerstone to the company's overall supply-chain management strategy. "The primary objective here was to obtain a demand management system that could be implemented within a distributed environment that would allow us to accomplish a series of key tasks that we outlined in the RFI [request for information] we issued to vendors," he says.

For example, the company sought a system that could facilitate the management and hands-on maintenance of demand history to ensure a true and non-overlapping picture of customer demand. There was a critical need for the ability to generate highly accurate point-of-sale forecasts, both short-term and medium-range. Also critical was the capability to translate point-of-sale forecasts from a store-level frame of reference to shipping forecasts from the distribution center's frame of reference. Other requirements included the ability to support the analysis of collaborative forecasts received from premier retail trading partners; facilitate new-item introductions; provide exception management capability; and support the multiple faces of a "forecasting pyramid," including sales forecasts, operational forecasts and budget forecasts.

An RFI containing a number of questions specific to functionality, ease of use, implementation timeframe, product support structure, flexibility, scalability, architecture and other critical factors was assembled and dispatched to a select group of eight software vendors, which included Prescient Systems. AAi.FosterGrant then enlisted the services of the Boston branch of CSC Consulting to assist in evaluating software vendor responses to the RFI. Once the field had thinned to 4 finalists, vendors were asked to provide in-house demonstrations of their products to display first-hand the feature functionality as outlined in their particular responses. After viewing the demonstrations of each vendor, the elements of the RFI were weighted for importance and the software vendors were re-scored.

The entire selection process took approximately 6 weeks, culminating in the October 2000 selection of Prescient Systems. Best known for its Demand Planning product, Prescient now offers, as a result of its acquisition last year of Proasis Inc., a multi-plant advanced planning and scheduling product, APS Workbench, as well as an advanced scheduling product for companies that need a more discrete scheduling solution. Prescient also developed the Forecast Collaborator product, designed to be used by companies that want remotely based employees, such as sales personnel, to participate in the sales forecasting and planning processes. The company also has two solutions that leverage the power of the internet. One is Collaboration Engine, a template-driven workflow management tool that has already had applications built for returns management, product promotion planning, and new product introduction. The other is a supply-chain portal product called Knowledge Foundation, a management desktop tool designed to deliver to managers not only information from internal and external sources, but the ability to drill down to the details behind a company's supply-chain performance.

Inventory Reduction
"Prescient's weighted RFI scored 40 percent higher than the nearest competitor," says Weigle. "Their standard feature functionality fit our business vision almost perfectly and they won hands down." An implementation workshop was held in late October, and the implementation actually launched in November 2000 and completed in February 2001.

"The Prescient Demand Planning solution has been in place for one full year now and is a major cornerstone of our supply-chain strategy, solution and workflow," says Weigle. "Our inventory is down 33 percent compared with last year on a month-to-month basis, and our year-to-year peak inventory level is down 46 percent. Order fill-rates for sunglasses have risen to 98.9 percent so far this year, up from an average of 94 percent in 2001 and 88 percent in 2000. We can now forecast consumer point-of-sale demand five months in advance with only 30 percent to 35 percent forecast error. DRP exceptions are down by 80 percent, stock transfer costs have been reduced by approximately 50 percent, and our required warehousing space has been reduced by 40 percent compared with 12 to 18 months ago. And we've closed all of our overflow warehouse facilities."

Today, AAi.FosterGrant forecasts its entire business - thousands of SKUs across hundreds of retail customers - with just three demand planners. The company collects point-of-sale data for the approximately 75 percent of its business generated through its premier retail partners. Using either EDI transmissions or web-based communications, that point-of-sale data arrives and is channeled to a dedicated data repository at the Smithfield headquarters, where it is used for forecasting, business analysis and future product design.

Product orders from the premier retailers also arrive via EDI. For other customers, AAi.FosterGrant often generates the orders as part of a vendor-managed inventory program, while the company's 800 field service representatives feed orders into the company's systems via telephone from customer locations. Some smaller customers call or fax in their own orders, all of which are received in Smithfield. The confusion that once surrounded types of orders and their respective value to demand planning has evaporated.

Order Fill Rates
On a recent vendor scorecard from Wal-Mart, in-store in-stocks, a measure of the ability to keep the retailer's shelves fully stocked with AAi.FosterGrant merchandise, were rated at 98 percent. Order fill rates for Wal-Mart so far this year hover at 97.8 percent. Target has cited order fill rates of 98.3 percent against its goal of 95 percent, with on-time shipment performance at 100 percent, matching Target's high goal.

Weigle and company plan additional changes on the supply-chain management front and currently are in the process of further expanding an already significant VMI program, based in large part on Prescient Systems' technology. The ongoing expansion includes incorporating dynamic "plan-o-gram" optimization using linear programming technology, a program already in operation for Meijer and Wal-Mart Canada, says Weigle.

"We will also be improving upon and refining our DRP strategies," he adds. "We have removed the DRP planning function from our ERP system in favor of a 'home grown' custom-developed system and currently are engaged in discussions with Prescient Systems regarding the new DRP planning tool now being developed."