As a major innovative manufacturer and distributor, Hasbro sought to increase profits by reducing logistics and distribution costs. Of primary concern were the number, location and size of distribution facilities and the impact on transportation costs resulting from movement of production capacity to China and the acquisition of new product lines. Finished goods and games were staged in dedicated warehouses rather than full-line distribution centers. Hasbro also sourced its various products both domestically and internationally. In addition, the company was interested in starting with a green-field approach to defining an appropriate number of and locations for distribution resources based on current and future customer demands. This complex scenario created the challenge to integrate planning, eliminate duplication of product storage locations and consolidate distribution resources.

The toys business is extremely seasonal - supporting heavy demand in the fourth quarter. Also, being very competitive, approximately 65% of the production each year involves new products. The top sellers are not known early in the year. Therefore, 90% of the forecasted demand is produced by the third quarter. This is called "pre-building" and is production-to-stock. This is done not only to meet demand, but also to "protect" the production capacity late in the year to meet late season priority customer orders for products that are selling well. Finished goods are stored at only one location per SKU in the US to facilitate rapid responsiveness to third and fourth quarter customer orders. The required production cycles for late-season production-to-order orders are typically five days. Furthermore, the client was experiencing a period of great instability due to wide shifts is sales volumes and the acquisition of several smaller companies.

Hasbro contracted Tompkins Associates, Inc. (Tompkins) and Automation Associates, Inc. (AAI) to perform a joint study of the existing and alternative distribution network costs and performance capabilities. To achieve this optimization, Hasbro wanted to consider all alternatives, including - but not limited to - consolidation of operations, regional distribution and warehousing in China. These large-scale changes to the way Hasbro was doing business required a methodology that had capacity to prove and implement a successful plan. So, as part of the study, a simulation model of the distribution network was developed. Simulation offers analysis capabilities and a greater degree of flexibility, to account for real-time factors such as seasonality, which is difficult to achieve using other methods. The benefits of simulation for this particular project included strategic/planning level of analysis (not tactical), the ability to produce a result with a tight project schedule and having the environment for a comparative analysis paradigm to evaluate several distinct scenarios. The simulation model was used to support the evaluation of alternative product delivery systems. The results of the study would be used to initiate detailed planning efforts to choose and implement the best solution.

Data collection comprised the first phase in the simulation model development. The categories of collected data were customer order details, manufacturing volumes and schedules, distribution operations details and transportation network details. Once organized, major categories of model inputs included customer representation and location, order patterns by location and product lines, manufacturing data and distribution center logistics detail - including inbound and outbound transportation data and inventory representation.

Historical customer order information was aggregated to capture the product specific order volume throughout the year. Customer order patterns varied by product line, by season and by customer organization. The top 80 customer locations were found to constitute the majority of the order volume for the company (80%+). These 80 locations had the potential to initiate product orders every day of the year.

As a result of the nature of the production-to-stock, manufacturing orders were placed well in advance of the actual customer order receipt. The lead-time varied based on whether the products were manufactured domestically or internationally. Consequently, the lead times for production and inbound transportation have significant impacts on the inventory storage costs associated with forecast volumes and lead times from product receipt to actual order receipt.

The model included a defined percentage forecast error into the production forecast generating inbound product to the distribution network. The result of this deviation was an increase in the total inventory in the system, an increase in the inbound shipment costs to the regional distribution centers and an increase in the storage and obsolescence costs experienced within each center. Alternatively, by selecting specific product families for distribution from restricted sites, the distribution network costs were reduced as a result of the forecast error on volatile products. The analysis also included products from five different product families. The orders, production forecast, storage locations and initial inventories were established for each of these product families. Various scenarios were executed to understand the impacts of full inventory stocking at each of the distribution center locations versus dedicated inventory fulfillment centers.

The central question - required number of facilities and their ideal locations - was analyzed through the execution and analysis of a number of scenarios. The scenarios considered included combinations of four distribution centers to just a single distribution center. The locations evaluated incorporated the experience of Tompkins Associates in locating facilities, as well as a desire to utilize existing locations. These scenarios were executed and compared to the existing system baseline model results. Additional scenarios evaluated included impacts of product stocking decisions within the distribution network, changes in the manufacturing locations, changes in the transportation network shipment frequencies and changes in customer order volumes.

Using these inputs, simulation provided the capability to accurately measure outputs. The outputs of this model proved to be highly effective due to the ability to use conditional inputs, measure effects of inventory policies and cash flow expenditures over time and use real-time production schedules.

Comparing the baseline simulation to actual costs and delivery times validated the results. Through a combination of selecting a three (3) DC network and implementing changes in the transportation planning, leading to higher consolidation of shipments, the models demonstrated a $4.6M savings in the annual logistics costs. This cost savings offset higher inbound shipping costs. This particular design of 3 distribution centers performed best due to land-bridge costs from current operations locations. The volumes used and the actual transportation unit costs produced simulated results that accurately represented the actual performance. Other alternatives provided similar savings potential, giving the client the opportunity to select among several different philosophies of managing their supply chain.

When compared to the base system cost, the alternatives executed resulted in savings profiles. Therefore, savings and simulated changes were considered to be valid and practical. Specific proven alterations to Hasbro's current supply chain included a different number and location of distribution centers, sourcing from China, mixed inventory management approaches, outbound shipment consolidations and improved product forecasting. Fluctuations in customer demand for high-risk products could best be offset with centralized distribution points servicing the entire country.

At the time this project was completed, Hasbro was conducting further detailed studies of cost savings associated with moving their manufacturing. Company decision-makers have taken ownership of the model and can continue to execute additional scenarios. Simulation gives these strategists the ongoing ability to represent the evolving policies that are used to manage the supply chain, providing an understanding of the range of performance and robustness of these policies through naturally dynamic conditions.