| Abstract: | An organization that has but a single product would be a rarity in the United States today. While single-product firms and tools such as break-even analysis may have been in vogue at the turn of the century, today jointness pervades the scene. The bank that issues a mortgage to a customer also may have managed his or her checking account or granted a student loan to his or her daughter. The department store that sold a suit one week earlier may have attracted the customer with its furniture, toys, and cosmetics. The hospital whose obstetrical team delivered the baby may have been chosen because its excellent orthopedics department set the husband's broken leg three years prior. The sporting goods manufacturer may sell its golf clubs because of its reknown for skis and tennis rackets. It makes no difference whether it is products or services or departments; all share equipment, time, and reputation. The list goes on forever; jointness is the rule, not the exception.Management can pretend that there is only one product and attempt to isolate its effects, as if it were in a vacuum. However, pretending does not eliminate jointness and the purifying assumption is apt to mislead analysis. So how should jointness be treated?Other than the mathematical programming models of the last forty years, only one model for dealing with jointness has been widely adopted for most of the twentieth century; that model is the retail inventory method. The retail system typically is used by department stores to control inventories that consist of 100,000 or more items.The retail system recognizes costs and prices, inflation and obsolescence, mistakes and successes. It tracks purchases and sales, deals with discounts and transfers, and controls investment while providing an estimate of ending inventory. Fortunately, the retail system, the favorite of the department store and the boutique, could be used in the factory as well as in a variety of service and nonprofit organizations.This article briefly reviews the retail method and then uses linear programming (LP) and postoptimality analysis to modernize the traditional conceptions—to attempt to make a good system even better.Many constraints are studied including overlapping seasons, sales quotas, target inventories, product life cycles, and spatial and monetary limitations, among others. In the process, one comes to appreciate the size of the inventories being controlled, the human and departmental interactions, and the usefulness of the LP-retail system (“knotty retail” is perhaps an appropriate name) not only to generate cash budget estimates and to manage stock levels, but also to select the most satisfying alterations of managerial policy.The article concludes by discussing further constraints and other types of organizations that might benefit from adapting the retail system. |
