This paper reviews some problems and solution strategies to overcome the problems of short forecast horizons, changing forecasts and short term demand variations in a long supply chain. The purpose is to identify the potential for and limits to the use of integral information for optimizing the performance of the chain as a whole.
The points highlighted in this paper are not new, but well described in literature and even older study books. In practice, however, these lessons seem to have been forgotten, and expectations of collaborative systems are based on false assumptions. One of the main false expectations is that the “bullwhip effect” (of amplification of demand variations to the upstream processes) can be avoided with integral information communication in the supply chain. Rather, the bullwhip effect is inherent in a long supply chain, and planners can only employ strategies to ward off the adverse effects.
Another false expectation is that integral availability of information throughout the chain will automatically lead to better performance. On the contrary, indiscriminate use of all new information will only cause a very nervous system. Integral information is only the beginning of improvement, it merely provides a sound basis for trade-offs that need to be made throughout the chain. The main trade-offs that supply chain planners must make are between forecasting effort, capacity change costs (order nervousness), and investment in buffer inventories. Existing work on these trade-offs should be taken into account when designing new supply chain management solutions. This paper tries to bring these trade-offs and the theory that drives them back into the focus of developers of modern systems.
In the internal supply chain of a large company the trade-offs can be made to optimize the whole. Useful future work can be done on supporting these trade-offs in networks of independent parties: often one party must make efforts to improve the performance and bring benefits for another participant. Management reporting must be carefully designed to support inter-company trade-offs, choices, highlight causes and effects.