Matching Assets to Supply Chain Demand at IBM Microelectronics
The Problem
The IBM Microelectronics Division is a leading-edge producer of semiconductor and packaged solutions for the networked world, supplying a wide range of customers in market segments like application-specific components, embedded controllers, wireless, microprocessors, memory, and storage. In the early 1990s, IBM decided its microelectronics division should expand from producing parts exclusively for other IBM locations to producing a range of products for diverse internal and external customers. To do so demanded an overhaul of its supply chain system, which required analytics expertise in supply chain optimization.
The Analytics Solution
IBM Microelectronics developed intelligent models for matching assets with demand to determine which demands it could meet and when, and to provide manufacturing guidelines. In 1994, an IBM team began applying analytics techniques to build these tools, interweaving linear programming with a traditional material resource planning algorithm and a heuristic matching process. The team deployed three core applications that determined customer commitments and manufacturing requirements, identified the best use of manufacturing resources, and facilitated fast response to customers placing orders.
The Value
The team's analytics-driven work improved manufacturing utilization and customer-order response time. It resulted in asset utilization improvements of $80 million when it was launched in 1999, and was expected to yield $200 million a year in improvements subsequently. It also reduced average order response time to less than one day, with on-time delivery increasing from 90% in 1998 to 97% in 1999. Meanwhile, customer satisfaction increased 20%.
