Overview

Recently, a manufacturer was facing the kind of problem that most companies want to face: an increase in orders was overloading the production scheduling process. Customers could not find out promised dates of delivery, order status, etc. The whole order entry process encompassed many functions (an engineering review, purchasing status, etc.) that affected production capacity and lead time. This company was concerned about customer retention: customer satisfaction was well below a six sigma level. Immediate action was needed, but their key question was whether process improvements could speed up the flow or if they needed to start hiring more staff.

Another company had a much different situation, but faced similarly dire consequences. Their manufacturing process relied on the precise alignment of high-frequency communication diodes, which are about the diameter of a pin. Because these diodes are so tiny, it s easy for them to land on an edge, creating a defective connection and thus becoming scrap. The yield at the key workstation was about 87% (about 2.7 sigma). This company wanted to be at six sigma levels, because that would mean tremendous savings in terms of scrap, rework, cost, and capex investment.

At first glimpse, these process improvement challenges sound quite different. One company is dealing with a transactional process where the goal is to quickly and accurately transfer information among groups; the other is concerned with highly specialized technology at a single workstation. Yet both of these situations can be (and were) improved using the same basic model of...