ABSTRACT
Organizations can be managed in different ways. The reactive “laissez faire” approach relies on consumer complaints, liability suits, shrink ing market share or financial losses to trigger efforts to remedy current problems. In this mode of thinking, there are no dedicated efforts to anticipate problems or improve current conditions. An alternative ap proach, that requires heavy investments in screening activities, relies on inspection of the product or service provided by the organization’s customers. One set of statistical tools that apply to such a screening, or inspection-based management style, is acceptance sampling. Using such tools enables decision makers to determine what action to take on a batch of products. Decisions based on samples, rather than on 100% inspection, are usually more expedient and cost effective. Such deci sions, however, only apply to the specific population of products from which the sample was drawn. There are typically no lessons learned by inspection-based management and nothing is done to improve the con ditions of future products. Quality control requires a different style of management. This third management approach focuses on the processes that make the products. Data collected is used to track process perfor mance over time, identify the presence of assignable causes and expose common causes that affect the process in an ongoing way. Assignable causes occur at specific points in time and usually require immediate corrective action, while common causes are opportunities for improve ments with long-term payoffs. Feedback loops are the basic structure for process-focused management. For a comprehensive treatment of in dustrial statistics in the context of three management styles: inspection based, process focused quality control, and proactive quality by design, see Kenett and Zacks (1998).
