Abstract
Abstract
Any working process is designed with the intent to create an output that will satisfy either predetermined requirement or a potential customer. One of the parameters that will influence customer satisfaction is the variation of the working process output that is within or outside the originally predetermined or expected range of variation. The meaning of ‘expected range of variation’ in this context is related to a relatively recently recognized dimension of quality ‘perceived quality’, whereas a ‘predetermined range of variation’ is a synonym for Crosby's quality definition of conformance to requirements and specifications (Crosby, Quality is free: The art of making quality certain, 1979, McGraw-Hill). The working process output is controlled by variables that are functions of time and/or working environment specifics. A change of variables may create an ‘uncontrolled’ output that shows a level of variation outside the predetermined or expected requirements. From a management prospective this is called either a problem or an opportunity.
The ‘uncontrolled’ outputs have a negative impact on the performance of the working environment, which is, in most cases, closely related to the production cost. The unknown assignable cause of variation (or root cause of variation) that is contained within the unknown critical set of variation (or family of variation) is the main reason for the uncontrolled outputs. In view of the cost as a function of time, in any working environment, it is important to identify quickly the critical set of variation which would lead to a quicker problem solution and better control of the financial figures.
Multi-vari is a statistical engineering method used to analyse a set of data acquired in an organized manner and analysed graphically or analytically. There are two basic applications of this method: (i) to determine statistically the homogeneity of the data distribution and (ii) to identify the critical set of variation that contains the assignable cause of the variation. This paper focuses on the second application of this method, which gives the greatest return on investment with respect to day-to-day operations in different working environments. The multi-vari method presented and application examples should help readers to understand the essential features, to evaluate the method, and to identify its potential for application in their areas of expertise.