Shingo’s Approach to Quality Management

Judgment Inspection: The kind of inspection which was discouraged by conventional TQM takes place at the final stage of production and ‘simply distinguishes defective from non-defective products and issues a ‘post-mortem certificate’. However, if the objective is to reduce the defect rate, defects must be uncovered as soon as they are produced. The owner (operator) of the process which produced them must immediately be informed so that corrective action is immediate. The idea behind the corrective action is to ensure that the root cause of the defect is identified and eliminated. Such a scheme of inspection is called Informative inspection.

‘Judgments inspections discover defects, while informative inspections reduce them. ‘The objective of inspection must be to reduce the defect rate and not to detect defects. This objective is realized when inspection provides ‘immediate, accurate feedback at the source of defects.

Shingo (1989) was greatly influenced by an observation made by the managing director of Matsushita Electric in 1965. This observation was that each consumer purchased only one television set out of the company’s total output of one million set. If the set produced by a consumer was defective, his faith in the company would be destroyed. However, under sampling inspection and SQC approaches, this would have been perfectly acceptable to the company. This led him to rea1ise the need for methods which could assure quality with the thoroughness of 100% inspection and the ease of sampling inspection. Shingo also noted that control charts, in practice, are a form of judgment inspection. This is because there is usually a time lag between occurrence of the defect and action being taken on it. As he put it (Shingo 1989), ‘somewhere along the line means and ends became confused; the notion that you cannot have quality control without using statistics supplanted the original, valid concepts. ‘

Shingo suggested two kinds of informative inspection:

Self inspection: the worker is given the responsibility of inspecting his or her products.
Successive inspection: the worker at a station inspects each piece passed on by the worker at a preceding station.

Though self-inspection will usually lack objectivity, it can be enhanced with the use of devices that automatically detect defects or mistakes. Such devices are called Poka Yoke or mistakeproofing devices. Interestingly, the term originally used by Shingo was baka yoke, which means foolproof. The term was abandoned because it could be misconstrued) imply that the operator who produced the defect was a fool. Poka Yoke devices can be used to shut down the machine or line when an abnormality detected, or to issue a warning signal.

Poka Yoke devices are typically simple and robust. Shingo provides an example of one used at Matsushita Electric. In a vacuum cleaner packing line, the packed box was weighed to ensure that no parts were missing. In spite of this, complaints of instruction manuals and small parts being missing were reported, the production manager decided to go in for a Poka Yoke devices. A stopper on the roller conveyor line would be retracted only if two limit switches– one activated when the stand containing manuals was pushed down to pick up a manual, and a similar switch for the stand containing the small parts-had been activated. Otherwise, the box would be stopped and a buzzer would sound. Complaints dropped to zero within two months.

Figure 3.9 shows another example of a Poka Yoke device. Such devices illustrate a basic idea: it is human to make errors-but errors need not lead to defects, if they are detected and prevented from occurring in time, Errors and defects are not the same thing. Source inspection is the term used by Shingo to denote inspection that prevents defects by controlling the conditions that influence quality at their source.

Figure 3.9: Poka Yoke for Bending Cover Edges. An automobile component processed for the left and right sides of the vehicle had identical contours except for the position of a hole.

The part was designed to be bent on the right side, but occasionally right- and left-handed components were interchanged. A limit switch was installed on the press at the point where the hole should be located. If the hole was not in that position, the limit switch would be activated, the press would shut down and a buzzer would sound to alert the worker.

Control charts may have the same purported objective of reduction of defects through highlighting special occurrences, but Shingo’ s system (which he called zero quality control) of source inspection and Poka Yoke consists of a more hands-on approach to quality management.

The efforts of Shingo and others propelled some manufacturer to what we would today describe as World Class quality. This achievement was reached two decades ago as related to Shingo (1987).

In 1977, the Shizuoka plant of Matsushita’s Washing Machine Division had succeeded in attaining a continuous record of one month with zero defects on a drainpipe assembly involving 23 workers. Five years later, Schonberger’s (1982) wrote that ‘statistical tools are not quite so important in the spectrum of concepts that constitute Japanese total quality control as are some of the more conceptual factors.’ He identified three key managerial traits that lead the Japanese to superior quality performance:

Making quality the responsibility of the production (not QC) department.
The habit of improvement-as opposed to the desire to keep things stable.
The drive for perfection.

Schonberger’s described seven basic principles of TQC as practiced by Japanese manufacturers:

Process control-every process was checked for control, but the responsibility was shifted to production personnel. Also, SQC was not the primary tool that was used for this.
Easy-to-see quality, leading to what would later be called the ‘visual factory’.
Insistence on compliance-‘urgency’ would not be accepted as an excuse for shipping out defectives.
Line stop-the authority to stop the production line in case of a quality problem was given to workers. However, the underlying philosophy was that each line stop must lead to a root cause analysis to eliminate the source of the defect.
Correcting one’s own errors-the worker or work group which made defective parts is responsible for rectification. There are no separate rework lines.
The cent check- acceptance sampling techniques for items lots were rejected by the Japanese because:

The notion of lots itself is in opposition to agile manufacturing concepts.
The statistical tables used for acceptance sampling require the acceptable per cent defective as an input; even in 1982, the Japanese were talking about defects in parts per million (PPM)
Sampling does not ensure that no customer gets a defective item.