Paf Model 1t4q51

49 Quality costs: PAF and the process model- are they compatible?

w. KEOGHa and J.F. DALRYMPLEb aLecturer, Centre for Management Studies, University of Aberdeen, Old Aberdeen AB9 2IT, United Kingdom bSenior Lecturer, School of Management, University of Stirling, Stirling, FK94LY, United Kingdom

1.

INTRODUCTION

This paper seeks to explore the relationship between the Prevention-Appraisal-Failure Model outlined in BS 6143:Part 2[1] and the Process Cost Model presented in BS 6143:Part I [2] and attempts to determine whether or not these are compatible standards. The simplicity of the PAF model, which emerged from early work by Juran[3], Feigenbaum[4], and Masser[S] among others, often presents real difficulties in implementation for organisations. For example, in identifying costs, in determining which costs to measure, and how to report these costs [6]. The Process Model is presented in a deceptively simple manner in the British Standard. The model is seen by some as a step forward as it gets people to examine processes in more detail[7]. However, questions arise as to its appropriateness in coping with systems in a state of change, in the knowledge and expertise required to operate the model, and reporting mechanisms which will aid the TQM programme or initiative. BS 6143 ptl makes it quite clear that Total Quality Management requires not only the management of processes, but also an identification of an aim for the outputs from the process. It is stressed that the management of processes is fundamental to improving quality in either manufacturing or service industries. Further, it is brought home that every person contributes to, and operates within, a process. From experience, it is clear to the authors that every process should have an owner and this is emphasised within the standard. The idea that people should 'own the process' implies that any problems associated with the process will be dealt with and brought to a swift conclusion by the owners. All business activities relate to processes and this is emphasized in that Quality Cost Models should reflect the total cost of each process rather than some arbitrarily defined cost of quality. However, arbitrary costs, collected according to some rationale, may be better than having no cost estimates at all. 2

PAFMODEL

The American Society for Quality Control (ASQC) investigated Quality Costs and, in 1967, issued a document which illustrated what they were and how they could be dealt with. A second edition followed in 1971 [8]. The importance of these costs and how they fitted four primary areas of cost breakdown; Prevention, Appraisal, Internal Failure, and External failure was explained. It was made clear that, in order to communicate the importance of these costs and gain within the organisation, they must be communicated in the language of management - money I On the surface, the guidelines appear to be straightforward and achievable with some detail included on the aims of running a pilot programme. In the UK, in 1978,[9] this sentiment was echoed as it was estimated that Quality Costs amounted to £1 0 billion of the 1976 manufacturing output (or 10%). BSI issued BS 6143 in 1981[10] modeled on the American guidelines and written in a more somber tone. The PAF Model as it presently stands appears to cover most situations and the notion that it will cost money to

G. K. Kanji (ed.), Total Quality Management © Springer Science+Business Media Dordrecht 1995

314

Part Five

Quality Systems and Learning Culture

operate the system makes it clear that quality is not free. However, if the standard is applied properly then costs could be offset against savings in much the way that Crosby indicated[ll]' Many difficulties do arise in practice. For example, access to data is not always forthcoming. There may be a number of reasons for this, including; no collection mechanism in place the wrong sort of collection mechanism in place cultural or organisational barriers. The treatment of costs can vary from company to company. Possibly even within the ing Department, as well as with personnel in relevant departments, where there may be disagreements about identification, allocation, and access to data for reporting purposes. Sampling issues are also very important; and for example, activity sampling may be prone to error and lor bias. Despite shortcomings, it should be stressed that the P AF Model provides a good starting point. In most organisations there is evidence of Failure Costs. For example, data on customer complaints, delayed orders, rework and scrap, are usually accessible if not fully documented. 3

PROCESS MODEL

The Process Model standard has many strengths and is very good in the explanation of the Cost of Conformance (COC) and the Cost of Non-Conformance (CONC). This is used throughout the standard as a way of illustrating how to develop a Quality Cost, Total Quality Management package. We may even regard the Cost of Conformance as the minimum cost of being in business, as things are currently done. Interestingly, it is also emphasised that COC and CONC are both capable of improvement. Further, the process itself may be considered at any level within the organisation. The model of Quality Costs should reflect every cost in the process and thus ultimately show the total cost. Improvements could then be identified; not only at the top level, but within every level of the organisation. Processes may even be identified as redundant.

3.1

Developing a Framework The underlying framework of the standard comes from ICAM which stands for Integrated Computer Aided Manufacturing and the methodology employed is known as IDEF or the ICAM definition method (IDEFo) [12]. It is a structured approach to problems i.e. identification as well as analysis. It is a systems approach where the whole system is viewed and then broken down into steps and sub-steps and it can be used in many situations[13]. The first thing to do is to determine the steps within the organisation being investigated. Next, relationships are determined e.g. How does one communicate from one department to the other? Where does the instruction come from? This can provide a basis for investigation. For example; Design and the Design Function, Manufacture and operating problems, Purchasing etc. Where do we begin? What are we looking for') The standard advocates that a number of teams be set up. These teams should include multi-functional, highly qualified people. Their brief is to identify processes; to break these processes down; to identify the costs in these processes; to determine the categories of costs i.e. COC or CONC; and to prepare cost reports. Cost Reports would be used to advise the senior managers about the problems associated with the process(es). Certain key points have to be understood about the IDEF methodology which are not necessarily apparent from the standard. i It is hierarchical and shows expansion from the top level to the bottom level of the organisation down through department to process. 11 A system of diagrams must be internally consistent.

Quality costs: PAF and the process model III

IV

315

A number of boxes represent activities on each level. A maximum of six boxes can be used in anyone process diagram. The boxes contain an active verb or phrase. Examples of this could include "Fabricate" or "Construct the Unit". There is also an association within the boxes as well as between diagrams. is used throughout, not just within each diagram or layer but within multi-layers. In this way the whole picture of control of the processes within the organisation begins to emerge.

3.2 Problems in Use Complexity: To begin with, the standard does not formally identify or explain the IDEF methodology in detail. Although the top level diagrams are illustrated well there is a great deal of complexity in the diagrams; going from the zero level diagram, or upper level, to the first level diagram means that the situation being described is expanded or exploded. When the analyst goes to the next level down, the activities from the previous diagram are exploded into a number of other activity diagrams and events. This is complex in practice and the standard stays comfortably at the top level. Training: In order to prepare work for using IDEF, the people concerned must be trained. It is not enough to say that the organisation has a good group of individuals who are dedicated to the task. In reality the team must be aware of the problems when trying to break down say, their own operations. If the people concerned do not understand what they are doing, or analysing, they may trivialise or miss key areas of the process they are investigating. Definition: This leads on to the actual definition of what the team are looking at and looking for. In order to define the system they have to be unbiased. A number of overlaps will occur between departments, between sections and between teams. These overlaps serve the purpose of determining where the real problems lie. Misinterpretation: Perhaps one of the main problems in use is misinterpretation of the findings. If the people concerned do not fully understand what they have discovered from the analysis and if the people they are reporting to do not quite understand the system, then it is fairly obvious that misinterpretation of the findings will arise. Consistency: Further, it is vitally important that there is consistency in the analysis and reporting of the systems which have been investigated. This may be achieved by using the same team, or, by having the teams trained to the same level. This means that teams will not only require the same training, but the same levels of understanding of process investigation. Ownership: Finally, the key area of ownership can be used to highlight many existing problems. Very few people want to own failure and a serious problem with any cost of quality exercise is that systems problems can be uncovered relatively quickly. The identification of problems can lead to the perception that people are not doing their jobs properly. Tact is required to ensure that ownership is retained and not rejected.

4

CASES

Firstly, consider the case of a company at the forefront of new technologies. The organisation designs, develops, and builds prototypes of new products; as well as producing standard products. New methods and processes are being developed which improve product output as well as internal communications and speed of response. The organisation structure also changes with duties and responsibilities of key individuals being redefined over time. Overall, the situation is one of rapid change. In such a situation, a computerised process model may work well. After careful investigation it was felt that the P AF Model would be more appropriate as requirements of building a process model would be onerous and would not repay the effort. The subsequent P AF Model allowed for flexibility in identification and quantification of costs.

316

Part Five

Quality Systems and Learning Culture

The second case considered is a company in the chemical industry. In this instance, the processes tend to be relatively stable, procedures are very well established and change is relatively slow. Although the processes could be modeled at the highest levels, it became apparent that detailed knowledge was required to apply the Process Model. It was felt that training was necessary to identify the relationships within the processes and to apply the methodology. The PAF Model was easier to apply, certainly on a superficial level, and gave some indication of the Quality Cost task.

5

SUMMARY

It is evident from the two parts of the standard that they could complement each other. However, this is not straightforward and, at best, the Process Model adds another set of difficulties to those experienced when using the PAF Model. Problems with both the PAF and the Process Model include; accessibility of data, treatment of some costs, and sampling errors and bias. Furthermore, the Process Model has the additional problems associated with understanding, complexity of the approach, and proneness to misinterpretation of the results. The Process Model is therefore a much more difficult starting point and requires much more sophistication. Recent evidence suggests that such effort is not readily repaid.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

BS 6143: Part 2, Guide to the Economics of Quality: Prevention, appraisal and failure model, British Standards Institution, London (1990). BS 6143: Part 1, Guide to the Economics of Quality: Process cost model, British Standards Institution, London (1992). lM. Juran, Quality Control Handbook, McGraw-Hill, New York (1951). A.V. Feigenbaum, Total Quality Control, Harvard Business Review, Vol. 34 Pt 6 (1956), pp 93-101. WJ. Masser, The Quality Manager and Quality Costs, Industrial Quality Control, Vol. 14 No.6, (1957) pp 5-8. l Camla, editor, Principles of Quality Costs, second edition, American Society for Quality Control, Milwaukee,Wisconsin (1990). B.G. Dale and J.J. Plunkett, Quality Costing, Chapman & Hall, London (I 992}. ASQC Quality Costs Committee, Quality Costs - What and How (2nd edition), American Society for Quality Control, Milwaukee,Wisconsin (1971). DP, A National Strategy for Quality, Dept. of Prices and Consumer Protection, London (1978). BS 6143, Guide to the Determination and Use of Quality Related Costs, British Standards Institution, London (1981). P.B. Crosby, Quality is Free. Mentor, New York (1979). O.T. Ross, et ai, Architects Manual ICAM Definition Method, 1DEFO Cam-i Inc, Texas (1980). GJ. Colquhoun, R.W. Baines and R. Crossley, A state of the art review ofIDEFO, Int. Journal Computer Integrated Manufacturing, Vo1.6, No.4 (1993) pp 252-264.