You are on page 1of 17

IJOPM 17,1

100

A framework for proactive performance measurement system introduction


Robin C. Daniels and N.D. Burns
The Morris Institute, Loughborough University, Loughborough, UK
Introduction As the drive to become world-class has progressed and intensified there has been recognition that, if manufacturing is to be used as a competitive weapon[1], then the shopfloor and the production cell will have key roles to play in the development, or survival, of many Western companies. The pivotal position of the cell leader or line manager in the implementation of novel strategies and tools has been recognized for some time[2]. Despite this, often little guidance is offered to the cell leader in terms of maximizing the results of the continuous improvement which is seen as essential to shopfloor and operations management. This is true especially where the management and leadership of people are concerned. Well-known tools of kaizen such as poke-ayoke, single-minute exchange-of-die (SMED) and the fishbone diagram, are relatively easy to implement and can result in large savings in a short space of time. However, many companies which have started out along the route of kaizen find that, once the shopfloor re-layouts have been completed, the Cshaped lines are in place and everything has been fool-proofed (or mistakeproofed), they then reach a watershed[3]. If kaizen, or TQM, is initiated without real attention to the more fundamental and sometimes abstract elements of the organization, then only superficial and, therefore, short-term improvements are achieved. Applying kaizen as an overlay on the existing culture represents a systematic lip-service to the philosophy which will often put paid to any future culture shift[4]. For companies which strive to become world-class and yet maintain their existing management accounting systems, this risk is particularly acute. The use of exclusively financial performance measures is now widely recognized as conflicting with the most fundamental aims of modern business[5]. Focus on the needs of the market and the customer is increasingly seen as the prime driver for successful manufacturing, and nowhere is the supplier closer to the customer than on the shopfloor, especially in a just-in-time (JIT) environment. The research underlying this paper took place in an automotive component manufacturer in the UK which supplies several domestic and European car manufacturers. The cellular manufacture and JIT systems in place formed the backdrop for the establishment of a cell-generated performance measurement system which was devised and run by the cell operators, under the facilitative guidance of the researcher. The customer knowledge and appreciation of cell

International Journal of Operations & Production Management, Vol. 17 No. 1, 1997, pp. 100-116. MCB University Press, 0144-3577

requirements led to a robust system which directed the team in its kaizen A framework for activities and led to substantial improvements in the cell-generated measures as performance well as of the existing productivity-based bonus system[6]. measurement In order to bring this about, however, the cell leader (who was also the researcher) found that commitment to the new system and real bottom-up, enthusiasm for kaizen activities could be brought about only through devolving 101 much responsibility to the operators and truly empowering them[3]. It is this pushing down of responsibility from the middle/line manager or cell leader to the shopfloor which is often a stumbling-block to fundamental culture shift on the shopfloor[7]. The traditional role of the supervisor, as director and autocrat, has long been outdated[8], and, while many companies recognize this, little practical guidance is offered in terms of ways in which the supervisor can stop directing and start to facilitate. The research methodology and findings presented here can be developed to form a framework to guide shopfloor management in the introduction and sustaining of self-regulating kaizen teams. Research methodology and findings The research took the form of an in-depth case study which was divided into two stages: Stage 1. Analysis of the existing performance measurement system from an operational viewpoint. Stage 2. The development of a cell-generated system to drive and direct cell kaizen activities. Stage 1 consisted of the application of two methods of analysis: (1) Questionnaires designed using pair comparison and semi-structured interviews were used to determine the perceptions of and attitudes towards the existing performance measurement system of a crosssection of the production function. Those questioned represented all levels of the function, from cell operators to production director, and so facilitated the testing of the extent of goal congruence in terms of performance-driver identification, continuous improvement and understanding of system dynamics. (2) Those assumptions identified were tested with reference to historical data and statistical analysis. Calculation of correlation coefficient provided quantitative evidence to support the initial qualitative evidence. The findings of stage 1 were not untypical, and shortfalls of the existing system included the following: lack of correlation between identified production drivers; different interpretations of driver interactions and cause and effect relationships, vertically and horizontally, through the organization; lack of ownership of the measures employed on the part of shopfloor operators;

IJOPM 17,1

102

lack of relevance of measures to world-class attributes and to the customer; and minimal drive and direction for kaizen activities on the shopfloor as a result[9]. These initial findings served to direct the stage 2 research towards the development of a system which would reverse the negative aspects identified. The principal objective of this system would be to drive and direct the continuous improvements of the research group. Once the group had decided on appropriate measures with associated goals and timescales, then the methods and frequencies of data collection and recording were finalized. As the research group was established, so an adjacent cell with similar characteristics, in terms of product and customer, was identified as the control group. The control group was allowed to continue operating as before, with kaizen directed from above and aimed at short-term technical problems. In this way a direct contrast between the two groups was made possible as time went on. Methodological triangulation was employed using questionnaires, interview and participant observation, augmented by historical, statistical analysis. Driver definition The first task of the research group was to define its operating parameters which would be drawn by the measures of performance selected. Under the facilitative guidance of the researcher, and through careful consideration of customer requirements, the research group decided on the following initial (and provisional) measures documented in Table I.

Measure Multi-skilling Quality (scrap and final rejects) Complete deliveries Tool and plant downtime

Frequency Weekly Weekly Weekly Weekly

Goal (%) 75 0 100 0

Timescale (weeks) 20 20 20 20

Table I. Research group performance measures

Some considerable time was spent by the research group on the definition of these measures. A sample calculation of each illustrates those definitions: Multiskilling Example: there are 24 individual jobs on the cell. Each person knows a percentage of these 24 jobs; so, if Joe knows 14 jobs, then his rating is 15/24 = 62.5 per cent. For week 1, if: Joe knows 15/24 = 62.5 per cent; Simon knows 10/24 = 41.6 per cent; and Thomas knows 12/24 = 50 per cent

then: percentage of multiskilling = (62.5 + 41.6 + 50) / 3 = 51.4 per cent. Scrap and final rejects Final weekly rejects Weekly rejects 100. Total weekly production volume Weekly percentage of scrapped parts : Weight of scrap parts produced 100. Weight of coil steel used Complete deliveries Number of assemblies required by customer 100. Number of assemblies shipped Tool and plant downtime Total plant and tool downtime . Total labour hours These measures were not arrived at through any statistical or Pareto analysis of existing cell performance (although some such measures were introduced during a later refinement process), but were based on the intuitive skills and experience of the group. It was felt that these measures best reflected the areas of weakness of cell operation at the time and the demands of the customer. However, it should be appreciated that the suitability of the measures chosen, as well as their associated goals, were secondary in importance to the process through which the group had gone in selecting them. The concern of the research was not with the selection of world-class or even good performance measures, but rather with the method of selection, the recording and use of the measures selected, and the behavioural factors involved. Effects on kaizen and performance The measures put in place by the cell began immediately to drive kaizen in a way that the existing measures had failed to do. In some cases (a quality problem for example) action had to be taken immediately; in others, trends in measures began to appear after a period of time (in tool and plant downtime, for example). In terms of the new measures, performance began to improve markedly within the two weeks of commencement. Over the period of stage 2 (40 weeks) reject levels fell considerably (zero defects achieved for two consecutive months) and downtime began to show a gradual reduction as underlying problems were exposed and addressed. Also training began to show tangible benefits in terms

A framework for performance measurement 103

IJOPM 17,1

104

of labour flexibility and in increasing awareness of quality problems within the cell. Table II gives a summary of the levels of improvement over the first 40 weeks of activity. As can be seen from Table II, labour efficiency, the prime driver under the existing system, improved by an average of 17 per cent (or 12 per cent in real terms) over the ten months of activity. This result is particularly significant since it provides evidence of the benefit of an approach where the factors underlying efficiency gains are the focus rather than efficiency itself. This result illustrates another finding of the research, namely that signs of improvement relating to a particular project often began to appear during the planning phase before the physical changes were made. Subsequent investigation showed that such premature improvement was driven by the improved understanding of the problem at hand brought about through the attention paid to it by the group and by the generally increased motivation and communication within the cell. So it was that the cumulative effects of improvements related to all the new measures were manifested in the substantive gains in overall productivity and efficiency. This finding is also important since it demonstrates the benefits of localized, direct and contingent measures in driving higher-level derived measures.

Measure Multi-skilling Scrap Final rejects Complete deliveries Tool and plant downtime Labour efficiency

Performance at week 5 25 14 1 100 8 69

Performance at week 40 46 6 0 100 5 81

Improvement 84 57 100 0 38 17

Table II. Research group performance (as percentages) at weeks 5 and 40

The centrepiece of the groups activity on the cell was the cell kaizen board which was used to record the daily data and for displaying the master graphs for each measure. Figure 1 shows an example of a graph developed and used by the research group. Training had been identified by the research group as a core need to improve internal communication, encourage teamwork and increase efficiency. However, while all of these aspects did show improvement, it proved very difficult to link training and skill levels explicitly, to these changes. What was demonstrated, however, was that the time allowed for training and other kaizen activities correlated strongly with labour efficiency, as Figure 2 shows.

A framework for performance measurement 105

Figure 1. Percentage scrap produced for cell X for weeks 41-80 (inclusive)

Figure 2. Relationship between cell LPI (including kaizen hours) and the percentage of total hours booked to kaizen for cell X for weeks 1-40 (inclusive)

Through ongoing interviews and the use of questionnaires with group members it was found that the fall in labour efficiency or LPI was closely associated with reduction in motivation and commitment to the change programme on the part of cell members when the cell leader was changed and the time allowed for training and general kaizen was severely reduced. Meanwhile the control group failed to make progress in relation to any of the existing measures (as had been the case prior to the start of the research) and morale in the cell was deteriorating as the research group became increasingly motivated. Table III provides a summary of control group performance in terms of the four common measures applied to each group, albeit from differing perspectives.

IJOPM 17,1

106

It was from the findings of stage 2, in particular, and from the experience of the researcher that the framework for the implementation of such a system was developed. Before discussing the philosophical basis of the framework, and the framework itself, the importance of cell management is worthy of mention. The role of supervisory management is a crucial one and the findings relating to this are of key importance.

Measure Scrap Table III. Final rejects Control group Complete deliveries performance at weeks 5 and 40 (as percentages) Labour efficiency

Performance at week 5 24 14 78 59

Performance at week 40 32 11 72 64

Improvement (33) 21 (8) 8

The importance of cell management As has been shown (Figure 2), the attitude of cell managers/leaders towards kaizen is crucial to the success or failure of the endeavour. Many of the points covered thus far raise many issues relating to the changing role of shopfloor supervision and training and continuing education for supervisors. Most important is the need for supervisors and cell leaders who have the confidence to delegate and devolve responsibility downwards. For those who prefer to direct rather than to facilitate, the shift away from an authoritative even paternalistic culture may prove to be impossible without additional support and training. The final part of the paper will cover two salient aspects of the framework presented: the balance required between the choice of the correct measure and the need to maintain the commitment and motivation of the team; and the importance of the choice of initial project. Finally, some requirements in terms of cell leader characteristics are listed. These requirements originate from questionnaire surveys of the research group which included elements of upward appraisal. Before discussing the framework it is necessary to outline the philosophical basis of the model. The underlying argument is based principally around contingency (as opposed to configurationalist) theory and the view that change should be driven by the internal and external forces pertinent to the business at the time. A contingent approach to driving improvement Contingency theory as applied to the design of organizations is based around the view that the characteristics of management structure and control should be contingent on the task in hand, the people involved and the surrounding

environment[10]. This is in contrast to the classical and human relations A framework for approaches which Bennis[11] defined as organizations without people and performance people without organizations respectively. measurement It has long been recognized that the complexity of organizational design precludes the universal applicability of more prescriptive approaches[12,13] and that taking a broader view of the world (including outside influences such as the 107 customer and market) will yield more and not less robust organizational designs. Contingency theory provides the facility to identify (and sometimes to quantify) internal and external forces at work and the resulting forms towards which the organization will be pulled[14,15]. In the field of performance measurement research there has also been a shift away from the prescriptive and highly formalized management accounting systems towards a more balanced approach which also takes account of internal and external forces through the application of non-financial, as well as financial, measures[16]. The profile of such a performance measurement system will be contingent on the situation of the organization including its history, internal politics and management style. Research has shown also, however, that by taking the decision process to a point as close to the customer as possible, and, at the same time, minimizing the inter-hierarchical interference in new measure development, it is possible rapidly to develop and implement new measures which largely are unencumbered by internal political or historical forces[17]. Such measures will override (in terms of behavioural consequence and, therefore, actions) existing highly formalized and prescriptive performance measurement systems. As stated previously, the aim of the research underlying this paper was not to define an absolute list of world-class performance criteria. Rather it was to examine the process of developing a new system and a set of measures by shopfloor workers which were contingent on their situation and which would drive improvement activities. A framework for the implementation of a cellbased performance measurement system, therefore, should be constructed on the premiss that the measures chosen, and at a deeper level the system configuration and operation defined, are contingent on the needs of the cell, the goals of the business and, most importantly, the needs of the customer. The requirements of suppliers to the subject company also are becoming increasingly important with the increase in inter-company kanban systems and other, wider, issues of supply chain management[18]. Performance measures, therefore, are contingent on the relative characteristics and strengths of: customer requirements; internal requirements; and supplier requirements.

IJOPM 17,1

108

Customer requirements are vital to the definition of the framework for all other requirement categories. All requirements, therefore, will provide direction and drive towards improvements to the cell which are relevant to the customer requirements. The action research described previously succeeded in generating and driving a demonstrable culture shift in the cell. This shift was driven ostensibly by injecting into the system, which at the time was driven purely by labour efficiency, a set of measures which would institutionalize the requirements of the customer into the culture of the cell. This shift in emphasis is illustrated as follows. Optimizing the measure-driven cycle It is apparent that in a given performance measurement system there are various cause-and-effect relationships in place which are driven by the system and which ensure that the measures employed become self-fulfilling (i.e. you get what you measure). The link between cause and effect in this case is the behavioural consequence both of the measure and of the system as a whole. Using this principle it is possible to map the progress of cause and effect relationships through a cycle. Figure 3 shows a measure-driven cycle for a company operating under a traditional accounting-driven system, such as was the subject company of this research. Figure 3 illustrates the essence of the introverted measure-driven cycle in which the behavioural implications of non-customer-focused measures drive the culture of the shopfloor away from world-class performance criteria and the requirements of the customer. Through the development of a cell-based performance measurement system it becomes possible to incorporate those customer needs into the cycle. Again, behavioural implications link cause-and-effect relationships, but this time the cycle becomes proactive and dynamic. Once the culture has moved far enough away from the introverted initial state, then the customer becomes the focus and non-financial measures increasingly drive the shopfloor along the path of continual improvement. Figure 4 illustrates this change of focus. While the above clearly focuses on the individual cell the model could be used also to represent an entire organization where world-class performance criteria, and therefore customer requirements, define the business and manufacturing strategies; research at this level is ongoing. In such a case the measures are adapted to be relevant to each hierarchical level, while a clear link between strategy and the shopfloor is assured through a definition of strategy in terms of direct measures of service to the customer. Such goal congruence[19] means that the benefits of local autonomy can be maximized and cell-based entrepreneurial flair (or deviance from the norm) will not detract from the central mission of the organization. Such a characteristic is defined by Peters and Waterman[20] who discuss the loose-tight properties of then successful companies such as 3M.

A framework for performance measurement 109

Figure 3. The introverted measure-driven cycle

Figure 4. The optimized measuredriven cycle

IJOPM 17,1

The theoretical model here outlined, which derives from the results of empirical research[4], underlies the practical framework and guidelines for the successful implementation of a self-sustaining cell-based performance measurement system. A model and guidelines for implementation The model described in what follows is based on the empirical research described in this paper and describes a procedure for introducing such a system elsewhere. The cyclical nature of the behavioural implications of performance measurement systems as discussed above means that the provision of a feedback loop for the continual reappraisal and improvement of measures is made relatively uncomplicated. The most suitable way to represent the sequence of actions required for the implementation of a cell-generated performance measurement system is the flowchart. However, it is necessary first to identify two distinct scenarios which the facilitator may encounter regarding the choice of appropriate measures by the team of operators (henceforth referred to as the team). It has been shown[4] that where knowledge of customer requirements is good the team is likely to select measures which are highly congruent with the requirements of the customer (see Figure 5). It may be the case, however, that the appreciation of customer

110

Figure 5. The expert measurement framework

requirements is insufficient to select valid measures. The question of what is A framework for valid is best answered by the customer, and so direct customer input would be performance favoured here. The second scenario demands modification to the implemenmeasurement tation flowchart, as is illustrated in Figure 6. The terms in parentheses at each stage describe the quality of the process at that point and indicate that sound progress to the next stage is possible. The 111 second scenario (Figure 6) depicts the situation at the other end of the continuum where the degree of customer appreciation at best is minimal. In this situation kaizen activity is initiated and driven by sub-optimal measures while an educational and training input is initiated in order to lead to the definition of a more suitable set of measures. The initial set is then refined after input from the customer and introduced at the measure review (which is initiated as soon as the refined set is ready). If a review decides that the existing measures are still acceptable, then this is fed back into the kaizen activity. If the set requires modification, the new measures are identified and these are fed either into the kaizen activity or into the training entity, as appropriate. It is possible to combine the expert and novice measurement frameworks to form a contingent measurement framework (Figure 7) which can be used whatever the characteristics of the team in terms of customer appreciation. The contingent measurement framework can be used to formalize the development of a team at any point along the knowledge continuum by means

Figure 6. The novice measurement framework

IJOPM 17,1

112

Figure 7. The contingent measurement framework

of a simple flowchart. Two feedback loops are employed for the input of existing measures either back into the kaizen starts/restarts activity box or the identify measures process box. A brief summary of the actions and activities associated with each entity in the flowchart will clarify the mechanism: Identify measures. Information of customer requirements, cell problems and areas of waste and information arising from the review of any existing measures all contribute to an initial definition of measures with associated goals and timescales. Kaizen starts/restarts. Improvement activity is triggered by new measures. Review measures. A periodic activity is needed to ensure the continued application of relevant and rigorous measures. Input to the activity will be either valid measures () or refined measures (). The output from this activity will feed into the kaizen starts/restarts activity, if the measures are found to be valid, or into the identify measures activity, if some amendment is considered necessary by the team.

Education and training. The application of team training with customer A framework for input is important in order to improve customer and cell awareness in the performance team and to facilitate the identification of valid measures and goals. measurement Refine measures. This process differs from the review measures process in that it is concerned with the refinement of sub-optimal measures initially chosen by the team. Input from the education and 113 training process will serve to enrich the process and increase the chances of a valid measure or set of measures being defined for input to the review measures process.

Using invalid measures to drive kaizen The contingent measurement framework includes two OK/NOK decision points (indicated by ticks and crosses in Figure 7) where the validity of measures is assessed. The first occurs at the identify measures process and the second at the review measures process. The case of the invalid (or NOK) measures driving the kaizen starts/restarts activity requires clarification. This situation will occur only once, at the inception of the system. At this point it is likely that partially (or totally) invalid measures will be selected by the team, especially where appreciation of customer requirements is poor. It is preferable, however, to allow these measures to begin to drive kaizen while education and training is carried out and prior to production of the measure set. There are two main reasons for this: (1) In order to continue the motivation of the team it is preferable to maintain some momentum in the process and not to introduce a pause in progress while the imperfections of the measures chosen by the team are analysed. (2) Some measures are better than no measures at all and, particularly in cases, where previously there have been no drivers for kaizen, it is infinitely preferable to demonstrate some practical improvements in the cell. Team members will usually be keen to start some practical kaizen and, in the early stages, encouraging this without provisos is important for morale and facilitator credibility as well as for the learning process. In the early stages it is more important to encourage the generation of ideas, the growth of confidence, enthusiasm and pride in the cell and the improvements carried out than it is to insist on absolutely valid measures and resulting activities. It is preferable, therefore, in the early stages to allow team members to make suboptimal decisions and carry out sub-optimal activities than it is to allow only those activities which the cell leader considers to be valid. For the cell leader to select unofficially the team measures in this way undermines the whole rationale of the approach and represents the kind of lip-service paid to the concept which will be immediately obvious to the team.

IJOPM 17,1

114

The choice of initial project The findings of the research in relation to the choice of initial project reflect, in many ways, the guidelines offered to TQM practitioners in the 1980s and early 1990s[2]: The initial project (and indeed all projects) should be as small and as focused as is possible. The initial project should have a high probability of success according to the parameters defined by the team. The results of the project should be widely publicized. Recognition drives commitment. The initial project should be quickly followed by a second to maintain momentum. Role of the cell leader As mentioned previously, the role of the cell leader, supervisor or line manager is central to the success of a system such as that described. While it is not possible to provide an exhaustive list of required capabilities and personality traits, the following are valuable: When offering guidance and support the cell leader is able to strike a balance between suggestion and direction. A good knowledge of the personalities of team members and the ability to maximize the contribution of each member are important. Effective liaison with other cells and supply functions is important where the teams activities interface with external areas. In an organization where a traditional finance-driven performance measurement system is in place, with all the behavioural consequences which this may have, the cell leader is able to act as an umbrella or shield for his cell. In this way drives to increase productivity (seemingly in isolation) are translated by the cell leader into a form relevant to the measures of the cell and the functioning of the team. In an organization that employs a customer-driven performance measurement system in which goal congruence is high, the need for this characteristic is reduced. As the team develops, the cell leader is able to yield an increasing degree of control and responsibility to the team. The more mature the team becomes, the less need it will have of a guide or a parent figure. Implications for further research The research presented here suggests that through the development of the kaizen team which is focused on contingent and relevant measures it is possible to advance the concept of self-managing teamwork in a production cell[21] and that this will lead to beneficial shifts in culture. What also has been advocated

is the need for a degree of control in the cell-generated system. As Mills[22] A framework for states: Much of the conceptual and empirical literature has focused on the performance relationship between external control mechanisms and organisation structure measurement (see also [23,24]). Slocum and Sims[25] have observed that there is a relative dearth of models specifically addressing self-control in organizations, and even less related to kaizen teams. The research documented here indicates such a 115 framework which has been empirically tested and which points to several areas where extensive further research is required not least the changing role of first-line management. It may be argued that the role of the cell leader is a prime focus in the development of a world-class organization. Performance measurement research certainly has a key part to play in the advancement of such companies, and empirical research and trials which marry up strategy formulation, information provision and control to the behavioural implications of performance measurement continue.
References 1. Voss, C.A., Manufacturing Strategy Process and Content, Chapman & Hall, London, 1992, pp. 4-6. 2. Oakland, J.S., Total Quality Management, Butterworth-Heinemann, Oxford, 1991, p. 27. 3. Daniels, R.C., Performance measurement at the sharp end driving continuous improvement on the shopfloor, Engineering Management Journal, Vol. 5 No. 5, 1995, pp. 211-14. 4. Daniels, R.C., The behavioural consequences of performance measurement in a cellular manufacturing environment: a case study, PhD Thesis, Loughborough University of Technology, 1995. 5. Tomorrows Company, The Royal Society for the encouragement of Arts, Manufactures and Commerce, London, 1995. 6. Daniels, R.C. and Burns, N.D., The design and operation of a cell-generated performance measurement system Proceedings of the 11th National Conference on Manufacturing Research, De Montfort University, Leicester, 1995, pp. 613-17. 7. Wickens, P., The Road to Nissan, Macmillan, London, 1992, pp. 162-5. 8. Trought, B., When are manufacturing managers necessary?, Proceedings of the 1st International Conference of the European Operations Management Association, Cambridge, 1994, pp. 517-18. 9. Daniels, R.C., The behavioural consequences of performance measures, International Journal of Operations & Production Management, submitted June 1996. 10. Buchanan, D. and Huczynski, A.A., Organisational Behaviour, Prentice-Hall, London, 1985, pp. 353-5. 11. Bennis, W.G., Leadership theory and administrative behaviour: the problem of authority, Administrative Science Quarterly, Vol. 4 No. 3, December 1959. 12. Fayol, H., General and Industrial Management, 1916, translated from the French by Storrs, C., Sir Isaac Pitman and Sons, London, 1949, pp. 12-13. 13. Lupton, T., Management and Social Sciences, Penguin Books, London, 1971, pp. 21-3. 14. Mintzberg, H., On Management, Prentice-Hall, London, 1987, pp. 34-7. 15. Daniels, R.C., Backhouse, C.J. and Burns, N.D., A contingent approach to the new product introduction process: two comparative cases, Proceedings of the 3rd International Conference on Concurrent Engineering and Electronic Design Automation, Poole, 1996, pp. 397-403.

IJOPM 17,1

116

16. Bhimani, A., Performance measures in UK manufacturing companies: the state of play, Management Accounting, December 1993, pp. 20-21. 17. Thompson, J.L., Strategy in Action, Chapman & Hall, London, 1995, pp. 131-55. 18. Slack, N., Chambers, S., Harland, C., Harrison, A. and Johnston, R., Operations Management, Pitman, London, 1995, pp. 529-33. 19. Beischel, M. and Smith, R., Linking the shopfloor with the top floor, Management Accounting, October 1991, pp. 25-36. 20. Peters, R.J. and Waterman, T., In Search of Excellence, McGraw-Hill, New York, NY, 1987, pp. 56-7. 21. Glaser, R., Classic Readings in Self-managing Teamwork, ODD, Pennsylvania, PA, 1992, pp. ix-xii. 22. Mills, P.K., Self-management: its control and relationship to other organisational properties, Classic Readings in Self-Managing Teamwork, ODD, Pennsylvania, PA, 1992, pp. 259-74. 23. Blau, P. and Schoenherr, R., The Structure of Organisations, Basic Books, New York, NY, 1971, pp. 12-14. 24. Ouchi, W., The relationship between organisational structure and organisational control, Administrative Science Quarterly, Vol. 22, 1977, pp. 95-112. 25. Slocum, J. and Sims, H., A typology for integrating technology, organisation and job design, Human Relations, Vol. 33, 1980, pp. 193-212.

You might also like