تأثیر متقابل اهرم های مختلف کنترل : یک مطالعه موردی از معرفی یک سیستم ارزیابی عملکرد

In this paper, different notions of control are investigated in the context of a longitudinal field study dealing with the introduction and use of a new performance measurement system at one case company. The control framework of Simons [Simons, R., 1995a. Control in an age of empowerment. Harvard Bus. Rev., 67(2), 80–88; Simons, R., 1995b. Levers of Control: How Managers Use Innovative Control Systems to Drive Strategic Renewal. Harvard Business School Press] is used as a theoretical frame of reference. In earlier studies, differences between diagnostic control and interactive control have been frequently addressed [e.g. Abernethy, M.A., Brownell, P., 1999. The role of budgets in organizations facing strategic change: an exploratory study. Account. Organ. Soc., 24, 189–204; Bisbe, J., Otley, D., 2004. The effects of the interactive use of management control systems on product innovation. Account. Organ. Soc., 29, 709–737; Vaivio, J., 2004. Mobilizing local knowledge with ‘provocative’ non-financial measures. Eur. Account. Rev., 13, 39–71]. In our paper, it is shown that strategic performance measurement systems can be used both diagnostically and interactively, but such systems have implications for beliefs control and boundary control as well. Interactive use of performance measures is apt to improve the quality of strategic management and to increase commitment to strategic targets. On the other hand, interactive discussion of specific performance metrics increases the visibility of actions which may initiate resistance. In addition, interactive use of performance measures may be costly in terms of time consumption both when collecting the data and when discussing the results. Two major differences were found in the actual use of strategic performance measures when compared to the normative literature. First, in contrast to ascertaining certain cause-and-effect relationships before implementing new measures, it was perceived that the measures themselves would be used over time to confirm or reject alleged relationships. Second, no tight connections between the new measurement system and managerial bonuses were made. This was mostly due to the development process, during which the top managers themselves developed the measures to reflect their belief about the best way of achieving the ultimate financial targets.

Discussion of strategy, management accounting systems and their relationship has intensified after the introduction of strategic performance measurement systems like the Balanced Scorecard (Kaplan and Norton, 1992, Kaplan and Norton, 1993, Kaplan and Norton, 1996a and Kaplan and Norton, 1996b) and Performance Pyramid (Cross and Lynch, 1989, McNair et al., 1990 and Lynch and Cross, 1991). But even though these well-promoted strategic control systems have rapidly diffused to several (big) companies (Ittner and Larcker, 1998 and Malmi, 2001), there are numerous hints of potential problems and challenges. A number of case studies have pointed out the persistence of management accounting systems (e.g. Granlund, 1998, Granlund, 2001 and Burns, 1996; cf. Burns and Scapens, 2000). It might not be that easy to engage in using a new performance measurement model, since several controllable or uncontrollable factors can act as barriers to full implementation of such systems (Kasurinen, 1999 and Kasurinen, 2002). In addition, Ittner and Larcker (1997) have suggested that the benefits from new performance measurement systems could be outweighed by increased bureaucracy. Implementation of complex measurement systems is costly and evidence on economic benefits of these systems is thus far limited. Detailed approaches to strategic performance measurement systems could induce inflexibility into strategic thinking (see Mintzberg, 1987 and Hamel, 2000). According to the normative literature, the construction of strategic performance measurement systems starts from the (vision and) strategy of the organization in question (Kaplan and Norton, 1993). But even the objectives behind the strategy are far from clear-cut. Different coalitions or stakeholder groups may dominate in goal setting (Ezzamel and Hart, 1987). The ultimate objectives of a company can be associated with, for example, creating shareholder value (Rappaport, 1998) but has also been argued that a company should satisfy the needs of all relevant stakeholders (Emmanuel et al., 1990). The Balanced Scorecard, for instance, has been perceived both as a tool for creating shareholder value (Kaplan and Norton, 1996a) and as a stakeholder model (Otley, 1999 and Ax and Bjornenak, 2000). In addition, the objectives of the performance measurement system itself can be unclear. Contradictory views of control purposes may lead to a rejection of a proposed new strategic control system (Kasurinen, 1999). The Balanced Scorecard is typically presented as a tool for introducing a new strategy to a business unit, and hence, it is logical that the Balanced Scorecard should be constructed in accordance with the new strategy. It has also been claimed that one major benefit from the Balanced Scorecard stems from the construction phase as it indeed helps in specifying the strategy (e.g. Kaplan and Norton, 1996a, Epstein and Manzoni, 1997 and Tuomela, 2000). While it is important to explicate strategy-based assumptions of means-end relationships, mapping such chains of events is by no means simple (Otley, 1999, Norreklit, 2000, Norreklit, 2003 and Wenisch, 2004). As a matter of fact, recent research suggests that a seemingly clear and uniform strategy can turn out to be much more complicated during the development of a new performance measurement system and consequently undermine the entire project (Kasurinen, 1999). Since it is likely that a realized strategy is a combination of both intended and emergent ingredients (Mintzberg, 1978), capturing the essence of strategy to a performance measurement system is quite a challenge (cf. Lipe and Salterio, 2002). Given the fuzziness of goals and problems with defining and implementing a successful strategy, one might claim that strategic performance management has more to do with randomness and retrospective sense-making than with rational decision making. Based on the work of March (1971) and March and Olsen (1976), Cooper et al. (1981) use the garbage-can metaphor and the concept of “technology of foolishness” to illustrate the potential and actual role of accounting systems in, what they call, organized anarchies. The garbage-can view postulates that organizational action and decision-making is a mixture of problems, solutions, participants and choice opportunities that come together, often by chance. Decisions are often based on imitation and coercion. The goals may be unclear when making decisions and they are uncovered only after ex post rationalization. Instead of consistency, technology of foolishness prefers creativity and playfulness in the decision-making process (Cooper et al., 1981). But even if one adopts a more traditional or rational stance, a multitude of challenges are encountered. Several authors have, for instance, suggested that strategic performance measurement systems should be altered if the strategy is changed (e.g. Eccles, 1991, Sellenheim, 1991 and Grady, 1991). Giving credit to the continuously changing environment and emergent strategies, it is likely that strategic performance measurement systems should be refined quite often (Otley, 1999). This is, however, problematic from the measurement continuity point of view, and it also raises questions like how and when should strategic performance measurement systems be changed. Lack of proper mechanisms for improving and updating the measurement system is one potential problem with regard to strategic performance measurement systems (Anthony and Govindarajan, 1998; see also Kasurinen, 1999). Even if the goals and strategies can be agreed upon, defining the measures, setting appropriate targets and altering reward systems can be problematic. The use of non-financial measures poses special challenges. Non-existent or short measurement tradition reduces confidence in non-financial measures. Lacking knowledge of appropriate performance levels may lead to too low or high performance targets with decreased motivation (Vaivio, 1995). It may also be difficult to estimate the tradeoffs between non-financial measures (Anthony and Govindarajan, 1998, Otley, 1999 and Fisher, 1992). While strategic performance measurement systems tend to balance financial and non-financial measures, the idea is typically that non-financial measures are leading indicators of financial performance (Kaplan and Norton, 1996a). The problem is that it is difficult to explicate the impact of particular non-financial measures to financial results (Anthony and Govindarajan, 1998). In the short term, it might be that results from financial and non-financial measures contradict (Fisher, 1992 and Ittner and Larcker, 1998). While lack of measurement history and poor understanding of tradeoffs make it difficult to set targets for non-financial measures and to change incentive systems, it has been claimed that measuring non-financial parameters is meaningless if rewards are not attributed to them (Eccles, 1991). But in addition to difficulties in determining challenging target levels, it should be acknowledged that non-financial measures are not free from gaming possibilities either1 (Fisher, 1992).

This study investigated the role of strategic performance measurement systems in respect of the interplay between different control levers. The 3K Scorecard was used for both diagnostic and interactive control purposes at the case company. Furthermore, the 3K Scorecard had specific implications for both beliefs systems and boundary systems (see Table 2). In the course of the development process, the main objective of the performance measurement system evolved from communicating customer focus (i.e. support to beliefs system) via strict strategic control (i.e. diagnostic control) finally to making sense of strategy and learning about strategic interdependencies (i.e. interactive control). The emphasis on an interactive (rather than diagnostic) use of the 3K Scorecard for strategic learning purposes has several implications. Already in the construction phase, managers frequently praised how enlightening the measurement team meetings were, giving strategy documents a specific content and nurturing intellectual discussion of underlying cause-and-effect relationships. When using the scorecard, it is possible to look at trends and evaluate the validity of anticipated relationships. This in turn accentuates the need to formulate and describe the measures in a way that the assumed cause-and-effect relationships are visible, and hence, it is possible to evaluate or question the validity of those assumptions. Our findings strongly support the use of strategic maps ( Kaplan and Norton, 2000 and Kaplan and Norton, 2001) and the definition of at least tentative cause-and-effect relationships as presented in other recent case studies (e.g. Wenisch, 2004). Two of the most intriguing findings in our case study relate to normative prescriptions that are commonly associated with the use of non-financial measures in such systems as the Balanced Scorecard, namely: (1) ex ante cause-and-effect verification; and (2) tight connections to managerial bonus systems. In our case study neither of these prescriptions held which could be explained by the bottom-up development process and interactive use of the measures. First, the connection between different measurement perspectives and two specific cause-and-effect relationship paths were not tested before the system was implemented. These relationships were based solely on managerial reasoning. But, ambiguity concerning the balance between different non-financial measures and the links between non-financial and financial measures was not considered to be a measurement problem per se. Rather, this was exactly the reason for using a balanced performance measurement system – i.e. to learn about potential cause-and-effect relationships, and the relative importance of different measures. Similarly, target setting was perceived as a challenge, and any meaningful target setting was only expected to follow the use of the system. Having said this, tentative target levels were set initially. Second, at the end of our research period only minimal changes had been made to the managers’ remuneration systems. Some non-financial aspects had been added to bonus contracts, but this was not necessarily directly linked to the use of the 3K Scorecard. One important explanation for this stems from the development process and the way in which the system was used. From the perspective of global ABB, managers of FinABB could be considered as middle management. The impetus for adopting a new performance measurement system did not come top-down, but rather it was FinABB's management's own idea to develop a scorecard that would assist them to improve strategic management. Thus, there was consistent belief that the new system would eventually lead to improved financial performance and there was strong commitment towards using the system (perhaps with the exception of the Domestic Sales Manager). In addition, the focus was on discussion and learning – interactive use – and there was no interest in creating a reward or punishment atmosphere. Finally, it was perceived that reward systems should not be tightly linked to the bonus systems when ambiguity about measures and appropriate target levels prevails. But, as suggested above, such a connection might be less important if the managers have themselves developed the system for their own learning purposes. While interactive use of performance measures sheds a little different light on performance management, it can also be problematized. At least two problems remain with performance measurement and possibly even worsen when strategic performance measurement systems are primarily used for interactive controlling. First, the introduction of new non-financial performance measures may initiate rather strong resistance. Non-financial measures improve the visibility of actions. Since interactive control initiates discussions about strategic uncertainties, it is likely that more information about subordinates’ and peers’ actions is disclosed when compared to diagnostic control. This is likely to intensify resistance to change. The introduction of new non-financial measures is also likely to disrupt the power structures within an organization. An in-depth knowledge of customers, for example, implies an informal power dimension which might result in those who possess such knowledge being reluctant to share it (Vaivio, 2001 and Tuomela, 2001; see also Markus and Pfeffer, 1983). The additional time spent on data-gathering and the actual use of information creates further problems. In relation to the interactive use of strategic performance information, middle managers and functional managers were instructed to collect some of the data required for the 3K Scorecard. While these reporting requirements were perceived as an important means of communication, it added to these managers’ workload. In addition, the business controllers at FinABB saw their workloads increase as a result of the 3K Scorecard – partly due to a need to continually remind and instruct middle managers and functional managers about their new reporting responsibilities. An interactive control, with its emphasis on discussion, also increased the number and length of meetings, consuming much managerial time. To conclude, the way in which strategic performance measurement systems are used has several implications for the benefits and problems related to its use. This finding is consistent with earlier results (Langfield-Smith, 1997, Abernethy and Brownell, 1999, Vaivio, 2001 and Bisbe and Otley, 2004) that it is not only the specific control tools (like the Balanced Scorecard) that are used but also the way they are applied that should be taken into account. Moreover, it should be taken into account that performance measurement systems have implications for all levers of controls and that the interactive use of performance management systems has some special benefits and challenges when compared to diagnostic controlling.