PERFORMANCE MYOPIA: THE EFFECT OF PAY-FOR- PERFORMANCE INCENTIVES ON EXPLORATION AND COORDINATION

Incentives are one of the core practices of strategic human resource management and their effect on motivation and performance has been studied extensively. Particular attention is devoted to pay-for-performance (PFP) incentives while research on the effect of PFP incentives on performance has produced contradictory results. More importantly, incentives are not an isolated process; the effect of incentives goes beyond individual or organisational performance because they affect an entire organisation. Although incentives are included in most organisational design frameworks, the effect of incentives on other organisational design components has been neglected. The study uses the organisational design framework to focus on neglected relations between incentives and other organisational design components. The purpose of the study is to explore what are the organisational design components and how they are influenced by PFP incentives. A  case study research design was used. The data was collected in a  small company in which the incentive system was changed to PFP incentives as a  part of substantial changes in the organisational design. Data was collected through interviews with employees, supported by internal documentation and observation. Thematic analysis was used to analyse the data. In the case, the PFP incentives led to higher performance although the PFP incentives restricted the new exploratory strategy and harmed cooperation. The effect of the PFP incentives on exploration and cooperation was slow, and hardly visible, predominantly as a  result of unintentionally deviated attention. The study points out that focusing solely on performance when designing incentive systems may be myopic because PFP incentives may have a detrimental effect on other organisational design components. Based on the results, the paper provides a  set of suggestions to consider when implementing PFP incentives.


Introduction
Incentives are one of the core practices of strategic human resource management, so it is not surprising that the effects of incentives have been researched for more than 40 years in human resource management and in personnel psychology (Cerasoli, Nicklin and Ford, 2014). Most of the research has focused on the effect of pay-for-performance (PFP) incentives on motivation and individual performance (Cameron and Pierce, 1994;Deci, Koestner and Ryan, 1999;Cameron, Banko and Pierce, 2001). Although hundreds of studies have been conducted, the findings are contradictory. Many studies found a positive effect of PFP incentives on performance (Delaney and Huselid, 1996;Jenkins, Mitra, Gupta and Shaw, 1998;Cameron et al., 2001;Akhtar, Ding and Ge, 2008;Alfes, Truss, Soane, Rees and Gatenby, 2013) while many other studies concluded a negative or no effect of PFP incentives on performance (Kohn, 1993;Pfeffer, 1998;Deci et al., 1999;Cappelli and Neumark, 2001;Frank and Obloj, 2014;). However, incentives are not an isolated tool or a process in an organisation; an incentive system should be integrated into the entire human resource management system (Wright and Boswell, 2002), which is part of the overall organisational strategy (Boxall, 1996;Boxall and Purcell, 2000). Thus, focusing solely on performance consequences may be myopic.
To study the wider relationships of incentives, this paper uses the organisational design framework of Burton, Obel and Håkonsson (2015), which views an organisation through a model of interdependent components such as strategy, structure, processes and people, coordination, and incentives. Performance of an organisation is dependent on the fit or alignment within the organisational design components. Any misfit or misalignment within the components results in lower performance than could be obtained otherwise (Donaldson, 2001;Donaldson and Joffe, 2014;Burton et al., 2015). According to Burton, Obel and Håkonsson (2015) organisational design framework, the PFP incentive system is assumed to fit in with exploration strategy, divisional organisational structure, and coordination based on high decentralisation and delegation. In comparison with human resource management and personnel psychology, significantly less attention has been paid to the relations of incentives with other organisational design components. Therefore, this article contributes to this area by identifying the neglected relations of incentives.
The purpose of this article is to explain the effect of changing incentives on other organisational components using the perspective of organisational design. In particular, the article poses the question: how do the changing incentives to a PFP system affect the other organisational design components? The article proceeds as follows. First, we briefly summarise the results of previous studies on the effect of PFP incentives on individual and organisational performance. Second, we explain the role of incentives in the organisational design framework and the suggested alignment of incentives with other organisational design components. Third, we present a particular case to demonstrate the effect of PFP incentives on exploration strategy and coordination. Based on the results, we provide a set of suggestions to consider before implementing PFP incentive systems.

Effects of PFP Incentives on Individual Performance
There is a vast body of research on the effect of incentives on individuals. The main body of research is focused on the relation between incentives and individual performance. Searching for the effect of PFP incentives on individual motivation or performance may be deceptive. If managers want to support their decision on incentives with research results, they can probably find solid evidence for the particular decision they want to implement. There is strong evidence both for the positive or negative effect of incentives on individual performance. In addition, some studies did not find any significant effect of incentives on individual performance (Cappelli and Neumark, 2001).
Traditional economic theory suggests that human behaviour is rational and that people maximise their utility. To do so, people need extrinsic motivators, and thus PFP incentives improve performance. According to this stream of literature, PFP incentives affect performance via such mechanisms as instrumentality, which is based on direct link between performance and pay (Baker, Jensen and Murphy, 1988;Jenkins, Mitra, Gupta and Shaw, 1998;Cadsby, Song and Tapon, 2007) and the importance of pay-forperformance incentives for motivation (Rynes, Gerhart and Minette, 2004;Garbers and Konradt, 2014), or setting and achieving difficult goals (Presslee, Vance and Webb, 2013). This traditional point of view is still topical and even the latest studies in leading management journals have unearthed new evidence that PFP incentives are positively associated with future performance (Maltarich, Nyberg, Reilly, Abdulsalam and Martin, 2017). These arguments are also strongly rooted in society, and PFP incentives are used in the majority of US companies (Rynes, Gerhart and Parks, 2005). PFP incentives generally work for quantitatively measurable performance and for situations where incentives may be directly tied to performance (Lazear, 2000).
On the other hand, behaviourists, psychologists and human resource specialists often challenge the traditional economic view that the higher the incentives then the higher the performance. Deci, Koestner and Ryan (1999) conducted a metaanalysis of 128 studies and examined the effect of extrinsic rewards on intrinsic motivations. They revealed that all types of rewards significantly undermine intrinsic motivation. This finding supports the self-determination theory (Ryan and Deci, 2000;Gagné and Deci, 2005), which emphasises the role of intrinsic motivation over external motivation and incentives. However, Cameron, Banko and Pierce (2001) conducted another meta-study and identified the negative effects of extrinsic rewards on intrinsic motivation only under a narrow set of circumstances. Only tangible and rewards offered beforehand, which were loosely tied to the level of performance, had a negative effect on tasks that were perceived as interesting.
Recent research aims to integrate economic and psychology-based explanations. For example, Maltarich et al. (2017) studied the moderating role of expectations on the effect of PFP incentives on individual performance. They concluded that the effect of PFP incentives on performance was moderated by accumulated underperformance in a PFP incentive system that was carried forward to future periods. The effect of PFP incentives on performance was negative for employees whose underperformance in the past had a negative impact on their future rewards. This effect could be eliminated by providing exceptions for underperforming employees. Setting easier goals and slowly increasing performance goals leads to improvement in performance as opposed to a decrease in the performance of frustrated employees if there are no exceptions.
PFP incentives also bring various risks. First, the implementation of PFP incentives can cause deviation from the attention to the outcomes towards the attention to reward objects while emphasising monetary rewards may lead to the desire for money rather than the desire for meeting goals (Hur and Nordgren, 2016). This deviation also hinders employees' intrinsic motivation because performance decreases significantly when the bonuses are removed later. Furthermore, employees in a PFP incentive system tend to demand higher salaries in subsequent negotiations (Cappelli and Neumark, 2001). The second risk is linked to changes in behaviour. There is both experimental and field research evidence that PFP incentives may support opportunistic behaviour (Burton and Obel, 1988;Håkonsson, Obel, Eskildsen and Burton, 2016). The desired outcomes can sometimes be reached by superior performance but also by unethical or even illegal behaviour, such as withholding, manipulating, or misrepresenting information (Frank and Obloj, 2014). The third risk is in determining an effective level of PFP incentives because paying too much or too little reduce performance (Kamenica, 2012). Finally, risk-averse individuals tend to choose positions without PFP incentive schemes (Cadsby, Song and Tapon, 2007) and thus, organisations with strong PFP incentive schemes may predominantly attract risk-takers or greedy individuals.

Effects of PFP Incentives on Organisational Performance
In a similar way to the individual level of analysis, research on the effects of incentives on the organisational level of analysis has focused primarily on organisational performance. Similarly, the research yields ambiguous results. To illustrate, we contrast the results from two studies conducted on a national survey in the US, both of which focused on the effects of high-performance human resource management practices. The first study concluded that PFP incentives had a positive effect on perceptual measures of organisational performance (Delaney and Huselid, 1996). The second study was based on data collected three years later and the study found that PFP incentives have no effect on labour efficiency because the increased performance was outweighed by raised labour costs per employee (Cappelli and Neumark, 2001).
The mechanism between incentives and organisational performance is sometimes perceived as a black box, hence some studies focused on uncovering the underlying mechanisms. Researchers found, for example, an important role of line managers in the implementation of PFP incentive systems by creating and maintaining a positive environment. In a positive environment, employees are more engaged and as a result, they behave and perform better (Alfes et al., 2013). Furthermore, the design of a PFP incentive system should be perceived as equitable by employees (Shaw, 2015). Equitable incentives lead to greater trust in management, to higher employee commitment, and greater job satisfaction, which results in higher performance (Whitener, 2001). Based on an analysis of numerous studies, it is apparent that there is a thin line between facilitative and disruptive effects of incentives on performance (McGraw, 2016). One of the ways to better understanding is to perceive incentives as a part of a complex system of management practices, not as an isolated tool or process. And this perspective is employed by the organisational design framework.

The Role of Incentives in Organisational Design
We applied the organisational design framework suggested by Burton et al. (2015). The framework provides a holistic approach and views an organisation as a set of interrelated components. A key notion in organisational design is to reach a "fit", or "alignment" among the organisational components. The organisational components can be differentiated as structural and human. Structural components comprise goals, strategy, and structure. Human components comprise processes, people, coordination, control and incentive systems. Research shows that about 30% of organisational performance can be explained by organisational design, i.e. by a fit or misfit among the organisational components (Burton et al., 2015). Any misfit or misalignment between two or more of the components may be detrimental to an organisation and inhibit organisational performance and a misfit is a serious reason for redesign. In addition, when changing any of the components, the fit with the other components should be assessed because a single change may bring misfits with all other components. Thus, each component of the organisational design framework is mapped onto a twodimensional graph with four quadrants for four possible options, and a fit is reached when the organisational component options are in the same quadrant.
Incentives are "means or instruments designed to encourage certain actions or behaviour on the part of employees or group of employees" (Burton et al., 2015, p. 217). Organisational design framework divides incentives according to two dimensions (Gerhart, Rynes and Fulmer, 2009;Burton et al., 2015), which are commonly used for classifying an incentive system (Gerhart, Rynes and Fulmer, 2009). The first dimension differentiates between rewarding behaviour or results. Behavioural incentives focus on how employees comply with standards, rules and routines. The results incentives focus on outcomes and meeting planned goals. An important aspect of the decision on this dimension is the feasibility of monitoring the performance because information on the performance may be difficult or costly to acquire. The second dimension differentiates between rewarding individuals or groups. The combinations of these two dimensions lead to four possible distinct types of incentive systems.
The first type of incentive system combines behavioural and individual-based incentives and is labelled as personal pay. Employees usually have fixed pay and are expected to follow the rules and do things as requested. The pay level tends to be lower than in other incentive systems. The second incentive system type combines behavioural and group-based incentives and is labelled as skill pay. Rewards are based on a structure of grades, which are organised into a system. The system is based on seniority, skills, education, and other variables that should predict the ability to perform tasks assigned to the positions. The third incentive system type combines results and individual-based incentives and is labelled as bonus-based. This type is enrooted in management-byobjective philosophy and employees are rewarded based on their individual performance. With the exception of some jobs (e.g. sales representatives), the bonus is in addition to personal or skill pay. Precise information on individual performance and a clear link between the performance and the outcome are elementary conditions for the bonusbased type of incentives. The fourth incentive system type combines results and groupbased incentives and is labelled as profit-sharing or gain-sharing incentives. This type is appropriate if a task depends on the joint effort of a team, division, department, or other sub-unit, and there is a certain level of interdependency among the work of individuals. Each incentive system type should be a fit or alignment with other organisational design components. Both bonus-based and profit-sharing incentive system types are representatives for PFP incentives, so next, we will specifically focus on the alignment of bonus-based and profit-sharing incentive system types with strategy and coordination. The overview of the alignment according to the organisational design framework is shown in Figure 1.

Alignment of Incentives with Exploration Strategy
The organisational design framework applies Miles and Snow typology of strategy (Miles, Snow, Meyer and Coleman, 1978), which has proven to be an effective tool in the strategic management and organisational theory domains (Hambrick, 2003). Again, the organisational design framework maps strategy onto a two-dimensional graph with four quadrants along two dimensions: exploitation and exploration (Burton et al., 2015). PFP incentives are in alignment with strategies based on high exploration, either in combination with high exploitation or low exploitation. Exploration embodies innovation, searching new opportunities, and exploring new possibilities, while exploitation is based on efficient use of current resources and exploiting current capabilities (March, 1991).
The alignment of PFP incentives with exploration seems to go against several assumptions about PFP incentives and exploration. First, there are serious concerns that for tasks involving creativity and innovation, PFP incentives do not work (Ryan and Deci, 2000;Gagné and Deci, 2005;Cerasoli, Nicklin and Ford, 2014). Second, PFP incentives may be perceived as an exploitative payment system (Ogbonnaya, Daniels and Nielsen, 2017), which is not in coherence with the organisational design framework. Third, exploration requires learning and it is difficult to incentivise learning, especially for those who learn slowly (March, 1991). These arguments were empirically tested with the conclusion that fostering PFP incentives leads to higher exploitation while weakening PFP incentives leads to higher exploration, particularly by high-performing individuals (Lee and Meyer-Doyle, 2017). Such a contradiction provides the grounds for this study.
The two most recent studies explored the conditions under which PFP incentives may foster exploration. First, the experimental study of Ederer and Manso (2013) tested an exploration PFP incentive scheme. The main feature of the exploration PFP incentive scheme is a combination of tolerance for early failures and the prospect of pay-forperformance after a learning period. The exploration PFP incentive scheme encourages employees to learn better ways of performing the task in the first period and they perform better later. The exploration PFP incentive scheme produced better results than fixed pay or standard PFP incentives. Second, a simulation of Baumann and Stieglitz (2014) used a computational model to test the effect of incentives on generating employee innovative projects. They used two types of incentives in their simulation: high-power incentives were tied closely to the value created, and low-power incentives were tied loosely to the value created. High-power incentives had dysfunctional effects on employee behaviour and motivation and thus did not foster exploration. However, low-power incentives proved to be an effective approach to induce exploration. Despite low-power incentives producing a lower number of innovative projects, the quality of the project was significantly higher. Baumann and Stieglitz additionally recommend reinforcing explorative behaviour by incentives provided later rather than agreed beforehand.

Alignment of Incentives with Coordination
Coordination is generally defined as "managing dependencies among activities" (Malone and Crowston, 1994). Coordination supports the integration of the organisational tasks, links together various sub-units of an organisation, and provides pathways for information sharing (Burton et al., 2015). Analogically to the previous sections, PFP incentives are in alignment with the right side of the model in Figure 1, which is characterised by a high degree of decentralisation and differs in the level of formalisation. The market model emphasises low formalisation and coordination relies on informal communication, information sharing culture and cooperation. The clan or mosaic model is generally based on a set of norms and values that underlines how organisational goals should be accomplished, yet there is some degree of freedom for employees.
There is surprisingly little research on the effect of PFP incentives on coordination, information sharing, or cooperative behaviour. An experiment in the US and Switzerland showed that employees under PFP schemes are significantly less cooperative (Burks, Carpenter and Goette, 2009). A certain role may be attributed to the equity of the incentives because individuals who are under-rewarded tend to behave less cooperatively and more selfishly (Harder, 1992). Another possibly harmful effect is connected to personality; PFP may increase active harmful behaviour toward co-workers, which is particularly distinctive for competitive employees with a strong desire for interpersonal comparison and high aspiration (Gläser, van Gils and Van Quaquebeke, 2017). Although the evidence is not vast, the research results indicate that PFP incentives have a high potential to harm cooperation.
Only a few studies focused on eliminating the risk of harming cooperation. For example, a job designer may lower interdependencies between employees by decoupling a task from employee interdependence, which will result in lower information processing needs (Puranam, Raveendran and Knudsen, 2012). However, this situation seems to be more suitable for exploitation, while exploitation or ambidexterity -i.e., pursuing both exploration and exploitation (Gibson and Birkinshaw, 2004;O'Reilly and Tushman, 2004; 2013)requires a higher degree of coordination. Another way to avoid the risk is the implementation of profit-sharing or gain-sharing incentives. Based on traditional economic theory, there are concerns about paying on a group basis because of the so-called free-rider problem. The free-rider problem posits that people do not work hard because they will share the rewards regardless of their effort. However, empirical evidence indicates that the occurrence of the free-rider problem is less common than theoretically predicted due to peer pressure and social relations with other workmates, especially in small face-to-face groups (Wageman and Baker, 1997

Methods
To be able to gain an insight into the complex situation where human behaviour plays an important role, a qualitative approach was chosen to gain an understanding of the dynamics of the researched contingencies within the organisational context (Bartunek, 2012;Ketokivi and Choi, 2014). A case study research design was chosen because of the exploratory nature of the research and to obtain an insight into the hidden contingencies among incentives, coordination, strategy, and organisation performance (Eisenhardt, 1989;Yin, 1994;Baxter and Jack, 2008;Myers, 2013). Case studies are often recommended as an appropriate research design for studying organisational design and change (Pettigrew, Woodman and Cameron, 2001;Pettigrew, 2013) and particularly for understanding relations among different organisation components within the entire organisation (Fiss, 2009).
Purposeful sampling is appropriate for pilot studies, extensive case studies, critical case studies or for studies of hard-to-find populations (Eisenhardt and Graebner, 2007;Bernard, 2011). The research was conducted in a small organisation providing recruiting services and there were two inclusion criteria as a necessary condition of purposeful sampling (Eisenhardt, 1989;Eisenhardt and Graebner, 2007;Bernard, 2011;Robinson, 2014). First, the organisation had changed the organisation structure as a basis for the new organisation design. Second, the organisation had changed the incentive system as a part of the new organisation design. Based on the inclusion criteria with the combination of the willingness of the organisation to share information on internal constitutions, the case study represents the hard-to-find population, which was strengthened by conducting interviews with both the two owners, with all (n=5) employees and with an external consultant hired to help with the new organisational design.
The quality of case studies is a frequently discussed issue (Eisenhardt, 1989;Meyer, 2001;Siggelkow, 2007;Grbich, 2013). The quality relates to the validity and reliability of the research. Validity in this research is strengthened by the use of the triangulation method, which was ensured by three possible approaches (Denzin, 1978;Creswell and Miller, 2000;Thurmond, 2001;Remenyi, 2012;Yin, 2013): First, the triangulation of data collection methods was used and the data was collected using multiple methods. The focal point of the study lies in interviews while other data sources, such as supporting documentation, observation and field notes, were included to increase the quality of the case study (Denzin and Lincoln, 1994;Yin, 1994;Creswell and Miller, 2000;Bernard, 2011;De Massis and Kotlar, 2014). Second, the triangulation of data sources was supported by collecting data from both owners, all employees and the external consultant as different research participants. This technique provided the opportunity to triangulate data among different participants with different roles and perceptions of the world. Third, both researchers were involved in the process of data analysis. A case study protocol helped researchers conduct analytical procedures, support transparency and reliability of the research (Eisenhardt, 1989;Yin, 1994;Lee, 1999;Gibbert, Ruigrok and Wicki, 2008;Gibbert and Ruigrok, 2010;Gioia, Corley and Hamilton, 2013). The case study protocol consisted of five parts (Runeson and Höst, 2009). First, the preamble described the purpose of the case protocol and summarised the administrative activities, tools and ethical considerations used for data manipulation. The second part described the research objective and the research questions. The third section provided a case description including the criteria for sample selection. The fourth section described the procedures and the tasks of the individual research phases such as the data collection style with the support of interview protocol and conditions. The last section specified the data analysis procedures.
Data was collected in a sequence of four visits to the organisation and a phone call with the external consultant within a one-year period. During the initial interview with the owners, we introduced the research goal, methods and conditions including permission for observation, and also checked the inclusion criteria for the case study. Semi-structured interviews were chosen for collecting the individual answers into related predefined areas, in this case, the organisational components from the suggested framework (Lee, 1999;Bernard, 2011). The collected documents described the new organisational structure, the goals and strategy of the organisation, and the people and the incentive system set out in the employees' contracts. Simple unstructured observation with an exploratory nature was used (Cooper and Emory, 1995), which is recommended for the purpose of a single case study in the natural conditions (Pretzlik, 1994;Mulhall, 2003;Flick, 2009). Considering the context of the study, the position of the complete observer was chosen (Bernard, 2011). Notes from the observation were recorded after each individual visit with the focus on the physical organisation, communication among employees, and behaviour during our presence in the organisation. The observation data provided an insight into the organisation's culture, people's behaviour and organisational context. All the collected data was organised in the case study database and created input to a thematic analysis.
Thematic analysis was used for data analysis for the potential of uncovering hidden contingencies among organisational components (Braun and Clarke, 2006;Braun, Clarke, and Terry, 2012;Grbich, 2013). Despite the current appraisal and the possibilities for using software applications, data analysis was conducted manually. The decision for manual processing was supported the fact that the main tool for conducting data analysis is the researcher due to their understanding of the data, their overview of the case study complexity and the potential to uncover hidden contingencies from the data (Denzin and Lincoln, 1994;Basit, 2003;Rudestam and Newton, 2015;Wicks, 2017).
We proceeded through six stages of conducting thematic analysis (Braun and Clarke, 2006;Braun et al., 2012) with the focus on people's actions and behaviour reflecting the organisational change. First, the data from the interviews was transcribed, printed and re-read several times in order to become familiar with the data (Braun and Clarke, 2006;Leech and Onwuegbuzie, 2007;Braun et al., 2012;Vaismoradi, Turunen and Bondas, 2013;Vaismoradi, Jones, Turunen and Snelgrove, 2016). This step included chronological organisation and unique identification of all the data in the case study database so as to be able to share the database by both authors (Langley, 1999). Subsequently, both authors discussed the story of the organisational change with the focus on the owners' need for organisational exploration and better coordination in the organisation. Thus, we began to explore the data regarding our research question. Second, after re-reading and discussing the data, we iteratively started to code the data based on the agreed coding scheme emanating from the organisation design components with the focus on organisation exploration and coordination within the organisation. The second step yielded 230 initial codes (Côté, Salmela, Baria and Russell, 1993;Attride-Stirling, 2001;Braun and Clarke, 2006;Braun et al., 2012;Gioia et al., 2013;Vaismoradi et al., 2016). The third step included the categorisation of codes to a clustering based on their similarities. We discussed each code individually with respect to the data revision from the raw data. Continuing the process of revising each code and discussing the similarities with the previously categorised codes, we clustered together the initial codes into 26 initial themes (Attride-Stirling, 2001;Braun and Clarke, 2006;Braun et al., 2012;Gioia et al., 2013). The revision of initial themes was processed in the fourth step (Eisenhardt, 1989;Braun and Clarke, 2006;Braun et al., 2012;Gioia et al., 2013). The number of initial themes was reduced to 23 and both authors agreed on the meaning of the codes with respect to the raw data. During the fifth step, the final themes were developed in relation to the research questions (Braun and Clarke, 2006;Braun et al., 2012;Gioia et al., 2013;Vaismoradi et al., 2016). The process of developing the final themes consisted of comparing the initial themes and related initial codes and searching for the relationship between each other in order to set the final themes. The fifth step resulted in five final themes: coordination/cooperation, exploration, trust, incentives, and change agents. The sixth step led to the full description of the researched case study (Braun and Clarke, 2006;Braun et al., 2012;Vaismoradi et al., 2016) provided in the results.

Findings
We present a case study research in a small firm providing recruitment services. As a result of previous growth, the owners decided to conduct an organisational redesign and hired an external consultant. We analysed the entire process of redesign from the initiation through implementation to evaluation after several months. First, we describe the situation leading to the organisational redesign and the three steps of the redesign process. Second, we analyse the effects of the change on performance, exploration and coordination.
Three main reasons drove the redesign. First, as a consequence of organisational growth, raising the number of customers and employees resulted in higher demands on coordination. Therefore, the owners needed to improve coordination through better information processing. The owners were also overloaded with too much decision-making on operational issues and thus planned to increase formalisation and decentralisation. Second, due to constant changes in regulations and volatility in the field, the owners intended to diversify the portfolio and to seek new business opportunities. Third, the external consultant suggested to focus on the efficiency of the employees by implementing a new organisational structure, several process improvements and a new result-based incentive system aimed to support decentralisation and to motivate employees to be responsible and accountable for their actions.
The sequence of changes followed the recommendations of the organisational design framework. First, a new strategy was discussed, and it was decided to develop new services that required knowledge and methods different from what was needed for their core business. Two employees, who were not experienced in a particular area, were assigned to develop the new services as a part of their job. They were highly motivated for the new mission, they invested a lot of time and effort in their self-development, and they quickly acquired new customers. The external consultant commented that their initial results were "unbelievable".
Second, a thorough job analysis was conducted by the external consultant to map out the job roles and information processing. He had several meetings with each employee as he wanted employees to feel involved in the redesign process. Based on the analysis, a new organisational structure was designed and job descriptions, including the roles and responsibilities for each position, were written. In addition, a system of meetings and notice boards was established for sharing information and regular reporting. Third, a system of PFP incentives was implemented. The first version of the PFP incentive system was designed as an individual bonus-based incentive system type with three quantitative criteria and a qualitative criterion, in which bonuses were calculated as approximately one-third of the total salary. The employees were disappointed by the system; some of them found the level of quantitative performance criteria unattainable and they felt heavily pressured because their incentives also depended on parameters that they could not really influence. The system was soon changed to combine individualbased and group-based incentive system types. Individual bonuses were based solely on the quality of work while the quantitative criteria were changed to group-based to support cohesion and teamwork. The PFP incentive system was formalised and included in each employee's employment contract.
After the redesign, the firm continued to grow according to the number of customers and revenues. Individual performance increased as did company performance. Thus, it would be possible to evaluate the change in the incentive system as successful and argue that the PFP incentives supported growth. A general view of the firm did not show anything wrong. However, only an assessment of performance would be myopic because other effects evolved over time.
The change in the incentive system had further effects, which were less visible, less tangible and which evolved over time. First, a significant deviation in attention was apparent. The employees focused primarily on activities connected with the individualbased incentives rather than on the group-based incentives, although all quantitative performance-based criteria were included as group-based incentives. Such a focus on individual-based incentives was probably driven by the first version of the PFP incentive system, which was designed as individual-based. The consequent shift to group-based incentives did not have the power to steer the attention back to cooperation.
The changes resulted in changes in interpersonal relations. Most of the employees felt a change in the organisational atmosphere. Their description of the atmosphere showed a higher level of employee individualism. Although decentralisation and higher employee self-reliance were some of the planned outcomes, individualism was an obstacle in tasks requiring cooperation and coordination. For example, employees described a decrease in helpfulness caused by the greater competition among them. They even described hiding resources in cases where sharing resources with a colleague could bring more revenue to the firm and thus greater incentives for all of them while hiding the resources could have a negative effect on the firm's reputation. Another effect was the lower level of trust, which lead to lower willingness to share information among employees and among the owners and employees because the employees were not sure if there would be any further changes in the incentive system. Hence, a change in interpersonal relationships and low information sharing counteracted the redesign activities aimed for better coordination.
Another effect, which evolved over time, was the change in pursuing the exploratory strategy. After a couple of months, the two employees assigned to develop the new services devoted significantly less effort to the development of the new services and mostly focused on core business activities. However, this effect was quite disguised and even the employees were quite surprised when they were asked about pursuing the exploratory strategy. The hidden effect can be best illustrated by the quote: 'I have no idea how that happened. I was thinking about it the other day, and I realised that now we [both colleagues assigned] do in essence the same things as the other [employees]. We do it [providing new services] only when a customer asks us'. Thus, the effect of PFP incentives is rather indirect and this effect may be caused by the unintentional diversion of attention from explorative goals to core business activities, which were incentivised.
In summary, the change in the incentive system had a positive effect on performance although the change also had detrimental effects on coordination and most probably an indirect effect on exploration. The main mechanisms were the deviation of attention to the determined criteria for bonuses and lower trust and less information sharing due to the greater competition among employees.

Discussion
This article does not intend to reconcile the disputes about the effect of PFP incentives on performance but instead draw attention to the neglected relationships of incentives. The study demonstrates how changing incentives to a PFP system may affect other organisational design components. Thus, the paper proves that the organisational design perspective provides an effective tool because the interdependencies of organisational components are essential for controlling and predicting the potential consequences of any organisational change.
However, we want to add two extending comments regarding the organisational design framework (Burton et al., 2015). First, incentives are a more nuanced component and especially individual-based incentives may not be in a good fit with a strategy focused on high exploration. Literature provides some suggestions of how to adjust PFP incentives for exploration. We dare challenge the finding that PFP incentives may foster exploration if they are set as a combination of tolerance for early failures and the prospect of payfor-performance after a learning period (Ederer and Manso, 2013). In our case, one of the mechanisms causing the detrimental effect of a PFP incentive system was the deviation of attention to pre-determined goals or criteria while the suggestion provided by Ederer and Manso only postpones the problems to the future, because employees may resort to exploitation after the learning period. On the other hand, the suggestion of low-power incentives (incentives loosely tied to the value created), which are not offered beforehand appears to have a high potential Baumann and Stieglitz, 2014) because there is a lower risk of deviation of attention from exploration. In other words, exploration may be fostered with incentives that are connected to behaviour rather than results and that are based on ex-poste evaluation rather than ex-ante agreements.
The second comment regards the difference between the structural and human components of organisational design. Structural approaches are usually more tangible and attract more attention. In our case, structural components such as strategy and structure gained more attention while the human components did not, particularly in the case of the external consultant. This might be an invisible reason for deterring coordination as well. Thus, we can fully support Bate, Khan and Pye (2000, p. 200), who stated: "On the one hand, design creates nothing. By itself, design is an empty vessel, waiting to be filled with people, meanings and actions -it is what. Yet on the other hand, it creates everything since the organisation design will have a fundamental framing effect on people's expectations and perceptions, setting the context for the organising activity -the social construction of roles and relationships -through which structure is enacted".
As a practical implication of the study, several suggestions for organisations planning to implement PFP incentives or to modify one can be provided. First, it is important to view incentives as part of a complex system of management practices. Otherwise, it is easy to get stuck in performance myopia, which is characterised by the focus on the relationship between changing incentives and performance only. Such a myopic view may indicate a positive effect of changing incentives on performance outcomes while concealing a slowly developing detrimental effect on employees, as we demonstrated on the effect on coordination and exploration. The organisational design framework is a valuable tool to broaden the focus. However, incentives are connected to further possible pitfalls and nuances, so we provide further suggestions.
The second suggestion is based on the attribute of incentives to narrow employees' attention. A direct link between results and incentives draws attention directly to money or to the incentivised goals. Thus, employees tend to limit their activity on a limited number of activities. The effect need not be intentional, but our case demonstrates that the attention may be diverted unintentionally. This effect is risky for exploration because explorative goals can be difficult to transform into measurable outcomes, especially if managers are uncertain about the future organisational course of action (Marengo and Pasquali, 2012). If there is recognisable intrinsic employee motivation for exploration, we suggest to base exploration on intrinsic motivation or to set more general explorative goals, which may be incentivised by subsequent bonuses (Gibson and Birkinshaw, 2004;Baumann and Stieglitz, 2014). Of course, such a suggestion implies higher demands on managers because of the need to achieve perceived equity of decisions about performance and bonuses.
Third, there is no turning back in changing incentives. Changing incentives may cause irreversible harm to human relations and coordination. Once attention is drawn to money or particular activities, it is difficult to forget money or to divert thinking. And once interpersonal relations are harmed by competition, it is difficult to fix the relations and to build cooperation. Following corrections of incentives do not restore the consequences of the initial change. Moreover, the removal of incentives can lead to decreased productivity (Bareket-Bojmel, Hochman and Ariely, 2017), which may finally result in the opposite effect than that expected. Hence, it is important to think about all the possible effects in advance.
Fourth, the amount of resources available is another important aspect to consider. If the total resources for rewards are fixed, PFP incentives result in the so-called zero-sum game, in which one employee's gain is another employee's loss. The zero-sum game is obvious when there is a fixed budget for rewards; however, the situation is similar when recruiters or sales representatives compete with colleagues for the same customers in a limited pool. In such a case, competition ignited by PFP incentives may lead to losing customers.
And finally, there is a risk of determining performance standards by, or in coordination with an external consultant. Companies with performance standards derived by external consultants generally have higher variability of bonus systems while having no significant difference in performance at the expense of organisational commitment and coordination (Murphy, 2000).
The paper provides some suggestions for future research. Similarly, to the organisation design perspective, strategic human resource management studies relations of different practices, such as selection and recruitment, performance appraisal, incentives, and training and development. For example, there are incentive systems recommended for high-performance (Appelbaum, Bailey, Berg and Kalleberg, 2000;Thompson and Heron, 2005), high-commitment (Whitener, 2001), and high-involvement (Wood and Ogbonnaya, 2016) human resource management although a deeper analysis may reveal further contingencies. Second, our study highlighted the key role of the so-called human components of organisational design (Burton et al., 2015). The relations (or fit) between human and structural components is probably more nuanced than usually stated in the literature. Moreover, the appropriate fit between the components may be influenced by top management and a change agent, which is quite an under-researched area. The third suggestion is based on the limitation of the study. The results are based on a single case from a small company. Thus, further explanatory studies may test the proposed relationships. For such research, we suggest controlling the effect for organisational size and industry.

Conclusion
Our case study provides a relatively unique insight into the interdependencies among organisational design components within the organisation redesign process. PFP incentives may be a powerful tool to enhance performance. However, looking solely at performance may be misleading. Even if the individual or organisational performance improves, related detrimental effects may subsequently appear. These effects are less tangible, difficult to measure and difficult to reverse. We demonstrated a case, in which PFP incentives diverted attention from pursuing an exploration strategy and harmed employee relations and coordination. Thus, we suggest employing a holistic view of an organisation based on an organisational design framework to be able to analyse the possible consequences of implementing PFP incentives. In addition, there is no turning back in changing incentives; it is easy to divert attention or ruin relationships, but it is difficult to repair them. Particularly for incentive systems, it is essential to think of the consequences before taking action.