ارزش مشروط یادگیری اکتشافی و بهره برداری برای عملکرد توسعه محصول جدید

This study endeavors to extend research on organizational learning by investigating the complicated effects between exploratory — exploitative learning and new product performance in a single new product project. Specifically, premised on contingency theory the authors investigate the negative nonlinear and interaction effects of project-level exploratory and exploitative learning behaviors on product development performance, and examine internal organizational and external environmental factors to recognize their differential moderating effects between the two learning behaviors and new product performance. Most of the hypotheses are supported based on questionnaire survey results of 253 new product projects. The results indicate that the two type of learning have curvilinear (inverted U-shaped) effects on new product performance, and suggest that product development performance will be enhanced when one learning is at higher level and the other is at lower level. Furthermore, the authors discover that process-based reward, encouragement to take risk, and environment dynamics strengthened the benefits of exploratory on new product performance. On the other hand, the advantages of exploitative learning on new product performance is further enhanced when output-based reward, project development formalization, and environment competitiveness is high. Finally, this study suggests that project managers should pay careful attention to employ the two learning behaviors during new product development.

For several years, many scholars have dedicated considerable effort on further understanding the nature of exploratory learning versus exploitative learning (e.g., Levinthal and March, 1993 and March, 1991). Thus, an accumulating body of research has established that exploratory and exploitative learning leads to better performance in new product development (NPD) and organizations (e.g., He and Wong, 2004 and Katila and Ahuja, 2002). On the other hand, some studies have examined the antecedents (i.e., market orientation or top management team's social capital etc.) and consequences (i.e., innovativeness or new product performance etc.) of exploratory and exploitative learning (e.g., Atuahene-Gima & Murray, 2007). Despite this growth of exploratory and exploitative learning literature, there are some oversights in the extant empirical research; and thus, the purpose of this study is an attempt to address these oversights. The first research gap is that prior research has focused on the more macro levels of analysis (i.e., firm to inter-firm). For instance, Atuahene-Gima and Murray (2007) studied the two types of learning at the firm level and found that the top management team's social capital was the key antecedent. Using patent data, Katila and Ahuja (2002) found that the firm's search depth and search scope were related to new product innovation at the firm level. As Gupta, Smith, and Shalley (2006) noted, empirical studies that examine exploratory and exploitative learning at a more micro level are relatively scant. The second research gap concerns the relationship between exploratory and exploitative learning and NPD performance. Although several studies have examined the nonlinear impact of exploratory and exploitative learning on NPD performance and the interaction effect of exploratory and exploitative learning at the firm level, empirical findings have been ambiguous on both fronts (e.g., Atuahene-Gima and Murray, 2007, He and Wong, 2004, Katila and Ahuja, 2002 and Nerkar, 2003). The unclear findings may have resulted from previous studies that took the firm or business unit as a unit of analysis and ignored the fact that the two learning behaviors might not have suffered from the constraint of scarcity due to the accessibility of potential external resources (Gupta et al., 2006) at the firm or business unit level. Perhaps, the two learning behaviors may explicitly differentiate the processing conditions and their mechanisms at the level of a single new product project. Although few previous studies has asserted that organizational antecedents differentially influence exploration and exploitation (Raisch & Birkinshaw, 2008), few if any studies in exploratory and exploitative learning literature have examined internal organizational factors that moderate the relationship between exploratory learning or exploitative learning and NPD performance. For these reasons, in this study, we focus particularly on the project level in an attempt to contribute to the exploratory and exploitative learning literature in three ways. First, we argue that given the potential risks involved in achieving an inappropriate level of exploratory or exploitative activities and the possibility of diminishing returns on their contribution to the NPD project, it is likely that the level of explorative or exploitative activities that is either too low or too high may harm performance; and a reasonable level may be optimal. From the perspective of contingency theory, we examined the nonlinear relationship between explorative or exploitative activities and NPD performance at the project-level unit of analysis. Second, based on the argument of Gupta, and Smith et al. (2006), we also examine if high (low) exploration needs to be associated with low (high) exploitation to enhance NPD performance at the project level. By focusing on a single new product development project, this study implies that the interaction effect for the two learning processes on performance will be negative. Finally, our study adds to the existing body of knowledge concerning exploration and exploitation literature by examining the moderating effect of the internal organizational factors between the two learning behaviors and new product performance. Given that the two learning behaviors involve complex information-processing tasks, we highlight process and system factors such as the reward system and the work environments (encouragement to take risks and project development formalization), all of which affect the information-processing capacity of the project members as internal organizational moderators between the two learning behaviors and NPD performance. Finally, we advance the extant literature through testing a contingent model (Hofer, 1975) by using a single new product project as unit of analysis, which stresses that competitive advantage is conditional upon external environmental conditions. Following this insight, we investigate two sources of external factors (environmental dynamics and competitiveness) so as to recognize their differential moderating effects between the two learning behaviors and new product performance. Through this richer framework and empirical assessment, we attempt to respond to Gupta, and Smith et al. (2006) call for research that examines a more micro level model of exploratory and exploitative learning. Consequently, this study addresses these oversights in the literature by investigating (1) nonlinear relationships between exploratory and exploitative learning and NPD performance, (2) their interaction effect on NPD performance, and (3) internal organizational and external environmental factors that moderate their effects on NPD performance.

The results of this study advance the exploratory and exploitative learning literature in several ways. First, one result is to show that exploratory learning has a curvilinear (inverted U-shaped) relationship with NPD performance at a single new product project level. This implies that there are positive returns to the exploratory learning activities but that it becomes detrimental to NPD performance beyond a certain level. This finding supports propositions in the organizational learning theory (Levinthal & March, 1993) that too much exploration in new information and emergent insights may be associated with high risk and cost such as expenditures for acquiring and using new knowledge (i.e., latent and emerging market needs), thereby harming product development performance. Given a rational level, it follows that project members who are able to employ exploratory learning have better product development capabilities than their competitors; and thus, they may offer a superior new product. Another possible explanation for this result according to the information-processing theory is that there may be cognitive barriers in processing many new ideas for the amount of time spent in the coordination of these new ideas that may decrease the reliability of NPD (March, 1991). Second, another finding also illustrates that exploitative learning has a curvilinear (inverted U-shaped) relationship for NPD performance at a single new product project level, suggesting an increase in the disadvantages of exploitative learning after a certain point. In other words, in-depth exploitation of existing product development capabilities may make it difficult for project members to adopt alternatives, and take newer directions in the development of a new product. This result supports the anecdotal finding that excessive exploitation in using the same ‘know-how’ of product development may injure new product performance (Katila & Ahuja, 2002). Similarly, given a rationale level, project members' exploitative learning generates better product development efficiency, such as, the reduction of errors in problem solving and the avoidance of mistakes in NPD processes. Furthermore, overemphasizing on exploitative learning may lack the novelty found in problem solving so that new products with no competitive dimension are able to compete with other offerings of competitors (Nerkar, 2003). Third, the finding of a negative interaction between exploratory learning and exploitative learning resonates with research that notes appropriate balance of exploratory and exploitative learning needs a high — low combination rather than a high — high combination. It implies that exploratory learning (exploitative learning) could be more valuable to NPD when it is coordinated with a rationale level of exploitative learning (exploratory learning).This finding is also consistent with the argument that both exploratory learning and exploitative require a little dose of each other to improve new product performance (i.g., Nerkar, 2003 and Atuahene-Gima, 2005). Furthermore, it appears that project members using a dose of exploratory learning may moderate those disadvantages of exploitative learning by generating some new insights in product innovation and problem solving (March, 1991). On the other hand, some exploitative learning through using existing knowledge may transfer new insights more efficiently, thereby tempering the risks and costs of exploratory learning (Danneels, 2002). The present study finds that the internal organizational and external environmental factors respectively moderated the impacts of exploratory learning and exploitative learning on NPD performance. Specifically, the reward system — process-based reward and output-based reward — moderated the impacts of exploratory and exploitative learning on NPD performance respectively. The results found that the impact of exploratory learning can be further enhanced when there is a high process-based reward system. It may be that because high process reward systems provide some psychological safeguard for project members, the members do not fear the risk of uncertainty and errors in NPD, and then, they are willing to undertake novel problem-solving approaches. In addition, the finding indicates that a high output-based reward system may elevate the effects of exploitative learning on new product performance. It indicates that project members may focus on innovation with low risk and more certain returns to attenuate the potential risk of new product ideas. Furthermore, the results provide support for arguments made in learning, management, and social psychology literature (e.g., Hitt et al., 1996) that the firm's new product reward systems are more important for learning behaviors and their outcomes due to the project members' assessment of new product performance risk. On the premise that environmental dynamics and competitiveness may have different implications for the effects of exploratory and exploitative learning on new product performance, the study examined their moderating effects respectively. Environmental dynamics positively moderates the impact of exploratory learning on new product performance. The result indicates that exploratory learning through gathering and assimilating new insight about product concepts and market information resists the threat of obsolescence of existing competences, thereby increasing the opportunities of a successful new product. On the other hand, the effect of exploitative learning on new product performance can be further enhanced when the level of environmental competitiveness increases. It shows that efforts of project members to respond to environmental competitiveness through new combinations and re-combinations of existing market and technological knowledge have a positive effect on NPD performance. The result supports the arguments (e.g., Day, 1994) that the nature of the external environment has an important role in the effectiveness of learning behaviors. The effect of exploratory learning on new product performance is moderated by the work environment of encouragement to take risks. In other words, project members will choose to be risky and innovative in NPD when they are given the green light to do so. The result supports the arguments (Cyert & March, 1963) that the firm should create a stimulating atmosphere and organizational routines, thereby facilitating the pursuit of new ideas and approaches. Furthermore, project developmental formalization positively moderates the effects of exploitative learning on new product performance. This implies that project developmental formalization may increase the benefits of exploitative learning. It may be that because such a work environment could facilitate members' information sharing, which could improve the efficiency of existing product development information among the members. Besides, our findings provide substantial support for the arguments (Burns & Stalker, 1961) that the ambidextrous configurations of the work environment have differential incentive mechanisms between learning behaviors and new product performance. 5.1. Managerial implications The findings of our study make important contributions to managerial practice. Managers may heed the recommendation that an absolutely positive view may be naively oversimplified, because exploratory learning and exploitative have curvilinear effects on new product performance. In this regard, project managers must be cognizant of the potential disadvantages of the two learning behaviors, in particular the costs of excessive exploratory learning such as the assimilation of unfamiliar knowledge, and the problems of excessive exploitative learning such as of a new product without adding the novel and competitive dimension. For a new product development to be successful, the managers should make a conscious choice to keep a rationale level of exploratory learning or exploitative learning without the potential costs and problems. Furthermore, while engaging on a high level of exploratory learning, our study indicates to managers that reliance on a high level of exploitative learning is detrimental to new product performance. For instance, a serious ultra-thin mobile handset of Motorola, RAZR, sold more than 100 million units since their introduction. More recently, those ultra-thin handsets' sales dropped and the market share of the company slid to third place because the ultra-thin handset of Motorola overemphasized in exploitation during the processes of product development. The processes had failed to invest in newer technologies (e.g., software for multimedia functions) for follow-up new handsets, thereby harming the new product development performance. In contrast, our findings suggest that while project managers make an effort to have a high level of exploratory learning, project members should be exposed to a little dose of exploitative learning in order to offset the inferiority stemming from new and emerging insights, and vice versa. Consequently, as the results indicate, managers may need to pay careful attention to allocating the project's limited resources between exploratory learning and exploitative learning during new product development. On the other hand, our study also cautions managers that the impact of both learning behaviors on new product performance is moderated by both internal and external factors. For instance, exploratory and exploitative learning are able to achieve better product development performance where they are matched with the appropriate organizational conditions (reward systems and work environments). Where engaging exploratory learning, project managers need to be concerned with ensuring process-based reward and greater stimulation of risky and innovative behavior in new product development so as to tap effectively into the new insights of product designs and the problem-solving domain in product development. These viewpoints also clearly suggest that to take advantage of exploitative learning project managers must work to develop output-based rewards in the project team that would let project members to focus on innovation with low risks and certain returns during new product development. Similarly, these findings imply that developing formalized product development routines that could facilitate the sharing of existing technological and market knowledge. Specifically, the divergent moderating effects of environmental dynamics and competitiveness suggest to managers that these sources of external environments create different resource-allocating and information-processing demand in new product development. 5.2. Limitations and further research Some limitations need to be recognized, and we shall note them while interpreting the results of our study. The first limitation of this study is that the result could not infer causal relationships due to its cross-sectional design. Further research might use a longitudinal design to examine these linkages more clearly. Second, self-reported data pose potential problems, such as CMV or limited retrospective. However, our results used on-site interviews for gathering consensus ratings to improve the reliability and validity of the research constructs. In addition, the post hoc examination with the Harman's one-factor analysis (Podsakoff & Organ, 1986) which indicated no serious common method problems provided evidence against the presence of common method bias. Third, our study controlled for only a limited set of the potential factors of the two learning behaviors and new product performance due to the limitations of survey research. Future research may include some potentially influential covariates that have been adopted in previous studies, such as research and development strength (Li & Calantone, 1998) and ease of market entry (Atuahene-Gima et al., 2005). Furthermore, further research should examine other dimensions of environmental dynamics and competitiveness such as the time dimension (the duration of competitiveness) and the differential source of uncertainty (technological dynamics and market dynamics). In addition, our study collected data from a sample of firms located in Taiwan, thereby limiting the generalized level of our findings. Fourth, some scholars indicated that there are different phases in the processes of new product development which are need of respective combination of the information and knowledge (Biazzo, 2009). Future research should examine the effect of exploratory learning and exploitative learning as well as the optimal combination in each of the different new product development phases. Finally, there is an important research direction at the analysis of project level that could uncover whether the team member attributes and the configuration of team members plays a significant role in exploratory and exploitative learning among new product development project.