آموزش و نوآوری: بهره برداری و اکتشاف مبادلات
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|20152||2012||6 صفحه PDF||سفارش دهید||5561 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Business Research, Volume 65, Issue 8, August 2012, Pages 1189–1194
This paper examines the relationship between learning and innovation outcomes, focusing on the trade-off between exploitation and exploration in learning and innovation. The study identifies two types of learning and two outcomes of innovation. Exploitation and exploration in learning are inversely associated with innovation rates and impact. While exploitative, localized learning is positively associated with innovation rates, but negatively associated with impact, exploratory learning-by-experimentation shows the opposite relationship. The study examines panel data of 103 companies in the global pharmaceutical industry over a 7-year period in an empirical test of our hypotheses. Results support the existence of the exploitation and exploration trade-off.
Innovation is one of the most important organizational processes and outcomes for value creation (Deeds, DeCarolis, & Coombs, 2000). Innovation is a central mechanism for strategic change and growth whereby organizations exploit, explore, and reposition themselves in changing internal and external conditions (Dittrich & Duysters, 2007). Both exploitative and exploratory learning govern innovation (March, 1991). Exploitation increases the efficiency of existing technologies, while exploration is required to produce new technologies of high quality and impact (Henderson, 1993). Thus, there is an inherent tension between exploitation and exploration in organizational learning in terms of outcomes of innovative activities (Sorensen & Stuart, 2000). A few empirical studies differentiate impact from innovation rates as innovation outcomes (see Gittelman and Kogut, 2003, Rosenkopf and Nerkar, 2001 and Sorensen and Stuart, 2000). Key to a firm's technology strategy is to strike the right balance between the two major types of learning—exploitative vs. exploratory—depending on what innovation outcomes—rates vs. impact of innovation—the firm is targeting. The study here asks: How do types of organizational learning shape innovation outcomes? Existing literature suggests that exploitative “localized learning” improves immediate innovation rates, but it often simultaneously reduces incentives for and competence with high-impact innovation (Ahuja & Lampert, 2001). Thus, firms must combine exploitative “localized learning” with exploratory “learning-by-experimentation” if they also want to enhance the impact of innovation. Although exploitation and exploration and their effects on innovation have been intensively examined (such as in Ahuja & Lampert, 2001), few empirical studies investigate the actual trade-offs between the two. Exceptions include Atuahene-Gima, 2005 and Auh and Menguc, 2005 in marketing literature. Unlike those studies, which used questionnaire methods, however, this paper employs longitudinal patent data to test empirically the trade-off between exploitation and exploration. Moreover, joint consideration of innovation rates and impact in this study with the inverse relationship between exploitation and exploration enables us to examine the discriminating effects of exploitation and exploration on outcomes of innovation that have not been tested before. A cursory look at our database in the global pharmaceutical industry shows an interesting pattern: science-intensive firms such as Genentech and Immunex, which focus on exploratory learning, appear to outperform others in terms of innovation impact (refer to Table 2). On the other hand, the most prolific firms in terms of the number of patents, such as Bayer and E. I. DuPont, are among the few that focus on exploitative learning based on strong technological competence. This interesting pattern is consistent with the inherent trade-offs between exploitation and exploration in organizational learning and innovation outcomes that we address in this paper. To provide more rigorous empirical findings, we constructed panel data of 103 companies in the global pharmaceutical industry over a 7-year period and then tested our hypotheses on relationships between types of learning and innovation outcomes. Results support the existence of the exploitation and exploration trade-off.
نتیجه گیری انگلیسی
This study advances the evolving research on innovation by considering the duality of any innovation process in terms of two outcomes: rates and impact. March (1991) highlights the inherent tension between exploitation and exploration in organizational learning and innovation processes. This tension is the constant focus of subsequent studies in innovation, management, and marketing literature. Indeed, there is a rich tradition in marketing of studying diverse aspects of innovation and new product development (Wuyts et al., 2004). Notwithstanding the merits of this tradition, however, marketing literature tends to rely on the questionnaire method via survey, which suffers from self-report bias in measuring learning and innovation (Sorescu, Chandy, & Prabhu, 2003). The longitudinal investigation here of patent data extends this line of research by providing additional hard evidence on the trade-off between exploitation and exploration. This finding implies that if a firm wants to improve its innovative capabilities and output in terms of frequency and impact, then it should strike a balance between exploitative, localized learning and exploratory learning-by-experimentation. This idea is, in essence, the popular “ambidexterity” premise (Tushman & O'Reilly, 1996), which says that firms must undergo the paradoxical strategic process of balancing between exploitative and exploratory innovation strategies (He & Wong, 2004). Indeed, seeking ambidexterity by conducting both types of innovation is a reality in the fast-evolving technological and market conditions of today. At any given time, firms may have to emphasize either exploration or exploitation, yet over time, a balance should be maintained. Incidentally, this remains a fruitful area for further research. The result for the R&D alliances variable deserves some explanations. By controlling for R&D alliances, we emphasize that technology learning occurs not only within a firm boundary, but also across firm boundaries. Results imply that R&D alliances contribute to innovation rates, but not to innovation impact. If we combine this finding with findings regarding technological competence and science intensity, we could suggest that exploration in the form of strengthening the external linkage to cutting-edge scientific knowledge and technology may be the only way to enhance innovation impact or improve the chances of impactful innovation. By juxtaposing internal and external R&D exploitation–exploration, firms may be able to overcome trade-offs in learning and innovation (Lavie & Rosenkopf, 2006). However, our data constraints prevented us from distinguishing between exploitation and exploration within the domain of R&D alliances. Even if the measure of the intent is elusive, this remains an area for further research, perhaps using more detailed survey-type alliance data. Relatedly, due to data constraints, this study could not examine specific environmental circumstances where either exploitation or exploration would be more effective in improving innovation rates and/or impact. In general, we predict that exploration efforts may become more value-adding when a firm faces greater complexity and uncertainty in its innovation. Indeed, Fleming and Sorenson (2004) show that science-guided search associates closely with high-impact innovations, especially for the highly complex innovations in pharmaceuticals. Future research along this line may enrich our understanding of how a firm manages innovation in a turbulent technological and market environment. Although the findings may be generally applicable to other technology-intensive industries where patenting innovative outputs is important, such a claim is an empirical question. Thus, conducting future research is necessary in other industrial settings to corroborate the findings in this study.