It's how broadly you search: The moderating effect of search span on the relationships between knowledge origins and innovative performance in R&D alliances

Innovation is a fundamental source of competitive advantage and a large literature has struggled to understand the drivers of innovation and how innovation processes should be managed in order to increase innovative performance (Brown and Eisenhardt, 1995; Crossan and Apaydin, 2010). Innovation is a search process leading to the creation of new knowledge (Nelson and Winter, 1982; March, 1991), which in turn typically results from novel combinations of existing pieces of knowledge having different origins (Schumpeter, 1934; Kogut and Zander, 1992). Based on this, we argue that, in order to deepen our understanding of the determinants of innovative performance, we should concentrate simultaneously on the origins of the existing knowledge that is combined to generate innovation and on how economic actors search for new knowledge. Previous research has focused on the innovation performance implications of combining knowledge from different geographic (Gomes-Casseres et al., 2006; Phene et al., 2006; Capaldo et al., 2012) and organizational origins (Rosenkopf and Nerkar, 2001; Miller et al., 2007), reaching however mixed conclusions. Thus, a first objective of this study is to shed more light on the influence of the geographic and organizational origins of knowledge on innovative performance. The origins of knowledge are but one of the determinants of innovation, however. The characteristics of the processes by which firms search for new knowledge also exert significant influence on innovation and have the potential to shape, or even subvert, the impact of the origins of knowledge on innovation outcomes (Fleming and Sorenson, 2001; Katila and Ahuja, 2002). Extant research has explored several dimensions of search (e.g., Katila and Ahuja, 2002; Laursen and Salter, 2006; Miller et al., 2007). We build on the observation that, as customer needs become increasingly multifaceted and the pace of competition intensifies, innovating firms increasingly resort to combining heterogeneous knowledge in order to generate complex innovations, spanning multiple different knowledge domains (Levinthal and March, 1993). In order to capture this innovative behavior we point to search span, that is, the extent to which firms search for new knowledge across different knowledge domains (Capaldo and Messeni Petruzzelli, 2011). Thus, a second objective of this study is to ascertain whether and how search span moderates the influence of the geographic and organizational origins of knowledge on innovation performance. We pursue the above two objectives at the interorganizational level of analysis, specifically in the context of R&D alliances. R&D alliances are increasingly important units of analysis for understanding competitive advantage based on innovation (Capaldo, 2007; Sampson, 2007). As intended here, R&D alliances are dyadic knowledge-intensive interfirm collaborative relationships aimed at developing innovation. Knowledge-based interpretations of interorganizational collaboration suggest that firms enter R&D alliances to jointly search for new knowledge by combining their respective knowledge (Powell et al., 1997; Sakakibara, 1997). Along this way, R&D alliances constitute tremendous sources of knowledge and learning, considerably affecting innovation results and competitive advantage (Mowery et al., 1996; Stuart, 2000). Based on the above, the present study aims to expand our understanding of the determinants of innovative performance in R&D alliances by focusing simultaneously on both (a) the geographic and organizational origins of the knowledge resources that allied organizations contribute to their joint innovative endeavors and integrate across their boundaries for the benefit of the alliance, and (b) the extent to which allied organizations jointly search for new knowledge across different knowledge domains, that is, joint search span (simply search span hereinafter). To do so, we focus on innovative performance at the alliance (i.e., dyad) level of analysis. Drawing on a sample of 1,515 R&D alliances we found that, although both the integration of geographically distant knowledge and the integration of organizationally proximate knowledge in R&D alliances negatively affect innovative performance at the alliance level, the alliance search span positively moderates both relationships.