Research in strategic management and innovation suggests that part of the base for heterogeneity in innovative capabilities resides in firms' search behaviors for new products (Cyert and March [1963]; Nelson and Winter [1982]; March [1991]; Katila and Ahuja [2002]). Firms tend to conduct local search around current areas of expertise (March [1991]; Stuart and Podolny [1996]) due to the effects of outcome predictability, cost efficiency, and technological confidence associated with internal exploitation (March [1991]). However, local search may constrain a firm's sensitivity to external changes in customer tastes, technological frontiers, and competitive dynamics, which may turn core competence into core rigidities (Leonard-Barton [1992]). To avoid such competency traps, firms attempt to explore knowledge elements that reside outside their current technological domains and their organizational boundaries (Katila and Ahuja [2002]; Rosenkopf and Nerkar [2001]).
Developmental resources are scarce for most firms, however, and this suggests that exploration activities will need to compete for resources against those projects focused on the exploitation of current resources and capabilities (March [1991]). This raises the issue of how to balance exploitation and exploration. Current research, however, has not provided clear answers on how to balance exploitation and exploration in technological search. For instance, both positive linear and curvilinear relationships have been found between exploration and innovative performance in empirical studies (e.g., Katila and Ahuja [2002]; McEvily and Yao [2004]).
Following Pisano ([1996]), this study takes a contingent perspective and argues that the effectiveness of search behaviors is moderated by the characteristics of a firm's knowledge base. As a firm explores new opportunities or solves new problems, its knowledge will be a starting point for future search. A firm's knowledge base influences the direction and effectiveness of its search behaviors (Yayavaram [2003]). We investigate how three characteristics of a firm's knowledge base (depth, scope, and technological opportunities) affect its innovative results.
Our hypotheses are tested with data from firms in the US electrical medical device industry (standard industrial classification (SIC) = 3,845) over the period from 1990 to 2000. Our findings indicate that the characteristics of a firm's knowledge base are closely related to its search behaviors. When a firm seeks to innovate, it must take into consideration its current knowledge base and must be cautious about the degree of internal exploitation.
Literature review
Firms can innovate by searching and integrating new knowledge elements (Cohen and Levinthal [1990]). Typically such search behaviors can be classified as exploitation and exploration (March [1991]; Stuart and Podolny [1996]). March defines exploitation as the ‘refinement and extension of existing competences, technologies, and paradigms’ ([1991]: 85) and claims that the returns from exploitation tend to be positive, proximate, and predictable. The alternative, exploration, is ‘experimentation with new alternatives’ (March [1991]: 85), and its returns are uncertain, distant, and often negative. Compared to exploitation, exploration is more likely to be associated with risk taking, uncertainty, and a long-term orientation.
Because of the risks associated with exploration, firms may prefer exploitation when it is possible. However, too much exploitation will reduce a firm's adaptive capabilities and, from a long-term perspective, may impair its competitive advantage. To avoid potential core rigidities and remain competitive, firms often experiment with new technologies (Leonard-Barton [1992]; Bierly and Chakrabarti [1996]). Firms that engage in exploration and integrate novel, emerging, and pioneering technologies into their operations will be more likely to generate influential knowledge than firms engaged in more limited local search processes (Ahuja and Lampert [2001]). This suggests the need for a balance between exploitation and exploration.
Recently Katila and Ahuja ([2002]) have proposed two dimensions to help better capture a firm's search behavior: search depth and search scope. Search depth refers to ‘the degree to which search revisits a firm's prior knowledge’ (Katila and Ahuja [2002]: 1184). This reflects a firm's capability in identifying new opportunities for existing knowledge elements. Search scope means ‘the degree of new knowledge that is explored’ (Katila and Ahuja [2002]: 1184), which captures a firm's willingness and capability in exploring and integrating new ideas, which is similar to the so-called ‘absorptive capacity’ (Cohen and Levinthal [1990]). These dimensions are extensions of March's ([1991]) distinction between exploitation and exploration and allow for the possibility that a firm might be active at both exploiting existing technologies and exploring new knowledge. Hence, the use of depth and scope ideas permits a more detailed and complete description of a firm's search behaviors.
Prior research suggests that search depth and search scope have different associations with the innovative results achieved by firms. For instance, Katila and Ahuja ([2002]) measure innovation performance by the number of new designs in the robotics industry and examine how a firm's search patterns affect its rate of new product introductions. They find that when a firm continues searching for new knowledge, more new products come out. Moreover, there is an inverted-U relationship between search depth and innovation performance, implying that too much internal search might constrain a firm's capability to find new ideas and designs. They also identify a significant and positive interaction result between search depth and scope which suggests a complementary relationship between them.
While a firm's search patterns may affect its innovative performance, research results are inconclusive and even contradictory regarding how. Some studies find that exploration has a positively linear relationship with new product innovation. For example, Katila and Ahuja ([2002]) observed that in robotics industry, the more new knowledge elements are searched, the more new products are created. Rosenkopf and Nerkar ([2001]) find a strong relationship between search scope and the impact of new inventions in the optical disk industry. As firms search beyond their technological domains and organizational boundaries (radical search), they are more likely to create influential inventions. Along with these results, however, studies also show that moderate exploration helps create more new products. For example, McEvily and Yao ([2004]) find an inverted-U relationship between exploration and new product innovation.
The above studies focus on the effect of search without considering the firm's knowledge base, which reflects a firm's competence and experience in special knowledge domains. This could be misleading as Penrose has argued, ‘… unknown and unused productive services [from existing resources] immediately become of considerable importance, not only because the belief that they exist acts as an incentive to acquire new knowledge, but also because they shape the scope and direction of the search for knowledge’ ([1959]: 77-79). Thus, the context of current knowledge should influence the relationship between search behaviors and innovative performance.
Pisano ([1996]) provides two insights into this question. First, the resource-based view emphasizes the value of knowledge and organizational competencies and conceives them as competitive assets, a rather limited vision. More attention is needed on the interaction between the firm's knowledge base and its competencies. Second, firms' knowledge bases are idiosyncratic and heterogeneous, even within the same industry or the technological area. In the setting of knowledge creation, then, there is no ‘best’ knowledge strategy. Rather, ‘the appropriateness of different practices and approaches may vary depending on characteristics of the knowledge environment’ (Pisano [1996]: 97).
To test this proposition, Pisano ([1996]) compares the influence of different learning types in two industries that differ in research settings: synthetic chemicals and biotechnology. Empirical results show that a greater input in the research phase of a chemical process development results in shorter lead times, confirming the proposal that learning before doing is more important in the chemical industry. On the other hand, results show that learning by doing is appropriate when organizations do not possess required knowledge, such as in biotechnology. Thus, the value of organizational capabilities is dependent on the characteristics of a firm's knowledge base.
This study applies Pisano's contingent perspective and argues that the relative effectiveness of exploitative search will depend on a firm's existing knowledge base. While previous studies have examined the interaction between explorative search and a firm's knowledge base (Wu and Shanley [2009]), this research attempts to focus on the role of exploitative search behaviors. The following section develops hypotheses on the main effects of search behaviors and the moderating roles of the firm's current knowledge base.
Hypotheses
Because of bounded rationality and limited resources, firms tend to search around their existing technological domains where their competences reside (Nelson and Winter [1982]; Cohen and Levinthal [1990]; March [1991]; Stuart and Podolny [1996]). As the depth of search increases, we would expect, at least technically, that more new products will be created, since the firm has already accumulated extensive experience in current knowledge domains (March [1991]; Fleming [2001]). Searching locally also is more likely to avoid technological mistakes and reduce experimentation time. As a result, with increasing search depth, more new knowledge is likely to be discovered.
Strategically, local search brings other benefits. First, it makes competitive imitation less likely since potential entrants will need to master early-generation knowledge in order to succeed. However, those early technologies may require a long time to build and understand or may be protected by patents. Second, the firm may enjoy economies of scope in the development of new products. When technologies are incrementally built up, current complementary assets will be useful in supporting next-generation products, which provides cost advantages to the firm relative to competitors. Finally, focusing search on a specific domain can enable the firm to build a reputation in the niche (Bierly and Chakrabarti [1996]). If a firm continuously improves existing products and introduces new but related generations, it is more likely that these products will be developed with customer needs in mind (Christensen [2000]). Customers, in turn, will be more aware of the firm's innovative capabilities and will be more loyal to its products.
However, the benefits of exploitation do not last forever. From a technological point of view, there are decreasing returns to physical scaling (Sahal [1985]). That is, given a set of knowledge components, the number of possible recombinations is limited. Kim and Kogut ([1996]) and Fleming ([2001]) suggest that when a group of technologies is repeatedly applied, the potential for future combinations eventually will be exhausted. Part of this effect comes from the diminished ability of developers to conceive new applications. The ‘imaginary life cycles’ (Henderson [1995]) of new product developers will tend to petrify, making them less likely to incorporate new components into their products. Finally, deep search within existing knowledge domain may form competency traps and lead to core rigidity (Leonard-Barton [1992]).
These considerations suggest that beyond a certain level of search depth, we would expect that the costs of local search will be higher than the benefits. This implies an inverted-U relationship between a firm's exploitation of existing knowledge and its capability in new product innovation and suggests the following hypothesis:
• Hypothesis 1: There is an inverted-U relationship between search depth and product innovation, given all else equal.
Moderating effects of the knowledge base on search depth
Although a deep search around areas of current expertise is less risky and costly than an exploratory search, there are still costs associated with it (Leonard-Barton [1992]). When should a firm conduct deep search within its current knowledge domains as opposed to exploring new domains? The innovation literature suggests that both a firm's external environment and its internal knowledge base affect the effectiveness of deep search. This study concentrates on the internal knowledge base while examining firms that share a general external environment (electrical medical devices). Under these conditions, the expectation is that when a firm's current knowledge base is diverse and rich in opportunities, deep search will be more productive.
Knowledge depth and search depth
Deep understanding in a specific technological area comes from deliberate investment, active learning, and historical accumulation. Due to resource constraints, a firm can invest in only a limited number of technological areas in order to establish such technological superiority. Once this knowledge expertise is developed, the firm tends to search within current domains and exploit potential opportunities. However, as the depth of the knowledge base increases, the chance of discovering novel applications and new knowledge elements diminishes (Kim and Kogut [1996]; Fleming [2001]). Deep search may lead to core rigidities that significantly hamper a firm's ability to adapt to competitive dynamics (Leonard-Barton [1992]). Thus, we propose that
• Hypothesis 2: The relationship between search depth and product innovation will be negatively moderated by knowledge depth.
Knowledge scope and search depth
A firm with a diverse knowledge base possesses expertise in many domains. This provides an opportunity for a firm to discover new linkages among existing knowledge elements. Given a diverse knowledge pool, a firm skilled in combining its competences can create more new knowledge than its rivals. In addition, the more diverse a firm's knowledge base is, the more linkages are created. This suggests that deep exploitative search will be most appropriate for a highly diverse knowledge base.
When a firm's knowledge base is very specialized, however, covering only a limited number of domains, internal exploitation may impair a firm's long-term performance. Given a limited space in which to discover new knowledge linkages, excessive exploitation may turn a firm's core competence into core rigidity. Therefore, when a firm has a limited number of knowledge domains, it is less likely to maintain competitive advantage if the degree of internal exploitation is high. When the knowledge base is highly diverse, internal exploitation will help the firm maintain high performance. This suggests the following hypothesis:
• Hypothesis 3: The relationship between search depth and product innovation will be positively moderated by knowledge scope.
Technological opportunity and search depth
Technological opportunity refers to the ease of achieving innovations and improving techniques (Fung [2004]). Industries differ systematically in terms of the strength and sources of technological opportunities (Fung [2004]; Klevorick et al. [1995]). Such differences can be represented by three measures (Fung [2002]): knowledge spillover, inter-firm research overlap, and scope of research. Fung ([2002, 2004]) operationalized these measures with patent information and found that technological opportunities were significantly associated with differences in research productivity in three industries: chemical, computers, and electronics. Similarly, Klevorick et al. ([1995]) identified three sources of technological opportunities: advances in scientific understanding and technique, advances originating from other industries and from other private and government institutions, and feedback from an industry's own technological advances. Differences in technological opportunities exist not only across industries but also across firms in the same industry, as evidenced by McGrath and Nerkar ([2004]) in their study of pharmaceutical firms.
We expect that firms will find it worthwhile to conduct deep search within their current knowledge base when that base is perceived to be rich in technological opportunities. This is likely for several reasons. To start with, the perceived benefits of a domain may more than offset the costs of exploitation and make the domain absolutely more desirable. In addition, technologies with high potential tend to require more novel development projects, whose first generations are primitive and rudimentary. They will often require additional deep search by first movers to make the necessary changes and produce the more commercially feasible versions that justify a firm's heavy initial developmental expenditure. If the inventing firm stops further exploitation of a novel technology, others can quickly catch up without the same risky initial investments. Finally, high levels of perceived opportunities may prove a constant spur to entry. This means that in order to stay competitive and prevent imitation and expropriation of rents, the inventing firm will need to remain innovative and continuously improve its technologies (Ahuja [2003]). These considerations suggest the following hypothesis:
• Hypothesis 4: The relationship between search depth and product innovation will be positively moderated by technological opportunities associated with the knowledge base.