The creation of impactful technical knowledge provides companies with a sustainable competitive advantage, significantly influences social welfare, and promotes economic growth. Because the generation of new knowledge mainly comes from the search for and recombination of existing knowledge, search and recombination techniques play an important role in generating both new knowledge and the subsequent impact of the recombinant outcomes. Indeed, existing research has shown that variations in search scope and search depth are viable strategies for investigators to create impactful knowledge. However, these studies have produced conflicting and contradictory findings on the relationship between search scope and knowledge impact. Furthermore, the literature has paid limited attention to the effects of knowledge search strategies on the commercial value of innovative outcomes.

This dissertation aims to address the aforementioned research gaps through three interrelated studies. First, the dissertation develops a novel theoretical perspective to reconcile the current debate. In particular, the first study refines the concept of search scope, conceptualizes two factors of neighboring components and distant components, and examines the boundary conditions of disciplinary interdependence and disciplinary visibility. Then, the second study investigates how this new theoretical approach effectively disentangles the debate on the relationship between search scope and knowledge impact and uncovers the benefits of knowledge cross-fertilization. Finally, the third study explores the relationship between knowledge search strategies and market impact, drawing on the literature about categories. As such, the dissertation contributes to the existing research in four principal ways:

First, the dissertation enriches the connotation of search scope by introducing the concepts of neighboring and distant knowledge components and identifying their differential effects on knowledge impact. Extant studies have produced inconclusive findings on whether search scope impedes or benefits knowledge impact. Some scholars suggest that search scope is conducive to improving knowledge impact, while others hold the opposite view. This controversy results from the literature overlooking the types of knowledge being recombined. To reconcile this debate, the dissertation scrutinizes the role of search scope and emphasizes the potential boundary conditions. Specifically, the study proposes that search scope triggers two types of novel knowledge component generation processes (i.e., neighboring components and distant components). Disciplinary interdependence and disciplinary visibility moderate the relationship between component recombination and knowledge impact. The findings show that search scope can either benefit knowledge impact through the increased use of neighboring components or impede impact through the increased use of distant components, depending on the types of knowledge components being recombined. As such, this research enriches the concept of search scope and extends knowledge search studies to a broad arena.

Second, the dissertation uncovers the benefits of cross-fertilization by investigating knowledge relationship intensity and neighboring knowledge concentration as well as the boundary condition of technological turbulence. Although the first study found that the recombination of neighboring knowledge components is beneficial to improving knowledge impact, the understanding of the effectiveness of neighboring knowledge recombination is still rather limited. This limitation lies in the abstractions of spatial separation and cross-fertilization of disparate knowledge when discussing the recombination of neighboring components. On the one hand, when the spatial separation is weak, the similarity between neighboring knowledge and specialized knowledge helps knowledge creators better understand heterogeneous knowledge, which reduces the uncertainty of resultant outcomes, but also foregoes novelty. On the other hand, the cross-fertilization of knowledge recombination may either stem from very few neighboring fields or multiple fields. Therefore, this dissertation proposes the concepts of knowledge relationship intensity and neighboring knowledge concentration to analyze the effectiveness of neighboring knowledge recombination. The findings show an inverted U-shaped relationship between knowledge relationship intensity and knowledge impact. Neighboring knowledge concentration significantly improves knowledge impact. In addition, technological turbulence negatively moderates these relationships. Therefore, this dissertation proposes and verifies the effectiveness of neighboring knowledge recombination.

Third, the dissertation integrates the knowledge search and market impact literature by examining how search scope and search depth affect the market impact of the novel outcomes. While a remarkable body of research has documented how different knowledge search strategies, such as search depth and search scope, affect knowledge impact as indicated by citation frequency, it overlooks how search strategies influence the market impact of recombinant results. Building on category-spanning literature, this study postulates that because search depth enables a novel outcome to have a clear identity, which exhibits reception benefits, search depth facilitates market impact. Conversely, search scope is likely to integrate distinct technological categories, leading to ambiguity and confusion for knowledge consumption-oriented stakeholders. As a result, a wide search scope decreases market impact. Furthermore, these relationships are contingent on knowledge relationship intensity. As such, these findings complement the existing knowledge search literature by extending the focus from knowledge impact to market impact.

Fourth, the dissertation develops a novel approach to construct knowledge networks and measure knowledge relatedness. Based on network function theory and a comprehensive examination of knowledge sources, the dissertation analyzes and constructs knowledge networks for each innovative outcome to approximate the pertinent recombinant search space. A recombinant search space is the foundation for studying knowledge search behavior. Existing research heavily relies on knowledge classification systems, such as paper and patent categories, to measure different knowledge search strategies properly. Whereas classification systems are helpful to compile, retrieve, and memorize different sets of knowledge, they overlook the potential connections among the knowledge components. Moreover, the annual update of the classification system makes knowledge components intractable and inconsistent. Therefore, this study builds on network function theory and identifies four types of knowledge relations: the intercitation relation, common-citing relation, common-cited relation, and indirect citation relation. Then, based on these knowledge relationships, the study develops a proper portrait of the recombinant search space. In so doing, the proposed method enriches the knowledge search literature from studying knowledge components to knowledge relatedness.