Sketching the Domain of Digital Innovation:
A Systematic Literature Review
Abstract:
In the new digital age, digital innovation has become one of the most important sources of competitive advantage for organizations. Accordingly, there is also a growing body of literature that research different aspects of digital innovation. Prior research addressed multiple facets of digital innovation and provided us a diverse, segmented picture on digital innovation. However, we still lack a comprehensive understanding of how different research streams fit together. Therefore, we conduct a systematic literature review to understand better what has been studied in the past decade and what we can learn from the findings. We identified 121 relevant papers in the Information Systems field and found a wide variety of theoretical perspectives discussed as well as an unusual trend regarding methodologies. Using the insights from this review, we develop two conceptual frameworks containing the digital innovation literature streams and guide future digital innovation research. Building on our findings, we propose that future research needs to focus on sustainable digital innovation, experimental methodologies, the fusion of digital innovation adoption literature, paradoxical effects, and digital innovation expertise.
Author: Dennis van Dijk Student number: S3523756
MSc BA Strategic Innovation Management Supervisor: Dr. J. Dong
Co-accessor: Dr. F. Noseleit Date of submission: 20-01-20
1. Introduction
The integration of digital technologies radically changes the architectures of products and services (Yoo et al., 2012). However, it also spurs new pathways for value creation, creation of new innovation processes, and transforming industries as a whole (Nambisan et al., 2017). Thus, digital innovation creates both opportunities and challenges for organizations (Armstrong & Sambamurthy, 1999). Seizing these opportunities may not be so easy for an organization as they may lack the capabilities or agility (Henfridsson et al., 2008; Lee et al., 2015). Hence, they need to develop new capabilities in order to initiate and develop digital innovations (Henfridsson & Yoo, 2014). Also, top management has to acknowledge the urgency to implement digital innovation for organizational improvement (Subramanian & Nilakanta, 1996; Fitzgerald et al., 2013).
This indicates the rising importance of digital innovation for both practice and academia. In the last two decades, we have seen a growing attention to a phenomenon of digital innovation both in Information Systems (IS) research (Henfridsson et al., 2008; Henfridsson & Yoo, 2014; Yoo et al., 2012) as well as for practitioners (Svahn, Mathiassen, Lindgren & Kane, 2017). Given this rising importance, both academic and practical attention to the topic of digital innovation, previous studies have addressed different topics. In these studies, the conceptualization of digital innovation ranges from a product, process, or business that is novel to the organization and requires alterations upon adopting while embodied in or enabled by information technology (Fichman et al., 2014). Recent research has increased our understanding of certain aspects of digital innovation. For example, scholars addressed digital innovation topics related to adoption & initiation (Hameed et al., 2012; Xu, 2017; Fitzgerald et al., 2013), knowledge management (Carlo et al., 2012; von Krogh, 2012), digital sustainability (Bose & Luo, 2011; Hanelt et al., 2017), implementation (Hameed et al., 2012; Ravichandran et al., 2017), organizational readiness (Lokuge et al., 2019), and organizational change (Volkoff & Strong, 2013; Orlikowski, 1996). However, while attention has risen for digital innovation, it is unclear what is known about the totality of digital innovation, how different research streams fit together, and the existence of possible worthwhile opportunities regarding (new) knowledge development (Kohli & Melville, 2019).
Lyytinen, Majchrzak, & Song, 2017). Thus, this study acknowledges this gap, as well as the concern expressed by Kohli & Melville (2019), by expanding the scope towards literature streams with a recent timeframe from 2010 to 2019. Hence, this study addresses the gap by initiating the following research question: ‘’What do we know about digital innovation at the
organizational level in the IS literature between 2010 and 2019?’’.
To address this question, we conducted a systematic literature review and identified 121 studies included in this thesis. Based on the analysis of the papers, we develop two conceptual frameworks describing two comprehensive models containing independent, dependent, moderator, and mediator variables.
This thesis has a two-fold contribution. First and foremost, we consolidate different bodies of knowledge on digital innovation in data-grounded, comprehensive frameworks consisting of independent, dependent, mediator, and moderator variables concerning digital innovation. Whereas Kohli & Melville (2019) propose a theoretical framework on research streams, we propose a different framework that sheds more light on relationships between different concepts. Therefore, future studies can easily understand what has been studied before as well as obtain constructs for their empirical analysis. Second, we identify the recent trend in digital innovation research over the last decade, which complements the literature review on earlier papers in Kohli & Melville (2019). Specifically, we provide insights into different contexts, methodologies, and theories used in digital innovation research.
This thesis is organized as follows. First, we discuss the main theoretical concepts related to digital innovation. Next, we discuss the methodology that involves a systematic literature method. After that, we discuss the main findings and present two conceptual frameworks. Following this, we discuss our findings and conclude with implications and limitations of this study as well as future research opportunities.
2. Theoretical background
2.1 IT innovation
The concept of IT innovation is described as the adoption of IT in organizations corresponding to their technological environment (Becker & Whisler, 1967). It is also defined as the embracement of IT in organizations for which IT is novel (Aiken & Hage, 1971). Therefore, IT could be already implemented within other organizations, but the implementation of IT is novel for the organization in question (Aiken & Hage, 1971). It is said that IT innovation adoption improves operational as well as strategic operations (Lee & Xia, 2006; Pervan et al., 2005). Moreover, organizations adopt IT innovations to be able to respond to changes in the external or internal environment and preemptive behavior to influence environments (Lee & Xia, 2006).
2.2 IT-enabled innovation
An IT-enabled innovation is a process and the result extending from the technology breakthrough (Ashurts et al., 2012). It is also referred to as benefits realization through or enabled by IT (Ashurts et al., 2008). For example, some research on social media suggests that new IT-enabled services change the way organizations innovate (Hagel, Seely Brown, & Davison, 2010). Specifically, it allows people to connect, and it creates new forms for sharing ideas both within and outside the organization (McAfee, 2006). Whereas IT innovation refers explicitly to the adoption of IT that is novel to the organization, IT-enabled innovation is not necessarily the technology itself that is considered to be innovative but the products, services, and processes enabled by the technologies.
2.3 Digital innovation
3. Methodology
3.1 Research design
Systematic reviews improve the quality of the review processes by enabling reproducibility and transparency (Tranfield et al., 2003). In order to provide a comprehensive and step-by-step plan to review articles systematically, we combined the approach of Tranfield et al. (2003) and Crossan and Appaydin (2010).
Specifically, we followed a three-stage procedure of Tranfield et al. (2003), namely planning, execution, and results. Furthermore, to conduct these, we also followed guidelines of Crossan and Apaydin (2010) on data collection, data analysis, and synthesis. Data collection consists of identifying relevant literature by using a pre-determined selection algorithm (e.g., pre-defined search terms). Data analysis represents the analysis of the selected literature. However, an analysis could proceed in different manners depending on the objectives. Ultimately, using synthesis, new insights are determined through data collection and careful analysis (Crossan & Apaydin, 2010).
At the first stage (i.e., planning) stage, we defined our research objective as well as the primary data source. Our objective was to address literature dating from 2010-2019. At this stage, we also ran several initial searches in the online database to get an initial understanding of the literature from different databases and journals. To ensure that we can get a manageable and useful sample, we chose the Web of Science as the primary database. This database was selected for the following reasons: (1) its comprehensiveness (over 12.000 journals available including Association for Information Systems (AIS) basket of eight journals (i.e., leading journals in IS field) and other IS journals); paper; (2) availability of Web of Science subscription via University of Groningen.
In our second stage (i.e., execution), we followed three systematic steps: (1) initial selection criteria (or inclusion/exclusion criteria) & pre-defined key- and searched terms, (2) assembly of a ‘consideration’ set, and (3) grouping of the articles. These three steps apply to data collection.
3.2 Execution: Initial selection criteria, key search terms & consideration set
As this study focuses on organizational-level as well as digital innovation, it had to meet two criteria elements before implementing: (1) organizational focus whereas the focus would rely on the field of study that involves solely organizational aspects. As Pugh & Dietz (2008) address: an organizational-level view involves no individual-level analysis of an organization (such as workgroups or departments). It is a unit above this level of analysis; (2) digital
innovation focus which relies on the definition by Fichman et al. (2014), digital innovation is a
product, process or business model that is novel to the organization and requires alterations upon adopting while embodied in or enabled by information technology.
First of all, in line with our research question, we focused only on the articles from 2010 onwards. Thus, that was our first limiter. Secondly, we searched for keywords in the title and abstract. This search was limited to solely articles that were incorporated in the Web of Science core collection1. Furthermore, we focused on finding the literature only from peer-reviewed journals and the articles that were in English. Following this, we generated, using Web of Science, two sets of articles. The first set was generated via the use of search terms "digital innovation*" OR "IT-Enabled Innovation*" OR "Information Technology Enabled Innovation*" OR "IT Innovation*" OR "Information Technology Innovation*" while focusing on the Information Management Journals list provided by Association of Business Schools (ABS). In this search, we excluded the IS basket of eight journals from the search. These search terms have been chosen given dominant conceptualizations of digital innovation described in the theoretical background. Furthermore, the choice of search terms was discussed and finalized in a meeting with the thesis supervisor. This search has resulted in 74 articles.
. The second set was generated via the use of the search term ‘’innovation*’’, and it covered only the IS basket of eight journals. We chose to have only one word and capture all articles that would contain word innovation. We made this choice as most of the literature on digital innovation is captured in these journals (Kohli & Melville, 2019). This has resulted in 432 articles. Appendix A contains all the ABS Information Management Journals that were used.
The two sets combined resulted in 506 articles for initial screening. These articles were inserted in Refworks to check for duplication. No duplicates were found.
We then used the data extraction form in order to check if the articles meet the criteria of the literature review. This data extraction form was inspired by the form of Cochrane (2014)
1 This consists of the following indexes: science citations, social sciences citations, arts & humanities citations.
as well as our input, which was thoroughly checked for completeness by the thesis supervisor. Each article was manually screened through reading the title, abstract, and keywords. If the article was not definite after this, the article was read in full to decide if it met selection criteria or not, which accounted for 168 articles. To account for potential bias in the selection of the articles and increase reliability, we asked another researcher to go through a random sample of 51 articles. Out of 51 articles, four achieved a different conclusion, which resulted in an excellent interrater reliability. These four articles were later discussed in a meeting with another researcher (i.e., coder of the random sample) and thesis supervisor. In this meeting, we achieved consensus about the four articles. One difference between these four articles was the inclusion of conceptual papers in the literature. As an outcome of the meeting, we decided to integrate conceptual articles as they create the foundation for the digital innovation literature.
After this check, we included 121 articles in the literature review that remained for the data analysis as well as synthesis. These 121 articles were further examined within the data extraction form. This extraction form contains all information that is necessary for further analysis. Appendix B displays the entire methodology procedure. Appendix C contains a list of all 121 included articles. The next section provides the procedure of data analysis and synthesis of the composed consideration set. For the grouping of articles, we captured the variables from all the quantitative articles, which resulted in the grouping of the articles. If variables show a similar overarching concept (e.g., organizational outcomes), they were categorized under the same classification (e.g., revenue and Tobin's Q in organizational outcomes). However, if variables do not show many similarities, they were individually discussed. This method has been approached to analyze whether individual research streams fit together.
3.3 Results: Data analysis and synthesis
To analyze the data, we first collected the data from each article. In particular, we collected the data on year of publication (from 2010 to today), journal of publication (name of journal), type of article (i.e., empirical or conceptual), theory used, type of methods used (qualitative, quantitative, mixed methods), research context (industry and country), the definition of digital innovation as well as details on the sample size, more details on the methods, variables included in article (independent variables, dependent variables, mediators and moderators), hypotheses and correlations, timespan of data collection, type of technology discussed and constructs provided by Kohli & Melville (2019).
of synthesizing the literature by contrasting and comparing our findings. The result of this process of the data analysis is conceptual frameworks capturing different relationships, which is presented in the findings section.
4. Results
In this section, we first present some descriptive results. In particular, provide the results below addressing the years of publication, number of articles per journal, type of methodologies used, type of theories, research context, type of digital technologies and, constructs by Kohli & Melville (2019). Furthermore, we present our conceptual frameworks containing variables and their relationships as they are included in the empirical papers.
4.1 Publications per year
Figure 4.1: Breakdown of the number of articles per year.
and it reached its lowest number in 2013 in which only six articles were published. From 2013 onwards, we observe an increasing trend except 2016.
4.2 Journals
We analyzed the number of articles in different journals. Figure 4.2 displays the proportion of studies per journal. For overview purposes, the IS Basket of eight has been displayed individually while the other journals, related to other academic information systems journals, have been categorized under other journals.
Figure 4.2: Breakdown of studies per journal.
Furthermore, we also categorized the proportion of the number of ‘other’ journals (figure 4.3). Surprisingly, the pie chart identifies that a select group of other journals express digital innovation literature. This categorization displays that Information & Management (20,83%) and Information and Organization (20,83%) are the dominant journals within this group.
Figure 4.3: Breakdown of other journals.
4.3 Methodologies
well as qualitative studies, both covered 56 (46.3%) articles each. However, we observed that the quantitative and qualitative studies become more apart in the last couple of years, meaning that there were more qualitative studies than quantitative studies.
Figure 4.4: Breakdown of methodologies.
Figure 4.5: Crosstab of Publication Date & Methodology.
From this figure, we can address that the number of qualitative studies has increased whilst the quantitative studies have decreased in the last few years. To verify this finding, we performed a scatterplot for both methodologies.
Figure 4.7: Scatterplot qualitative studies.
For qualitative studies, the scatterplot shows consistency, indicating that the number of qualitative increased over the last few years. On the contrary, quantitative studies show a negative trend, indicating that the number of quantitative studies decreased in the last few years.
4.4 Data collection method
Figure 4.8: Breakdown of data collection methods.
4.5 Theories
With regard to the theories in the consideration set, a lot of diverse theories have been adopted. Among this set, the most frequently used theories were innovation diffusion theory (10 papers), institutional theory (6 papers), absorptive capacity theory (5 papers), and disruptive innovation theory (5 papers). Moreover, resource-based view theory, agency theory, and knowledge-based view were used in three papers individually. However, all other papers addressed different theories or perspectives that were used not more than two times. This indicates rather high diversity in the use of theories in studying the topic of digital innovation. A list of all used theories can be found in Appendix D.
Figure 4.9: Breakdown of articles by theory.
Twenty-nine papers specifically addressed theory-building practices as their intention and did not employ a specific theory. This especially cases more recently (2016-2019), which is also in line with the rise of qualitative papers.
4.6 National context
Figure 4.10: Breakdown of national context.
4.7 Industry context
Figure 4.11: Breakdown of industries
4.8 Theoretical constructs by Kohli & Melville (2019)
Finally, to compare our findings systematically with those of Kohli & Melville (2019), we analyzed their constructs within our consideration set (figure 4.12). To analyze this, we used theoretical constructs presented in their theoretical framework, namely activities related to digital innovation that includes initiating, developing, implementing, exploiting, factors in the
internal organizational environment, and external competitive environment as well as outcomes
of digital innovation such as new products, services, and processes.
Figure 4.12: Breakdown of constructs adopted from Kohli & Melville (2019).
4.9 Digital technologies
Figure 4.13: Breakdown of technologies.
5. Conceptual frameworks
Drawing from the initial coding, we developed two models in which we framed and interpreted the research on digital innovation. This method has been chosen as it brings a presentation of all research embodied in an overview. Future researchers can interpret the model with relative ease and continue with empirical methods. In order to be able to come up with such models, we excluded qualitative studies from this stage of analysis. This is because it is difficult to identify the variables in qualitative studies.
5.1 Independent variables
were only individually examined as they could not be grouped into a meaningful category. We present and illustrate the following group of independent variables: organizational characteristics, digital innovation embeddedness, type of innovation, managerial characteristics, knowledge management, and capabilities. Each variable will be discussed via its provided central concept as well as the central argument. After this, we have presented the list of uncategorized variables.
Organizational characteristics. Organizational characteristics is a category that
encompasses the organizational influences on digital innovation. This category includes concepts as IT expertise (Hameed et al., 2012; Rondovic et al., 2019), organizational readiness (Lokuge et al., 2019; Rondovic et al., 2019), organization size (Hameed et al., 2012; Im et al., 2013), IT affordance (Chatterjee et al., 2015), R&D use (Kleis et al., 2012; Bardhan et al., 2013), organizational resources (Ordanini & Rubera, 2010), innovative climate (Leidner et al., 2010), diversification (Banker et al., 2011), IT Unit (Tarafdar & Tanriverdi, 2018), innovation
capacity (Ravichandran, 2018), knowledge dispositions (Gaskin et al., 2018) and IT strategy
(Leidner et al., 2011; Xu et al., 2014). IT expertise refers to the ability of the staff to contribute via their knowledge pool to develop and use digital innovation (Hameed et al., 2012; Rondovic et al., 2019). Organizations that have implemented IT-specialists adopt digital innovation more promptly and develop requirements for the effective use of digital innovation (Rondovic et al., 2019). Organizations should stimulate IT expertise as it could benefit the entire implementation process of digital innovation (Rondovic et al., 2019). Organizational readiness is determined by Lokuge et al. (2019) as: ‘’an organization’s assessment of its state of being prepared for
effective production or adoption, assimilation and exploitation of digital technologies for innovation, p. 446’’. Lokuge et al. (2019) treat organizational readiness for digital innovation
as a dependent variable but list various readiness antecedents that directly influence readiness (e.g., IT readiness, resource readiness, strategic readiness). Rondovic et al. (2019) address that readiness enhances the adoption and diffusion of IT within organizations. Furthermore, it could also be paired with IT expertise to enhance the positive influence (Rondovic et al., 2019).
Organization size refers to the number of employees within an organization at any given point
direct positive relationship between organizational size and digital technology adoption. In addition, Hameed et al. (2012) address that large organizations tend to adopt more digital innovation, which strengthens the posits by Im et al. (2013). IT affordance refers to the possibilities for ambitious actions recognized by a community (Chatterjee et al., 2015). IT affordance reflects the goals of the community and how IT is appropriated by the community to realize these goals (Chatterjee et al., 2015). Chatterjee et al. (2015) address that IT affordances affect improvisational capabilities, which affect organizational innovativeness via organizational virtues. Chatterjee et al. (2015) propose three sub-components of IT affordances: collaborative IT affordance, organizational memory affordance, and process management affordance. Amongst all different sub-components, collaborative IT affordance shows the highest impact as organizations engage in coordinated conversation. Therefore, collaborative IT affordances are important in stimulating collective thinking and wisdom (Chatterjee et al., 2015). R&D use represents the expenses of R&D per year (Kleis et al., 2012; Bardhan et al., 2013). Bardhan et al. (2013) argue that R&D, through interaction and enablement with digital technologies, improve firm value. This is caused by IT-enabled routines that leverage internal and external knowledge, which enhances the R&D process (Bardhan et al., 2013). Similarly, Kleis et al. (2012) predict that the interaction effect could stimulate innovation output. Both articles argue that R&D, in combination with IT, stimulates the effect on outcomes.
Organizational resources are defined by Ordanini & Rubera (2015) as resources that are
organization operates. If an organization meets the decision to either acquire or invest in R&D, diversification simplifies this decision (Banker et al., 2011). IT unit is an IT function that features all IT employees and has both decision rights as well as responsibility for an organization's IT resource (Tarafdar & Tanriverdi, 2018). Tarafdar & Tanriverdi (2018) reason that IT units can sense whether technologies are relevant and useful for new product innovations as well as enhance the matching of technologies due to enhancement of the organization's capacity to imagine a wide range of product innovations. Furthermore, IT units also support the organization's relative knowledge and with the aid of technology standards and can help the organization to improvise. Innovation capacity is determined as an organization's ability to produce new products, services systems or processes over an extant period of time (Ravichandran, 2018). This variable measures the innovativeness of an organization. Ravichandran (2018) show that innovation capacity directly influences organizational agility due to the contribution to the flexibility for reconfiguration of resources, which enhances agility. Knowledge dispositions refer to a bundle of preferences, behaviors, and cognitive models (Gaskin et al., 2018). Gaskin et al. (2018) address that knowledge dispositions are drivers for innovation in software organizations as they can express who and what of organizational learning, which influences innovation outcomes. Gaskin et al. (2018) also address that they tend to stick to one innovation at a time and neglect the others. Regarding
organizational characteristics, literature shows that this category acts as an antecedent for
outcomes such as innovations, improved firm value, digital innovation adoption, and diffusion.
Digital innovation embeddedness. Digital innovation embeddedness is another category
that encompasses IT-related attributes such as adoption, diffusion, and use. This category is composed of the following variables: digital innovation use (Pavlou & El Sawy, 2010; Setia et al., 2011; Kuegler et al., 2015; Dong & Yang, 2019), IT intensity (Mueller et al., 2018), digital
innovation ambidexterity (Lee et al., 2015), IT integration (Rondovic et al., 2019), adoption rate (Carlo et al., 2014; Escobar-Rodriguez & Romero-Alonso, 2014), IT investments (Mani et
makes a knowledge base more explicit and easier to use. Concept of IT investments is operationalized in many ways: IT hardware in organizations (Kleis et al., 2012), IT annual budget to the organization's annual sales (Havakhor et al., 2019), insourcing and outsourcing of IT (Qu et al., 2010), investment in IT portfolio (Xue et al., 2012), Organizational-level IT spending (Bardhan et al., 2013), investment orientation (Ravichandran et al., 2017) and investment in IT fashion (Wang, 2010). As mentioned above, Kleis et al. (2012) and Bardhan et al. (2013) both in organizational characteristics mention the interaction effect between IT investments and R&D use. Additionally, Havakhor et al. (2019) also describe the interaction effect of IT and R&D investments. Similarly, Havakhor et al. (2019) & Bardhan et al. (2013) used the dependent variable: Tobin's Q. Apart from Tobin's Q, Xue et al. (2012) motivate that (explorative) IT investments are aimed at enhancing innovation by accelerating the development of new product and process innovations. They also motivate that (exploitative) IT investments aim at enhancing efficiency (Xue et al., 2012). A different aspect of IT investments is insourcing and outsourcing of IT (Qu et al., 2010). They describe that both insourcing and outsourcing have an impact on the development of IT-enabled business processes (Qu et al., 2010). The outsourcing of digital innovation can result in IT capabilities (Mani et al., 2010). IT
embeddedness shows that digital technology can generate several outcomes capabilities,
knowledge recombination, innovation outputs, and business value. Furthermore, the characteristics of adopters have also been addressed where controversial arguments address that they early adopters quicker adopt in the initial stage. In contrast, late adopters adopt a higher adoption rate in mature stages of innovations.
Type of innovation. The type of innovation consists of elements that refer to descriptions
that measure any given innovation (e.g., incremental innovation and architecture). After review, this category consists of two variables: Disruptive IT innovation (Lui et al., 2016) and level of
radicalness (Carlo et al., 2011). Disruptive IT innovation is defined as an architectural
radicalness refers to the uniqueness and originality of innovations (Carlo et al., 2011). Carlo et
al. (2011) address that the level of radicalness of base innovations is an antecedent for the level of radicalness of both process and service innovations. This indicates that both Lui et al. (2016) and Carlo et al. (2011) address that disruptive IT innovations are antecedents for (radical) process innovations when adopting disruptive IT innovations.
Managerial characteristics. Managerial effects involve managerial influences on digital
innovation. This category consists of the following variables: Top management attitude (Leidner et al., 2010; Hameed et al., 2012; Moghavvemi & Salleh, 2016), managerial
perception (Stratopoulos & Lim, 2010; Sharma & Rai, 2015), job tenure (Sharma & Rai, 2015), type of leaders (Lu & Ramamurthy, 2010), Role of CIO’s (Leidner et al., 2010), decision factors
digital innovation opportunities that a stable environment supports (Lu & Ramamurthy, 2010).
The role of Chief Information Officer (CIO) refers to the responsibility and impact a chief
information officer has on digital innovation use within an organization (Leidner et al., 2010). Leidner et al. (2010) argue that CIO's shape the digital innovation agenda by providing thought leadership to TMT, which makes them aware of possible digital innovation support that could enhance organizational strategies. Decision factors concern the decision that involves the adoption and diffusion of digital innovation by organizations (Li et al., 2011). Li et al. (2011) formulate a theoretical framework including three elements for decision factors: entity factors,
object factors, and contextual factors. These elements affect the intention to adopt or continue
to use for digital innovation (Li et al., 2011). Some examples of commonly discussed variables:
perceived usefulness, compatibility, ease of use (Li et al., 2011). The social capital of the founders refers to social networks and personal relationships (Spiegel et al., 2016). In this
article, they couple social capital with the initiation of an IT-start up, whereas they argue that social capital could be particularly crucial due to the availability of knowledge, resources, and investors (Spiegel et al., 2016). Spiegel et al. (2016) proved that social capital indeed permits the success of start-ups. Managerial obstacles point out the obstacles that disable managers from conducting in digital innovation (Picoto et al., 2014). Some examples of obstacles are insufficient top-management support, lacking capable staff, poor integration with existing digital innovation and processes (Picoto et al., 2014). As addressed in the literature (e.g., Basole, 2007; Leidner et al., 2010), top management support is essential for the adoption and diffusion of digital innovation. If that support is insufficient, it will decrease the ability to either adopt or diffuse potential digital innovation, which is compatible with the arguments of Picoto et al. (2014).
Knowledge management. In this review, knowledge management answers to the
due to the exploitation and transformation of external knowledge (Joshi et al., 2010; Cooper & Molla 2017). Moreover, Cooper & Molla (2017) took it a step further and explained that these innovations also impact organizational performance due to improved sales. In addition, Trantopoulos et al. (2017) confirm that deep digital-enabled external search positively affects service innovation. The knowledge base is defined as the knowledge level an organization possesses at any point in time (Roberts et al., 2017). Roberts et al. (2017) argue that a knowledge base consists of detailed knowledge and knowledge diversity. Related knowledge helps an organization to acquire and assimilate external search with more ease as its familiar (or related) to the organization's knowledge base (Roberts et al., 2017). Whereas, knowledge diversity is essential for external knowledge searching as a diverse knowledge base increases the possibility that external knowledge will relate to what is already known by the organization (Roberts et al., 2017). They imply that the knowledge base helps to assimilate digital innovation (Roberts et al., 2017).
Capabilities. Within this review, capabilities consist of two variables: capability factors
(Bui et al., 2019), dynamic capabilities (Karimi & Walter, 2015), and IT. Capability factors consist of two levels: service-level strategy and supplier strategy (Bui et al., 2019). They address that capability factors, together with governance structures, enable IT outsourcing (Bui et al., 2019). In the article by Karimi & Walter (2015), dynamic capabilities extend, modify, change, or create organizational capabilities that enable responsiveness. They address that dynamic capabilities enable responsiveness to digital disruption (Karimi & Walter, 2015); they measured this through the response performance of newspaper organizations.
Individual variables. We have earlier discussed all categories for the independent variables.
We now address each variable that could not be easily categorized. These variables consist of:
collaboration asymmetries (Michalski et al., 2014), VC investments (Breznitz et al., 2018), business process reengineering (Altinkemer et al., 2011), product stack composition (Han et
strategies) the organization that they invest in (Breznitz et al., 2018). Business process
reengineering is the redesign of business processes, through the integration of IT, in order to
gain significant improvements (such as efficiency, quality, service) (Altinkemer et al., 2011). When organizations integrate BPR projects, organizations first experience a drop in their performance as implementation problems arise (e.g., lack of communication) (Altinkemer et al., 2011). However, after initiation BPR will significantly improve organization performance and productivity as employees become more comfortable with the processes (Altinkemer et al., 2011). Controversially, the positive effect will decrease over time (Altinkemer et al., 2011).
Product stack composition indicates the proportion of hardware, software, and application
software in the IT portfolio (Han et al., 2013). Han et al. (2013) specifically address the importance of product stack composition related to service expansion. Product stack can deliver customer knowledge, which is needed in order to participate in service expansion (Han et al., 2013). In the article by Kung et al. (2015), environmental pressures consist of three elements (coercive pressures, mimetic pressures, and normative pressures). Kung et al. (2015) express environmental pressures as (another) antecedent for digital innovation adoption. Under high coercive pressures, organizations tend to quickly adopt new structures due to the enforcement of the external environment (e.g., customers). Under mimetic pressures, organizations tend to follow fads and fashions. In contrast, they mimic an already existing (and working) structure, under normative pressures, organizations adopt due to the bandwagon effect of other organizations already adopted the innovation (Kung et al., 2015). Customer involvement is defined as the extent to which an organization interacts with the customer in the product development process (Saldanha et al., 2017). As a result, customer involvement can enhance the generation of innovations (Saldanha et al., 2017). Organizations could harvest information, opinions, and feedback from customers to implement (eventually) in product development (Saldanha et al., 2017). Work practices refer to different perceptions of employees in different roles in organizations (Avgar et al., 2018). They suggest that work practices complement digital innovation investments by facilitating workforce learning (Avgar et al., 2018).
5.2 Dependent variables
The same structure for independent variables is applied towards dependent variables. We first begin by describing the categories with their variables and close with individual variables that could not be characterized. We categorized dependent variables into the following groups:
Organizational outcomes. Organizational outcomes refer to the outcomes such as
productivity, revenue, and Tobin's Q. Variables that are included within this category are
revenue (Lu & Ramamurthy, 2010; Ordanini & Rubera, 2010; Qu et al., 2010; Wang, 2010;
Altinkemer et al., 2011; Leidner et al., 2011; Setia et al., 2011; Michalski et al., 2014; Dong & Wu, 2015; Karimi & Walter, 2015; Chuang & Lin, 2017; Cooper & Molla, 2017), Tobin’s Q (Qu et al., 2010; Altinkemer et al., 2011; Bardhan et al. 2013; Shama & Rai, 2015; Havakhor et al., 2019), organization productivity (Altinkemer et al., 2010; Picoto et al., 2014; Mueller et al., 2018), efficiency (Xue et al., 2012), customer loyalty (Xu et al., 2014), service expansion (Han et al., 2013), organizational readiness (Lokuge et al., 2019), task performance impact (Kuegler et al., 2015) and organizational agility (Lee et al., 2015; Ravichandran, 2018).
Revenue answers to the organization's gain, such as return on assets and sales. This variable is
Indicating that all studies define productivity as 'labor productivity,' it can be addressed that this is a common method of measure for productivity. It can also be stated that all studies address digital innovation as their antecedents (e.g., big data analytics (Mueller et al., 2018)).
Efficiency refers to an organization's ability to manage its supplier relationship, internal
operations, and customer operations (Xue et al., 2012). Xue et al. (2012) measure this outcome based on four measures of efficiency: (1) payable turnovers, (2) inventory turnovers, (3) receivables turnover, (4) sales and costs. The study of Xue et al. (2012) is focused on IT investments as its antecedent, which, as determined above, should result in more efficiency. According to Xu et al. (2014), customer loyalty consists of three phases: (1) cognitive loyalty, (2) affective loyalty, (3) conative loyalty. Xu et al. (2014) also express the need for coupling IT innovations with customer loyalty. Service expansion is referred to as the ability of IT companies to develop service revenue (Han et al., 2013). Han et al. (2013) address several antecedents (e.g., performance in the product business, product stack composition, and product stack scope). They express that an IT's organization product portfolio is associated with service expansion (Han et al., 2013). Service expansion is measured as a ratio of service revenue to total revenue (Han et al., 2013). Organizational readiness is already determined in the
independent variable section. Different from Rondovic et al. (2019), Lokuge et al. (2019) treat
organizational readiness as a dependent variable as they explain the antecedents (such as IT readiness). Task performance impact refers to productivity, job performance, and job effectiveness (Kuegler et al., 2015). They identify the use of a digital innovation (ESSR) in teams (intra and inter) as the antecedent for task performance (Kuegler et al., 2015). They measure task performance impact through quicker task performance, job effectiveness, job performance, and productivity (Kuegler et al., 2015). Organizational agility enhances the ability to respond to environmental changes and opportunities in a more promptly, accurate, and cost economic manner (Lee et al., 2015; Ravichandran, 2018). Both articles express digital innovation as an antecedent to enabling agility (Lee et al., 2015; Ravichandran, 2018). Lee et al. (2015) measured their agility based upon the proactiveness, responsiveness, and radicalness abilities. Moreover, Ravichandran (2018) measured agility via customer responsiveness, operational flexibility, and strategic flexibility.
Digital innovation practices. Digital innovation practices encompass all digital
digital innovation assimilation (Roberts et al., 2017), digital innovation sustainability outcomes
(Cooper & Molla, 2017). Digital innovation adoption indicates that digital innovation has been selected for implementation by the organization (Hameed et al., 2012; Kung et al., 2015). Hameed et al. (2012) address that there are many antecedents for the eventual adoption of digital innovation. However, Li et al. (2011), Sharma & Rai (2015), Kung et al. (2015) describes prior steps regarding adoption, namely: intention to adopt (Li et al., 2011; Kung et al., 2015) and adoption decision (Sharma & Rai, 2015). Sharma & Rai (2015) address that an adoption decision is an antecedent for adoption, which is influenced by numerous factors (e.g., environmental, individual). Li et al. (2011), Kung et al. (2015) & Moghavvemi et al. (2016) acknowledges the intention to adopt. To measure this dependent variable, researchers could examine whether participants would intent or plan to adopt digital innovation (e.g., Li et al., 2011). Hameed et al. (2012) takes it a step further and discusses the antecedents for actual adoption instead of the intention or decision. According to Rondovic et al. (2019), digital
innovation diffusion indicates the application of digital technology. They test digital innovation
diffusion through three streams of antecedents (technology, environment, and organization) (Rondovic et al., 2019). Rondovic et al. (2019) use similar organizational characteristics variables as tested by Hameed et al. (2012) for digital innovation adoption, which shows consistency. Digital innovation assimilation refers to the incorporation of digital innovation within organizations. Within assimilation, there are two categories: full and partial assimilation (Roberts et al., 2017). Full assimilation addresses the entire cycle (awareness, adoption, implementation, and deployment), whereas partial assimilation refers to only post-adoption stages (implementation and deployment) (Roberts et al., 2017). Like its antecedents, Roberts et al. (2017) solely focused on the knowledge base. According to Cooper & Molla (2017), digital
innovation sustainability outcomes refer to what extent sustainable technologies and practices
have been assimilated both to improve the sustainability of the infrastructure as the organization as a wider lens.
Innovation outcomes. Innovation outcomes refer to outcomes such as innovation output,
process innovation performance, and commercialized innovation. The included variables are
innovation output (Joshi et al., 2010, Carlo et al., 2012; Kleis et al., 2012; Carlo et al., 2014;
Sharma & Rai, 2015; Gomez et al., 2017; Saldanha et al., 2017; Ravichandran et al., 2017; Breznitz et al., 2018; Gaskin et al., 2018; Dong & Yang, 2019) and process innovation
in patents). A substantial amount of studies argued that innovation as an outcome but did not necessarily specify the outcome in general. Numerous papers address 'innovation output' and had patents as their primary observable output indicator (e.g., Kleis et al., 2012; Sharma & Rai, 2015; Gomez et al., 2017; Ravichandran et al., 2017; Saldanha et al., 2017; Dong & Yang, 2019). However, Gomez et al. (2017) identify both patents and the number of product innovations introduced by an organization as an observable output indicator. The review also showed four studies that identified a different innovation output indicator. These studies are the study by Carlo et al. (2012) which measured for number of internet computing innovations, the article by Breznitz et al. (2018) as they measure product innovation based on keywords and a text-mining tool in new product introduction articles, and the study by Gaskin et al. (2018) as they measured level of innovation through the measured number of web-based innovations (base, service, and process) developed and adopted by respective SDOs. At last, the study by Joshi et al. (2010) also measured commercialized organizational innovation through the use of newspaper databases. Process innovation performance is a dependent variable use by Trantopoulos et al. (2017). They define process innovation as the first-time introduction of novel or enhanced technologies for the production of products and services (Trantopoulos et al., 2017). Subsequently, they develop a process innovation performance measure based upon cost reduction achieved due to process innovation.
Capabilities. Capabilities refer to the capability elements in the stream of studies that
have been reviewed. This category consists of four different capabilities, namely:
improvisational capabilities (Pavlou & El Sawy, 2010), operational capabilities (Pavlou & El
measured managerial via the use of measures by Sethi et al. (2001). Capability enhancement is an outcome determined in the article by Bui et al. (2019) is the improvement of core competencies within organizations (e.g., access to skilled labor and refocus on core business). Bui et al. (2019) address that capability enhancement is frequently discussed in the IT outsourcing literature.
Individual factors. We have discussed all categories for the dependent variables that
could be categorized. We continue by addressing individual dependent variables that could not be categorized. The variables that will be discussed are support request (Avgar et al., 2018),
collaboration satisfaction (Bala et al., 2017), BPO performance (Mani et al., 2010), contract structure (Susarla et al., 2010; Susarla & Barua, 2011), abnormal cost of equity (Lui et al.,
2016), the extent of B2B markets (Mishra & Agarwal, 2010), internet-enabled interfirm
communication (Bell et al., 2012), propensity of a firm to in acquisitions over R&D (Banker et
al., 2011), success of early start-ups (Spiegel et al., 2016). Support requests refer to the number of support requests numbers issued each month (Avgar et al., 2018). They focus on end-user learning (or workforce learning) as they try to measure whether digital innovation implementation and work practices complement each other by facilitating workforce learning. Their measurement, however, does have a limitation: they cannot distinguish whether data from declining support requests means greater familiarity with digital innovation (Avgar et al., 2018).
Collaboration satisfaction refers to the appraisal of collaboration capabilities enabled by digital
innovation used to achieve NPD-related tasks (Bala et al., 2017). Therefore, Bala et al. (2017) issue that this assessment was an accurate representation of digital collaboration technology in an organization. Bala et al. (2017) used a five-item scale to determine collaboration satisfaction.
BPO performance is addressed in the article by Mani et al. (2010), whether the fit between
(2010) discuss an organization that seeks services from service providers, and Susarla & Barua (2011) discuss a service provider who distributes its services to organizations. The abnormal
cost of equity capital refers to the abnormal cost of equity capital of organizations over a while
(Lui et al., 2016). They addressed the impact of disruptive digital innovations on the cost of equity while moderating for the chief executive officer (CEO) incentives and institutional pressures (Lui et al., 2016). To measure the abnormal cost of equity capital, the conceptualization of abnormal performance was adopted, which is the ratio between actual performance and expected performance (Lui et al., 2016). The extent of B2B market use refers to the use of internet-based markets that bring buyers and sellers together on a virtual platform to exchange information, products, services, and payments (Mishra & Agarwal, 2010). This variable is measured by the total percentage of purchases that organizations make on B2B markets (Mishra & Agarwal, 2010). Internet-enabled interfirm communication is defined as the extent to which an organization uses the internet to enable interactions with current and prospective customers (Bell et al., 2012). Bell et al. (2012) determine that internet-enabled interfirm communication is a separate element of the digital innovation strategy of an organization. Bell et al. (2012) measured internet-enabled interfirm communication by means of a 5-point scale with questions related to how organizations value information and knowledge sharing via the use of the internet. The propensity of a firm to in acquisitions over R&D refers to the digital innovation strategy of organizations that decide whether they acquire or invest in R&D (Banker et al., 2011). This strategy is adopted as organizations have to respond to the potential threat of new IT entrants (Banker et al., 2011). This variable is measured via a PROP value that is determined as the total sum of investments in acquisition divided by the total sum of investment in acquisition plus investments in R&D. In contrast, a higher PROP-value determines that an organization employs more acquisitions than R&D and vice versa (Banker et al., 2011). The success of early start-ups refers to the degree of whether a start-up has received funding (Spiegel et al., 2016). The variable is measured via a binary approach, whereas one means that they received funding, and 0 means that no funding has been acquired (Spiegel et al., 2016).
5.3 Moderators
The moderators consist of two categories and individual variables, which will be elaborated accordingly. Moderators have been categorized in the following groups: environmental
conditions and organizational characteristics. Other variables that could not be categorized are
Environmental conditions. This category encompasses all variables that determine an
external environmental influence. Variables included are environmental dynamism (Lu & Ramamurthy, 2010; Leidner et al., 2011; Xue et al., 2012; Lee et al., 2015; Chuang & Lin, 2017), competition (Mueller et al., 2018) and stable environment (Lu & Ramamurthy, 2010).
Environmental dynamism represents the unpredictability and speed of change from the external
Organizational characteristics. Organizational characteristics enclose variables that
determine organizational influence. This category is comprised of: innovation capacity (Ravichandran, 2018), organizational capabilities (Mishra & Agarwal, 2010), perceived
complexity (Kung et al., 2015), and performance in product business (Han et al., 2013). Innovation capacity is determined as a moderator in the study by Ravichandran (2018). Highly
innovative organizations are more likely to learn and engage in experimentation resulting in coping with high uncertainty and taking more risks (Ravichandran, 2018). This reason, accompanied by many more, ensures Ravichandran (2018) that it moderates the effect between digital innovation and agility. This moderator was measured via a 5-item scale that has been adapted from a scale proposed by Hurley & Hult (1998). Organizational capabilities refer to concepts as technological opportunism and technological sophistication (Mishra & Agarwal, 2010). Mishra & Agarwal (2010) indicate that opportunism refers to a sense and response capability of an organization with respect to new technologies, and sophistication refers to the value-added ways of digital innovation. Mishra & Agarwal (2010) acknowledge that organizational capabilities reinforce the relationship between technological frames and the extent of B2B markets use. The conceptualization of organizational capabilities has also been addressed for measurement practices using 4-item scales (Mishra & Agarwal, 2010). Perceived
complexity is the degree of the perceived difficulty in understanding and use of digital
innovations (Rogers, 1995; Kung et al., 2015). It moderates the effect between environmental pressures and intention to adopt digital innovation in the article by Kung et al. (2015). Organizations that perceive high complexity tend to imitate other organizations more repeatedly hence predicting the effect between mimetic pressures and adoption (Kung et al., 2015). They adopted perceived difficulty, difficulty in understanding from a business, and technological perspective as measurements for intention to adopt (Kung et al., 2015). Performance in product
business moderates the effect between product-market experiences and an organization's
service expansion. Accordingly, poor business performance will make the resource-based effect stronger (Han et al., 2013). Sales growth is the measurement that has been adopted (Han et al., 2013).
Individual variables. The non-categorized variables are listed as follows: chief executive officer incentives (Lui et al., 2016), task equivocality (Kuegler et al., 2015), innovation type
(Roberts et al., 2017), assimilation stages (Roberts et al., 2017), information processing
executive officer (CEO) incentives moderate the effect between the cost of equity and disruptive
digital innovation as incentives can, e.g., reduce opportunism. Therefore, CEOs with higher incentive-based compensations acquire lower cost of equity when adopting disruptive digital innovation (Lui et al., 2016). Task equivocality is defined as the extent to which the tasks that an employee performs are non-routine (Kuegler et al., 2015). If task equivocality is low, employees' tasks are mostly routine-based and standard procedures, whereas high equivocality demands a higher need for information. Therefore, they address that task equivocality has a moderating effect between digital innovation use (ESSP) and task and innovative performance impacts (Kuegler et al., 2015). They adopted the measurement scale by Campion et al. (1993).
Innovation type refers to the type of innovation discussed (Roberts et al., 2017). Roberts et al.
(2017) address that digital innovation assimilation differs per innovation (e.g., e-business or software process innovations) as different innovations desire different resources (e.g., external knowledge). Therefore, they address that innovation type moderates the relationship between the knowledge base and digital innovation assimilation (Roberts et al., 2017). Assimilation
stages refer to full and partial assimilation of digital innovation (Roberts et al., 2017).
Therefore, as stages might differ, Roberts et al. (2017) argue that additional variance between the knowledge base and digital innovation assimilation when measuring for assimilation stages.
Information processing capability consists of both relation information processing capability
(RIPC) and analytical information processing capability (AIPC) (Saldanha et al., 2017). RIPC moderates the relationship between product-focused customer involvement and the amount of organization innovation as RIPC enhances the ability to manage customer relationships, identify personalized needs, and identify appropriate customers (Saldanha et al. 2017). AIPC moderates the relationship between information-intensive customer involvement and the amount of organization innovation as AIPC helps the organization to manage information overload and combine external with internal information (Saldanha et al., 2017). They measure RIPC as customer service and support, customer loyalty, product marketing, and personalized marketing and AIPC as data mining and data warehousing in organizations (Saldanha et al., 2017). Individual characteristics refer to gender, age, and propensity to act (Moghavvemi & Saleh, 2016). With this moderator, Moghavvemi & Saleh (2016) test whether the relationship between an entrepreneur's intention to use and, e.g., performance expectancy will be affected.
IT-enabled social integration capacity (IT-SIC) refers to the ability to help augment
connect and understand market needs. This shared understanding increases the change for organizations to commercialize their inventions (Joshi et al., 2010). News announcements, related to conferencing or messaging, are measured for IT-SIC (Joshi et al., 2010).
5.4 Mediators
The mediators in the articles we studied were extremely diverse. Due to this, it was not possible to categorize them at a higher level as it was the case with other categories of variables presented earlier. Thus, all mediators are individually addressed. These mediators are: trust (Michalski et al., 2014), IT application architecture synergies (Setia et al., 2011), routines (Carlo et al., 2012), community of practice for learning orientation (Bell et al., 2012),
coordination costs (Im et al., 2012), brand equity (Xu et al., 2014), operational ambidexterity
(Lee et al., 2015), information-value offering (Chuang & Lin, 2017), knowledge recombination (Dong & Yang, 2019). According to Michalski et al. (2014), trust refers to the partner's reputation and dependence, the propensity of information sharing, the partner's honesty, and charity. Michalski et al. (2014) determine trust as a variable that is critical for a successful relationship with partners. They use trust as a mediator between asymmetries and innovation, plus asymmetries and results (Michalski et al., 2014). IT application architecture synergies refer to the interaction between architecture spread as well as architecture longevity (Setia et al., 2011). It serves as a mediator to organizational performance (Setia et al., 2011). It is measured by satisfying three ratios: overall application architecture growth to spread, overall architecture growth to longevity, and overall application growth across domains (Setia et al., 2011). Routines involve the behavioral dimension of what an organization does (Carlo et al., 2012). In the article by Carlo et al. (2012), they measure two types of routines: sensing and experimentation. Sensing affects how organizations acquire external knowledge utilizing scanning and focused search (Carlo et al., 2012; Cohen & Levinthal, 1990). Experimentation refers to the ability to increase the accuracy of causal relationships between organizational actions and outcomes (Carlo et al., 2012; Brown and Eisenhardt, 1997). Carlo et al. (2012) measured sensing through the four-item construct of Srinivasan et al. (2002) and measured experimentation employing the 5-item construct by Brown & Eisenhardt (1998). Community of
practice for learning orientation is determined as group of professionals who informally share
costs increase when the organization grows in size (Im et al., 2012). Coordination costs have been measured by the use of SG&A expenses, which can be categorized in selling, administrative, and general expenses (Strassmann, 1999; Im et al., 2012). Coordination costs mediate the effect between organization size and digital innovation use, indicating that when coordination costs increase, an organization adopts more digital innovation (Im et al., 2012).
Brand equity is defined as consumer identification with the focal brand and the relevance of the
brand to their circumstance (Johnson et al., 2006; Xu et al., 2014). Brand equity mediates the effect between ICT service innovation strategy and customer loyalty (Xu et al., 2014). Brand equity is measured via the use of measurement scales of Johnson et al. (2006). Operational
ambidexterity is characterized as an organization’s ability to simultaneously pursue operational
exploitation as exploration (Lee et al., 2015). In the article by Lee et al. (2015), operational ambidexterity mediates the effect of IT ambidexterity on agility. As operational ambidexterity is an antecedent for agility, which is highly influenced by IT ambidexterity. Lee et al. (2015) proposed their own measured scales for measuring operational ambidexterity. Information
value offering is addressed in the article by Chuang & Lin (2017), who determines the
antecedents of information value offering (e-service capability and service innovation orientation) and the consequences (customer relationship performance and organizational performance). Therefore, in their conceptual model, they determine information-value offering as a mediator (Chuang & Lin, 2017). They measured information value offer based on three concepts (useful service, know-how service, and timely service) that have been adopted from prior research (e.g., Ulaga & Eggert, 2006). At last, knowledge recombination refers to two concepts, namely: knowledge recombinant intensity and diversity (Dong & Yang, 2019). These two concepts mediate the effect that IT use has on patent quality, patent quality breadth, and patent quality depth (Dong & Yang, 2019).
Figure 5.2: Conceptual framework with moderators.
6. Discussion
In this section, we discuss the findings of our thesis in light of our research question and objective of this study. Following this, we further discuss main implications of our thesis and propose some directions for future research.
Independent
variables
Organizational characteristics Digital innovation embeddedness Type of Innovation Managerial characteristics Knowledge management Capabilities Collaboration asymmetries VC investments Business process reengineering Customer involvement Work practices Environmental pressures Product stack compositionDependent
variables
Organizational outcomes Innovation outcomes Digital innovation practices Capabilities Support request BPO performance Abnormal cost of equity Internet-enabled interfirm communication B2B markets use Collaboration satisfaction Acquisition propensity Start-up success Contract structure Moderators Environmental conditions Organizational characteristicsThe primary objective of this research was to collect the different digital innovation research streams and combine these research streams in order to create possible coherence which could spur new future research directions hence the research question: ‘’What do we
know about digital innovation at the organizational level in the IS literature between 2010 and 2019?’’. In an effort to answer this research question, a systematic review has been conducted.
This review has resulted in two conceptual frameworks presented earlier that contain the variables discussed in empirical papers.
6.1 Main findings and implications
In our analysis, we first examined the descriptive statistics based on the consideration set. This analysis had a few main findings. First, we observed that in the last few years, there is an increasing trend in a number of articles that engage in qualitative research. This trend could occur due to the complex and changing nature of digital innovation that still requires theory-building studies. Thus, digital innovation research is far from saturated as new concepts are introduced in an increased manner (Colquitt & Zapata-Phelan, 2007). Another interesting finding is a very diverse set of theories that are used to study digital innovation. Only a handful of papers used similar theories such as innovation diffusion theory (Bose & Luo, 2011; Li et al., 2011; Escobar-Rodriguez & Romero-Alonso, 2014; Michalski et al., 2014; Oni & Papazafeiropoulou, 2014; Picoto et al., 2014; Kung et al., 2015; Sharma & Rai, 2015; Overby & Ransbothan, 2019; Rondovic et al., 2014). This finding is in line with the one of Kohli & Melville (2019), who also concluded that those scholars do not show a unified perspective in studying digital innovation. Furthermore, our findings indicate that most of the research focuses on outcomes. In contrast, Kohli & Melville (2019) found that the activities of developing digital innovation were a dominant subject in digital innovation research up to 2010. Therefore, there seems to be a shift of subjects within digital innovation research. Analogous to Kohli & Melville's (2019) framework, the activity of implementation is still a highly discussed subject in digital innovation research.
and digital innovation positively affects organizational outcomes, which is measured as Tobin's Q. In another study, Ordanini & Rubera (2010) found that organizational resources (slack resources, external ties, and innovative orientation culture) positively affect organizational performance, which is measured as, e.g., return on assets and sales. Interestingly, this indicates that here is coherency regarding the measurement of organizational performance. Surprisingly, other indicators such as efficiency (Xue et al., 2012) or agility (e.g. Lee et al., 2015) were not used as much. Hence, scholars tend to focus more on turnover measurements than other business values.
Apart from organizational performance, innovation outcomes are also one of the main discussed dependent variables. Within innovation outcomes, innovation output was the dominant indicator. This variable is (usually) measured by means of patent quantity (e.g., Kleis et al., 2012). Interestingly, few studies (Gomez et al., 2017; Breznitz et al., 2018) determined their innovation outcome as measured via product innovations introductions. This addresses that the popularity of measuring innovative performance is done through patents as an output indicator.
Furthermore, we found that studies rarely examined the diffusion and assimilation of digital innovation as a dependent variable (Roberts et al., 2012). Assimilation, as an element of absorptive capacity, has been addressed as an independent variable. Interestingly, digital innovation adoption research has received relatively more attention (e.g., Hameed et al., 2012; Kung et al., 2015). This shows opportunities for future research directions as adoption precedes diffusion and assimilation (Rogers, 1995).
Surprisingly, only Cooper & Molla (2017) addressed green digital innovation empirically within the consideration set. Nevertheless, literature (Mingay, 2007) discloses the increased concerns about climate change. Mingay (2007) addressed that green (or sustainable) digital innovations are introduced to counter these concerns. The concerns, as well as low empirical evidence, spurs future research directions.
