Master Thesis
MSc BA Strategic Innovation Management
An empirical analysis of the relationship between IT Ambidexterity and Firm Performance
AmbidexterIT: The influence of Explorative and Exploitative Information Technologies on Firm Performance, in the presence of a Digital Strategy.
by
Marc Tuinier S2929139
Supervisor: Nicolai E. Fabian Co-assessor: dr. John Dong
University of Groningen Faculty of Economics and Business June 2020
1
Abstract
The topic of enabling digital capabilities has become an important discussion among researchers and managers in the fields of strategic innovation management. It has become essential for organizations to reconfigure, integrate and build their IT capabilities in order to effectively capitalize on the digital environments in which they must now thrive. Therefore, this project aims contributes to this discussion by evaluating ambidextrous IT capabilities and the moderating effect of a digital strategy on firm performance. A unique data set covering over 100 million job postings over the course of a decade and an advanced dynamic panel estimator were used to analyze the relationships. The findings suggest that the pursuit of explorative IT capabilities has a negative effect on firm performance, advancing the notion that positive effects may be felt in the long-term. It is also found that exploitative IT capabilities negatively impact firm performance, proposing that the pursuit of both activities simultaneously may not be possible. Lastly, the findings imply that digital strategy may assist in improving firm performance through exploitative IT behavior but does not facilitate the expedition into unchartered territory.
2
Preface
I am grateful for the opportunities that have come across my path since moving to the Netherlands at the end of 2010. Before that, I was (and still am) the son of farming family with roots in Friesland that left on an adventure from France to Texas where I ended up being raised for the biggest part of my life. From education in Texas to being flung into preparatory scientific education in the Netherlands, I decided to always gather work experience on the side. In the past few years, these experiences included working at Red Bull, tutoring students privately in Shanghai, cofounding various startups and running the marketing for a tailor-made software vendor in Groningen. It had and will always be the intention to learn more and build with the knowledge I accumulate while studying. This is why the topic of ambidexterity intrigues me. With a knack for anything technologically related, I quickly realized the importance of getting everything one could out a specialization in Strategic Innovation Management. Taking part in the courses sufficient for the Focus Area in Digital Business would be the minimum. So, I took part in courses outside of our faculty on Neural Networks, Multi-Agent Systems, Philosophy of Neuroscience and Data in Digital Societies. There is still some time left, which I will dedicate to topics in Robotics, Smart Factories, Global Dynamics and Human Machine Communication. For the time that has passed, I would like to thank the organization that financed the last two years of my studies, specifically. Beeproger provided me the opportunity to organize their marketing efforts and work closely with the CEO in order to bring tailor made software solutions to more organizations around the Netherlands. Working this closely to what are often the beginnings of digital transformations, has allowed me to understand and extrapolate digital capabilities in both practice and theory. Better yet, I have an extremely great time doing just that.
On a personal level, ambidexterity is a process of learning how to get better at what you’re doing and getting better at looking for and applying what is new. Fortunately, my interests and work experience bring me to topics such as Innovation, AI and Human-Machine Interfaces. For this Master Thesis it was time to bring my interests together and find out how ambidextrous behavior can contribute to organizations. On a final note, I would like to thank Nicolai for challenging and improving my academic view in writing and interpretation; all while keeping this project fun.
3
Introduction
The accelerating adoption of digital technologies has led to transformations across all strands of society. In particular, the trend has proven to be of crucial strategic importance for organizations. Research is still unravelling the plethora of relationships between digital capabilities and performance (Kuntonbutr, 2020; Martínez-Caro, et al., 2020; Viete & Erdsiek, 2020) but has yet to show how digital capabilities fundamentally enable a firm to innovate: a simultaneous pursuit (e.g. ambidextrous) of becoming better at what they do and creating something new. As new digital technologies have become relatively cheaper over the course of time, firms’ use of technologies has changed such that control over assets has become less critical than the capabilities that emerge from them (Teece, 2017). This transition in viewing resources can also be explained through changes in conceptual thinking. While Barney (1991) argued for resource control as a way to gain competitive advantages, Teece (2010) proposed that firms build dynamic capabilities to deal with technologies.
These dynamic capabilities enable resource configurations as markets evolve, through their organizational and strategic imperative (Teece, 2010; Zahra et al, 2006). Research makes it clear that in order to adapt to their environment, firms need to reconfigure, integrate and build their capabilities. This concept is important to this project as it explains that capabilities evolve and emerge as an effect of the strategies firms set in place. As such, dynamic capabilities are higher-order capabilities (Collis, 1994; Winter, 2003) which should be understood as those that operate to enhance, reshape and generate ordinary capabilities. The significance of dynamic capabilities and the opportunities that arise from digital have left room for ample research regarding the significance of IT capabilities. This started when Wade and Hulland (2004) first acknowledged IT capabilities as key to mobilizing and deploying digital resources.
As firms became more involved in a digital environment, the term ‘IT Capabilities’ emerged as researchers sought to explain and deduce the digital phenomena in order to allow companies to structurally improve their capabilities (Aral & Weill, 2007, Kim et al., 2011, Pattij et al., 2019). Researchers unraveled that variations of IT capabilities were manifold. IT capabilities included IT Infrastructure and IT Human Skills (Pattij et al., 2019); IT Management Capability, IT Infrastructure Flexibility and IT Personnel Expertise (Kim et al., 2011); IT Integration and IT Business Alignment (Lyver & Lu, 2018) to Social Commerce-IT Capabilities (Braojos et al., 2017). While each variation enables the possibility to deepen our
4
knowledge, it leaves the opportunity to discover a fundamental dichotomy in the context of enhancing a firms’ ability to innovate.
The direct contribution this project makes will be to the topic of ambidextrous IT capabilities as it aims to unravel their effect on firm performance. The more interesting contribution is advancing current knowledge on the significance of a digital strategy to overcome the tension between both activities. In line with the uniqueness of these contributions, it must also be mentioned that the main data used in this project is also particularly novel. Also used by Hershbein & Kahn (2018), they show that skill requirements differentially increased in metropolitan areas relative to less victimized areas during the Great Recession. This skill data will also shed light on the extent to which firm performance is affected by explorative and exploitative behavior in their IT capabilities. The empirical findings will also contribute to literature on ambidexterity and IT capabilities. To address this gap concretely, two research questions will be stated to give form to the empirical analysis of this thesis. The first posits the question as to how ambidextrous IT capabilities affect firm performance, while the second dives into the mechanisms behind a digital strategy in overcoming the tensions rising from ambidextrous activity.
The dichotomy is that of embracing IT ambidexterity – the extent to which an organization exploits their current IT resources and practices (IT Exploitation) and explore new IT resources and practices (IT Exploration) – which fundamentally enables a firm to increase its agility and performance (One-Ki et al., 2015; Subramani, 2004). Exploitative activities require growing technical knowledge of realized systems and processes, for the purpose of optimizing these resources. On the other hand, exploration delves into the abilities encompassing a creative and broad understanding of the market and state of technology, in the effort to increase firm agility (Benner & Tushman, 2003; Ferraris et al., 2018; Ling et al., 2009; Lu & Ramamurthy, 2011; Wade & Hulland 2000;). The immediate downside to this concept is the imminent threat in which a firm mismanages these activities, leading to either a “competency trap” or “failure trap”. The success of repetition drives out the demand to explore while inexperience, constant deviation and failure engulf commitment, respectively (Leonard-Barton, 1992; March, 2003; Siggelkow & Rivkin). The following main steps are to unravel the relationships between these explorative and exploitative activities; elaborate on the IT aspect of those capabilities; build the hypotheses for this project and expand on the literature such that a comprehensive review and analysis contributes to the understanding of IT ambidexterity and digital strategy.
5
Literature Review
"Innovation is the design and development of something new, as yet unknown and not in existence, which will establish a new economic configuration out of the old, known, existing elements.”
- Peter F. Drucker, 1999
Ambidexterity
Fitting well into the orientation of this project, strategic innovation management, Drucker’s (1999) work transcended conventional thinking and touched upon the very core of dynamic capabilities. Building on the seminal work of Teece (1997), the entrenchment of these capabilities can originally be found in organizational ambidexterity (Duncan, 1976) and the concept of dual structures: commence and act on innovation (Burns & Stalker, 1961; Thompson 1967). This duality still drives ambidexterity theory, redefined as explorative and exploitative capabilities (Holmqvist, 2004; Lee et al., 2003, Nemancich & Keller, 2006). Exploiting and exploring entail employing and leveraging knowledge in well-understood and new ways (Taylor & Greeve, 2006). Exploitation deals with the demands surrounding efficiency and convergent thinking to exploit current capabilities. The flip side, exploring, entails efforts in searching and experimenting to produce novel recombination of knowledge (Andriopoulos & Lewis, 2009). Excelling in both activities is vital to long-term performance (Tushman & O’Reilly, 1996) and superior product development (Sheremeta, 2000).
Challenges that rise from the duality are innate to innovation. The forces that drive discovery and synthesis are challenges such as the tug-of-war between centripetal and centrifugal forces (Sheremata, 2000); the importance of knowledge breadth and depth (Taylor & Greeve 2000) and “competency and failure traps” driven by leaning towards one activity over the other (Gupta et al., 2006; Leonard-Barton, 1992). For example, organizations in the pursuit of exploring attempt to nullify failures and discount core competencies at the cost of current operations (Gibson & Birkinshaw, 2004). In an attempt to cross this chasm, the concept of ambidexterity encouraged top management to facilitate this balance through supportive structures and strategies (Gibson & Birkinshaw, 2004; Smith & Tushman, 2005). Along with dynamic capability theory, which provided an alluring framework to understand this change process, this project defines ambidexterity theory as emphasizing the mechanisms that enable a firm to explore and exploit; and its ability to reconfigure, integrate and build these capabilities.
6
Interestingly, the discussion that arose was whether the explorative and exploitative activities should be pursued simultaneously (Tushman & O’Reilly, 1997) or be viewed as temporal sequencing (Eisenhardt & Brown, 1998; Venkatramen et al., 2006). In the case of the latter, a firm assumes an environment in which the rate of change in markets permits it to act so. Inevitably, it becomes difficult for the firm to transform organizational strategy and alignment consistently in accordance with its focus (ie. exploit or explore). (Charles & Tushman, 2008) On the other hand, a simultaneous pursuit places a crucial importance on articulating a clear vision and strategic intent that justifies the operation of organizing capabilities ambidextrously (Rotemberg & Saloner, 2000). As such, this project aims to take from and contribute to this thinking, by elaborating on the importance of strategy and its moderating effect in the context of ambidexterity. In the context of digital, ambidexterity theory will be merged with information technology literature; emphasizing the importance of a digital strategy.
IT Ambidexterity
To act ambidextrously in combination with IT capabilities was introduced in 2003 when Benner and Tushman argued that explorative variation-decreasing and exploitative efficiency-oriented business process management practices were a pursuit of ambidexterity, through the IT department. Research initially tended to view IT capabilities from a functional perspective by placing an emphasis on the extent to which IT departments could enable rapid innovative releases while maintaining reliable architecture (Gregory et al., 2015, Leonhardt et al., 2017). Although useful, this project will view IT ambidexterity as an organizational pursuit through the measure of financial performance rather than functional, as IT departments can be leveraged for their specialized knowledge to assist in the reconfiguration, integration and development of IT capabilities (Chang et al., 2019; Sebastian, 2017).
The paradoxical conditions proposed earlier in the section on ambidexterity theory, can also be integrated into IT ambidexterity. The conditions are based on the premise that organizations are a set of intertwined business processes as an effect of organizational strategy and performance (Benner & Tushman, 2003). Value from IT is generated through business processes that find their root in organizational and technological resources such that business processes mediate the influence of IT on organizational performance (Ray et al., 2005; Schryen, 2013; Gattiker & Goodhue, 2005). These specific processes, defined as ordering of work activities (Davenport, 1998), can be defined in a digital strategy that encourage the pursuit of explorative and exploitative IT capabilities. In that respect, the strategy can also elaborate
7
on whether to use existing processes or create new ones (Gregory et al., 2015); to what extent a process should be automatic or manual (Gebauer & Shober, 2006; Kumar & Stylianou, 2013); the architectural choice of designing a modular or fully integral system (Tiwana et al., 2010; Shilling, 2000); and the use of either existing and known or unknown and impactful technology (Gergory et al., 2015). As this project places an emphasis on the latter, this section will continue to elaborate on explorative and exploitative IT capabilities; the mechanisms underlying their respective effect and the importance of a digital strategy in overcoming the tension.
Explorative IT Capabilities. As an extension of the explorative activities found in
ambidexterity theory, explorative IT capabilities involve creating a broad understanding of the market and building with the current state of technology (Wade & Hulland 2000). Explorative IT capabilities enhance a firms’ agility, innovativeness and growth (De Guinea, 2020) and are consequentially associated with outcomes such as new products or services that meet emergent customers’ or markets’ needs (Benner & Tushman, 2003; Cembrero & Sáenz, 2018; Popadiuk, 2012). These outcomes are the result of appropriating value from state-of-the-art and emerging technologies, which in turn enable a firm to become nimbler with regards to the development of its explorative IT capabilities (Ferreira et al., 2020). IT exploration brings with it the possibility to leverage diverse knowledge that in turn facilitates the identification of novel opportunities (Amit & Zott, 2001). Experimenting with new technologies introduces new and alternative knowledge domains yielding unique perspectives that help understand and simplify existing business transactions (Subramani, 2004; Tang et al., 2010). In the juxtaposition of novel outcomes and increasing appropriability of new technologies, firms can reduce costs while achieving efficiency, impacting firm performance through their explorative IT capabilities.
When considering the technological capabilities that enhance firms’ agility, growth and innovativeness, emerging technologies come to mind. Developments in big data emerge as an exploratory approach that yields positive firm performance results (Ghasemaghaei et al., 2018). Additionally, technologies and methodologies that allow for rapid prototyping on services and products are fundamental and prioritize exploration (De Guinea, 2020). Moreover, deploying novel IT architectures, such as Artificial Intelligence, enables firms to extend their knowledge base’s reach and richness (Kleis et al., 2012). Another form of an explorative IT capability is the application of topics such as Internet of Things, which also enable exploration of new business models (Glova et al., 2014). These examples in research on IT explorative capabilities
8
reinforce the conceptualization of competence exploration originally set in stone by March (1991): a firm dedicates resources to acquiring novel knowledge so that it can enhance product innovation, experimentation and its flexibility in exploring (McGrath, 2001; Raisch & Birkinshaw, 2008). To associate the capability with information technology within the context of this project, explorative IT capabilities are enabled through the appropriation of emerging technologies. In the next section, exploitative IT capabilities will be discussed in similar fashion in order to reveal the role of dependent technologies.
Exploitative IT Capabilities. Exploitative IT activities require growing technical
knowledge of realized systems and processes, for the purpose of optimizing these resources (Wade & Hulland 2000). Exploitative IT capabilities are those with the goal to strengthen a firms’ efficiency and productivity (De Guinea, 2020) which lead to the modification of existing services and products to meet customers’ or markets’ needs (Benner & Tushman, 2003; Cembrero & Sáenz, 2018; Popadiuk, 2012). The results are inherent to the mechanisms that underly exploitative capabilities, reflecting the extent to which a firm invests in its resources in order to reinforce its existing skills, processes and structures (Ferreira et al., 2020). IT exploitation improves the integration of existing knowledge (Avlonitis et al., 2001) and subsequent repackaging of said proficiency in different domains (Di Gangi et al., 2010). The enhanced integration and applicability of IT resources through IT exploitation (Im & Rai, 2014) enables firms to overcome differences in standards of data exchange and semantics (Rai & Tang, 2010) which in turn facilitates the integration of intelligence from nodes in the organization. Similarly, between strengthening efficiency and modifying existing services and products, organizations can effortlessly integrate knowledge through building on exploitative IT capabilities, reducing business uncertainty and impacting firm performance (Im & Rai, 2014; Malhotra et al., 2005).
Exploitative types of activities are commonly associated as being routinized in controlled environments (Gibson & Birkinshaw, 2004). To achieve this, it would be possible to increase task automation with a focus on the integration between existing IT resources (Shang & Seddon, 2002). Another direction would be to build on stable technologies within the resource base of the organization (Eisenhardt & Brown, 1998; Lewin & Phelan, 1999). A third option are software applications, such as Enterprise Resource Planning software, which place an emphasis on cost-reduction (Aral & Weil, 2007). Another perspective corresponds with the notion that automation (e.g. billing, reports, inventory management) is a form of exploiting
9
current IT resources (Kirtal et al., 2010; Gualandris et al., 2018). These emergent forms of IT exploitation as depicted in research, naturally build on the exploitative competencies (March, 1991) indicating the extent to which a firm invests in its resources in order to optimize its existing skills, processes and structures (Ferreira et al., 2020; Raisch & Birkinshaw, 2008). The connection to this project is made through the importance of dependent, or existing, technologies that facilitate a firms’ exploitative IT capabilities. The final section in this literature review will enlighten our current understanding on the role of a digital strategy in mitigating the imbalance brought by the simultaneous pursuit of both Explorative IT and Exploitative IT Capabilities.
Digital Strategy. In an attempt to embrace digital opportunities, firms enact digital strategies
to deliberately push organizations through a process in which information technology is a crucial lever of success (Mithas & Lucas, 2016). As a result, the terms digital and information technology are used interchangeably through research and in the context of organizations. The ambiguous character of digital and IT are a result of the emergent capabilities that firms derive from them. In both cases, digital and information technology strategies enable firms to build on their dependent technologies and allow organizations to experiment with emerging technologies (Catlin et al., 2015). In line with the transition in viewing a firm from a resource-based to a capability-resource-based perspective, McDonald (2012) introduced the notion that IT strategies force firms into regarding technology in isolation. Riding along McDonald’s proposition, this project considers looking through the “Digital Lens” in which digital strategies encapsulate the innovative ways to create value where digital information and physical resources merge. This effort is driven by the digital imperative of capitalizing on the increasing amount of digital connections will enable firms to interact with their customers and business in novel ways.
In an attempt to understand the importance of strategy in the context of ambidexterity and before elaborating on the distinction between the opportunities that arise when firms view their resources through the ‘Digital Lens’, it is paramount to recognize strategy as the crucial motivator behind the logic in this project. Generally, strategy involves a vision and a guide in terms of business goals to achieve said vision (Ahlstrand et al., 2001; Barber et al., 2006; Lubatkin et al., 2006; Raisch & Birkinshaw, 2008; Tushman, 1996; Vinekar et al., 2006). Within the realm of strategic management literature, the difference is stark among theorists who posit that strategy must involve a time component (Ahlstrand et al., 2001; van der Heijden,
10
,2004; Mintzberg, 1985) and those who feel that strategy should act as a roadmap (Barber et al., 2006; Eisenhardt, 1998; Hunger & Wheelen, 2003). Nonetheless, a general discipline has been found which entailed distinguishing strategy as being deliberate or emergent (Ahlstrand, 2001). However, it was posited that deliberate strategy could discount the future as it places an emphasis on control and planning (Eisenhardt, 1998; Veliyath, 1992) but should leave room for emergent strategies to emanate. Emergent strategies were set as critical components in turbulent business environments (King, 2008) as the results of deliberate decisions to focus and pursue aspects of strategic intent (Jet & George, 2005). To overcome these turbulent markets, a digital strategy should encapsulate a deliberate intent and leave room for emergent strategies to take place. To tie this in with ambidexterity theory and IT capabilities, the intent of digital strategy should be to assist in the reconfiguration, integration and development of Explorative and Exploitative IT capabilities as to overcome the paradoxes of ambidextrous theory, while leaving room for “un-thought-ofs”, or emergent strategies, to occur (Ahlstrand, 2001; Bodwell & Chermack, 2010). In this project, it is subsequently important to note how digital merges with information technology in order to provide the foundation for understanding the reconfiguration, integration and development of IT capabilities.
The tendency to view IT capabilities from a functional perspective was more prevalent when organizations had yet to discover the potentially transformational impact of integrating digital (Sebastian, 2017). It is possible to develop a suitable holistic theory, based on recent work from Chang et al (2019), who propose that firms should embrace the potential of cloud computing. Within reason of this project, for a firm to embrace the potential of cloud computing is equivalent to adopting a digitally oriented strategy. In order to appropriate value from “cloud computing” within the current IT ecosystem, organizations have or are ready to operate modular IT systems allowing for a smooth integration and a consistent ease of use. Viewing the firm as a modular IT ecosystem brings with it the consequence and importance of systemically governing relationships with the nodes in the ecosystem (Schilling, 2000). Whether it be employees, developers, suppliers or customers; a digital strategy unavoidably takes each of those who contribute to the IT ecosystem into account. This view has also been substantiated by research in “platform thinking”, which aim to view organizations as being modularized and working as components (Cusumano, 2013). Chang et al (2019) conclude that the deployment of cloud computing is still in its infancy; giving metaphorical purpose to digital pursuit at hand.
11
As such, “digital” is a variation of the entrenched “information technology” (McDonald & Rowsell-Jones, 2012; Tumbas et al., 2015) indicating the expansive impact of digital technologies. Digital innovation has become critical in organizations, further than the functionalized and organizational provenance fueled by technology (Grossman, 2016; Yoo et al., 2010). By merging these aspects, digital technologies have enabled organizations to look beyond the conventional utilization of technology, yielding results such as the appropriation of digital products by embedding software into products (Henfridsson et al., 2014) and enhanced customer profiling (Müller et al., 2016; Tambe, 2014). To enable such outcomes, this project acknowledges the importance of enhancing an organization’s ambidextrous IT capabilities. To recapitulate, value from IT finds its roots in organizational and technological resources; business processes mediate the influence of IT on organizational performance (Ray et al., 2005; Schryen, 2013; Gattiker & Goodhue, 2005); and these specific processes can be defined in a digital strategy (Davenport, 1998). Through strategy, it is possible to reflect both exploitational and explorational techniques, fundamental in dynamic capability theory (Harreld et al., 2007) and the reconfiguring, integration and development of IT capabilities. This project aims to contribute to this strand of research, by exploring the effect of digital strategy on the relationship between ambidextrous IT capabilities and firm performance.
12
Hypothesis Development
Figure 1: Conceptual Model
Explorative IT capabilities refers to the extent that a firm identifies and deploys new IT resources to support business process and strategies (Lu & Ramamurthy, 2011; Lee et al., 2015). However, a paradox rises when explorative IT capabilities are being pursued on their own such that explorative abilities enable firms to create novel knowledge in order to acclimate to environmental change and diminish the risk of extinction (Uotila et al., 2009); yet they come with the risk of locking into the “failure trap” (Levinthal & March, 1993). These capabilities are crucial to prosper and survive in the long term. Thus, the radical nature of the explorative output results in subpar short-term firm performance due to do the inapplicability of the innovation in the market (De Guinea, 2020; Lennerts et al., 2020; Norman & Verganti, 2014). Although it has been acknowledged that explorative IT capabilities enable novel outcomes and increase appropriability of new technologies; it is possible for firms to reduce the expenses made to appropriate and increase efficiency, positively influencing firm performance through their explorative IT capabilities (Subramani, 2004; Tang et al., 2010). Building on this, the research of Liu et al (2019) empirically justify that the effect of explorative behavior is felt in the long term. As part of this project, our dataset consists of a time span smaller than a decade; from the launch of iPhone 3G, to where we are now. For this reason, this project hypothesizes a negative relationship between explorative IT capabilities and firm performance. To conclude, building explorative IT capabilities comes with a downside: an emphasis on the long term indicating that short term firm performance may be inferior. Thus, the first hypothesis reads as follows:
H1: There is a negative relationship between explorative IT capabilities and firm performance.
H1
H2
H3
13
Exploitative capabilities encourage a process-oriented look at the firm in which the main aim is to optimize it current IT capabilities. This project assumes a positive relationship between exploitative IT capabilities and firm performance. From an exploitative IT capabilities perspective, firms are able to continuously update and extend their current infrastructures by constantly exploiting existing IT resources (Premkumar et al., 2005; Rai & Tang, 2010; Saraf et al., 2007). This helps a firm develop its operational competence (Benitez et al., 2018) and leads to improved firm performance (Saraf et al., 2007; Yao & Zhu, 2012). Underpinning a fundamental effect of ambidextrous behavior; through optimization, organizations can integrate more efficiently by building on exploitative IT capabilities, reducing business uncertainty and positively affecting firm performance (Im & Rai, 2014; Malhotra et al., 2005). To keep the use of research consistent, Liu et al (2019) also find that exploitative behavior has a significant effect on short-term financial performance. This is done by placing an emphasis on decreasing diversity, increasing efficiency and improving the appropriability to an organization’s internal environment such that it positively effects short term performance (Uotila et al., 2009). Having mentioned the limitation of our dataset, it does offer the opportunity to analyze the short-term effect of exploitative IT capabilities and firm performance. Building on this logic, the second hypotheses discerns the effect of exploitative IT capabilities on firm performance:
H2: There is a positive relationship between exploitative IT capabilities and firm performance.
The presence of a digital strategy places the importance of strategy in the context of firm performance (Bharadwaj et al., 2013) and ambidextrous capabilities (Schryen, 2013; Gattiker & Goodhue, 2005). It is argued that digital strategy should not be that of a functional-level strategy but that it is fused with business strategy (McDonald & Rowsell-Jones, 2012; Tumbas et al., 2015). In line with this theory, Mithas & Rust (2016) present empirical results that suggest a strategic IT emphasis plays a significant moderating role between IT investments and firm performance. To expand on the research of Tumbas et al (2018), they propose that a digital strategy can be encouraged through the installment of chief positions. Similarly, IT departments have been considered significant in enabling digital innovation (Guillemette & Pare, 2012; Weill & Woerner, 2016). This entails that the implementation and execution of higher-level digital strategies allow firms to create and appropriate value more effectively when that firm is found in a digital setting (Chi et al., 2016). Thus, a digital strategy directs digital resources to support the strategic need of the business and to apply existing IT capabilities to
14
discover new business opportunities (Tallon, 2007). This key mechanism encourages the understanding that Explorative and Exploitative IT capabilities can be reconfigured, integrated and developed through a digital strategy. This logic holds when a digital strategy is set in place such that this strategy moderates the relationship between the ambidextrous IT capabilities and their effect on firm performance. To advance, the moderating effect of a digital strategy on the aforementioned relationship will be hypothesized such that the digital strategy weakens the supposed effect in H1 and strengthens the assumed effect in H2. These relationships are depicted in Figure 1 as part of the conceptual model. Thus, the presence of a digital strategy moderates the relationship between explorative and exploitative IT capabilities on firm performance as follows:
H3: The negative relationship between explorative IT capabilities and firm performance is positively
moderated by a Digital Strategy such that the presence of a Digital Strategy weakens the negative relationship between explorative IT capabilities and firm performance.
H4: The positive relationship between exploitative IT capabilities and firm performance is positively
moderated by a Digital Strategy such that the presence of a Digital Strategy strengthens the relationship between exploitative IT capabilities and firm performance.
15
Data and Methodology
Data Source
The dataset that will be used derives from an analytical product that covers over 100 million vacancies posted online in the United States, amassed by Burning Glass Technologies (BGT): an employment and labor market data and analytics organization. Probing an estimated 40,000 employability and organization sites, Burning Glass parsed and deduplicated the vacancies to an accessible market analytic product (Hershbein & Kahn, 2018). This product enables researchers to scour an approximated 70 fields of detailed information such as skill requirements, which are crucial for this project. Specifically, BGT refers to these skills as essential or forming for the particular position. Additionally, the dataset allows to comb through a detailed set of job requirements, from soft skills (e.g. negotiation) to hard skills (e.g Python and SQL). The effort put into codification yielded a blueprint for this project. The outcome enables this project to analyze a crucial and interesting margin of firm demand: the skill requirements needed for an occupation (Hershbein & Kahn, 2018). A helpful step was that BGT categorized the dataset in three parts: Main Clusters, Sub Clusters and Skill Clusters from which we extracted the following insights exclusively for the S&P 500. The overview contained ±5,500 skills, forming ±600 sub clusters grouped in 27 main clusters1. To explain the clustering
and structure of the data, please view a subset of the data used in this project in Table 1.
Table 1: Example of Clusters
Burning Glass has categorized the labor market making it is possible to distinguish specific skillsets to accompany those responsibilities as sub clusters. Sub clusters will be used in this
1 The 2,400 skills that were not clustered, were excluded from this project.
Main Cluster Sub Clusters Skills
Analysis Business Intelligence Data Analysis
Business Intelligence Report Tools (BIRT)
Business Operations Management
Process Improvement 5S Methodology
Information Technology
Data Warehousing
16
project, as the scope they operationalize fits well with the organizational scale on which the analysis will be done. In line with the literature review, the explorative and exploitative capabilities should not be departmentalized but viewed from an organizational perspective. To connect the elements, Burning Glass’ sub clusters allow this project to operationalize exploitative and explorative capabilities on a firm level over the entirety of the Standard & Poor’s 500 index, nearing the span of a decade. However, this would only consider the independent variables in the hypotheses. To elaborate, the dependent, moderating and control variables are not part of BGT’s dataset. Hence, supplemental data was found in order to also operationalize firm performance, digital strategy and the control variables. For firm performance and control variables, financial and firm-related data was captured through the Compustat database provided by Wharton Research Database Services. The measure for digital strategy was obtained through BoardEx provided by Management Diagnostic Limited. More specifically, BoardEx is organized as a time-series of hypertext-linked individual curriculum vitae (Engelberg et al., 2009) which allowed for the extraction of current employment status of high-ranking positions in firms based in the United States. The aggregate of this data was indexed on Ticker and Year, for the comprehensive and efficient analysis.
Variables
Explorative IT Capabilities. Consistent with previous structure, the first of two
independent variables to operationalize will be a firm’s explorative IT capabilities. These capabilities are best reflected by the extent to which firms explore new IT resources and practices (Lee et al., 2015). To develop the logic in order to operationalize these variables reliably, this project develops a more grounded approach of the independent variables. For example, the extent to which firms explore new IT resources and practices is consistent with the notion that emerging technologies such as Artificial Intelligence (Kleis et al., 2012), Internet of Things (Glova et al., 2014) and Augmented Reality (Rauschnabel, 2018) inspire explorative behavior. This tendency builds on the theory that unabsorbed slack resources, or technologies for which the appropriability is yet to be understood, contribute to explorative behavior (Singh, 1986; Liu et al., 2019). With a comprehensive dataset at hand and an elaborate literature review on explorative IT capabilities, it is possible to operationalize the variable as one that considers emerging technologies. In this case, six sub clusters were selected based on research in which scholars used similar constructs, placing an emphasis on a firm’s explorative behavior and use emerging technologies. Table 2 depicts an overview of the selected sub clusters for Explorative
17
IT Capabilities and the research that substantiates the selection. For the analysis, these sub clusters were summed to operationalize Explorative IT Capabilities as Explorative IT.
Exploitative IT Capabilities. On the other hand, an organization’s exploitative IT
capability is best regarded as the ability to use existing IT resources and practices (Lee et al., 2015). This ability to use existing IT resources is coherent with the understanding that building on and utilizing dependent technologies through Business Intelligence (Harrison et al., 2015), Data Management (Shang & Seddon, 2002) and IT Automation (Kirtal et al., 2010; Gualandris et al., 2018) motivate exploitative activity. Following the logic formed in the section about Explorative IT Capabilities, research also shows that the inclination of relying on absorbed slack resources, such as operational slack rooted in the existing routines of firms, facilitates exploitative behavior (Singh, 1986; Liu et al., 2019). Justified through this logic, building on a firm’s Exploitative IT Capabilities will ensure exploitative behavior. Similar to the previous paragraph on this variable’s counterpart, the exhaustive dataset provides the opportunity to operationalize Exploitative IT Capabilities such that it coincides with the appropriation and reconfiguration of dependent technologies. Identically, the six sub clusters which were selected are consistent with research on the exploitative nature of the dependent technologies as appropriated by firms. Building on the previous section, Table 2 gives an overview of these dependent technologies and corroborated research that measure a firm’s Exploitative IT Capabilities as Exploitative IT.
Table 2: Independent Variables
Sub Clusters Scholars
Explorative IT
Capability Data Visualization (Chang et al., 2020; Demšar et al., 2007)
Artificial Intelligence (ML, NLP) (Chang et al., 2020; Demšar et al., 2004) Big Data (Bagozi et al., 2017; Bøe-Lillegraven, 2014) Cloud Solutions (Chang et al., 2019; Cusumano, 2013) Software Development
Methodologies (Harms & Schwery, 2020; Vavpotič et al., 2019) User Interface and Experience (Jessen et al., 2020; Kou & Gray, 2018)
Exploitative IT
Capability Business Intelligence (Popovič et al., 2019; Wieneke & Lehrer, 2016)
Data Science (Rialti & Marzi, 2020: 30; Tirunillai & Tellis, 2014) Data Management (Benitez et al., 2018; Wei et al., 2019) IT Automation (Kirtal et al., 2010; Gualandris et al., 2018) System Design and
Implementation
(Beniteze et al., 2018; Shang & Seddon, 2002)
18
Digital Strategy. The moderating role of a digital strategy within the realm of ambidextrous
IT capabilities performs such that it relocates digital resources to support the strategic need of the business and appropriates existing IT capabilities in order to observe novel outcomes (Tallon, 2007). Digital strategies are used to push organizations through a process in which information technology is a crucial lever of success (Mithas & Lucas, 2016). As part of this project and in line with Tumbas et al (2018), a digital strategy can be encouraged through chief positions such as those in IT departments. Similar to research by Nwankpa & Datta (2017), Chief Technology Officers were analyzed as they were individuals with comprehensive knowledge of IT innovations and digital business. Thus, this project aims to unravel the effect of employing chief positions in technology, as a proxy for the presence of a digital strategy, on the relationship between ambidextrous IT capabilities and firm performance. To operationalize this measure, the previous logic was applied, and supplemental data was obtained through BoardEx. Organized as a time-series of hypertext-linked individual curriculum vitae (Engelberg et al., 2009), it was possible to extract when Chief Technology Officers were employed at firms in the S&P 500. Subsequently, a dummy variable was made in which a firm that employed at CTO was labeled with 1; vice versa with 0. With the data at hand, the data set used for this project now consists of S&P 500 firms, their explorative and exploitative behavior, aligned with data on the employment of Chief Technology Officers that facilitate the role of a Digital Strategy.
Financial Performance. In order to comprehend the effects of the predetermined measures
on financial performance, this project will use return on assets as its dependent variable. Used to quantify a firm’s financial performance, ROA emphasizes the effectiveness and efficiency in which firms employ resources to generate income and maximize profit (Daily & Johnson, 1997; Gallo, 2016). Profitability ratios are used to effectively analyze, enhance and control the operations of a firm (Leskava, 2007), inherently making it an appropriate operational measure (Gallo, 2016). For the purpose of robustness checks, ROA is accompanied by Tobin’s Q. As a measure commonly used for research in strategy, Tobin’s Q assesses the extent to which firms are valued by financial markets with respect to their replacement cost (Carpenter, 2002). This is valuable as it takes short term performance into account, while weighing in on the future performance of the firm (Allen, 1993). These measures were collected through the Compustat database provided by Wharton Research Database Services.
19
Control Variables. In this project, two control variables are used: Firm Size and R&D
Intensity. Firm Size is measured as the natural log of total assets (Cantor, 1990) and R&D Intensity is measured as the ratio of R&D expenditures to net sales (Billings, 1999). Both are used as they are strong indicators that influence a firm’s preparedness for innovation (Framback & Schillewaert, 2002; Moohammad et al., 2014). More specifically, R&D Intensity is included in this project because of its known contribution to employment growth (Hall, 1986), output (Hall et al., 2010; Parisi et al., 2006) and firm performance (Falk, 2012). Firm Size is also appropriate as a control due to its acknowledged contribution to relative firm performance (Amato & Amato, 2004; Lee, 2009).
Data Analysis
Preliminary Analysis
Before running the regressions, the data will be tested for assumptions and then two overviews will be given regarding summary statistics and a correlation matrix. To test for normality, the Shapiro-Francia tested the untransformed variables in Table 3 and indicated unnormal distribution (p = .0001) across the dataset. To solve this issue, the variables were transformed to create a lognormal distribution. As this project also intends to analyze the effect of the moderating term, interaction variables were created based on log transformed variables. An overview of these variables can be found below.
Table 3: Summary Statistics
VARIABLES N mean sd Year 6,029 2,014 2.889 IT Exploration 4,998 2,626 47,056 IT Exploitation 4,998 8,403 155,243 Digital Strategy 4,998 0.533 0.499 RD Intensity (log) 4,963 0.0213 0.0442
Firm Size (log) 6,029 2.352 0.144
DS x Exploration 4,998 2.452 2.931 DS x Exploitation 4,998 3.463 3.499 IT Exploration (orthog) 4,998 0.00525 1.055 IT Exploitation (orthog) 4,998 -0.000466 1.055 ROA 6,019 0.0511 0.0926 Tobin’s Q 5,984 1.067 0.355
Next, a correlation matrix was produced in order to understand the strength and direction of association between the continuous variables in the dataset. As part of this matrix, the correlation coefficient, denoted as r indicates how well the data surround the line of best fit.
20
The relationship r ranges from -1 to +1 implying a perfectively negative linear relationship and positive linear relationship, respectively. Due to both high correlation in the untransformed and transformed independent variables, the original independent variables were orthogonalized to decrease correlation and accompanying statistical significance. To elaborate, log transformations of the independent variables still indicated relatively high correlation (r = 0.67). After the orthogonalization of IT Exploration and IT Exploitation, the correlation between the two variables was low (r = 0.17), ruling out multicollinearity for the independent variables. The statistical significance indicators and corresponding correlation coefficients are included in the matrix below for the purpose of further analysis and discussion.
Table 4: Correlation Matrix
Variables 1 2 3 4 5 6 7 8 9 1. IT Exploration 1 2. IT Exploitation -0.17*** 1 3. Digital Strategy -0.03 -0.03* 1 4. DS x IT Exploration 0 -0.07*** 0.78*** 1 5. DS x IT Exploitation -0.01 -0.05** 0.93*** 0.89*** 1 6. RD Intensity 0 -0.04* 0.12*** 0.18*** 0.12*** 1 7. Firm Size -0.02 -0.03* 0.11*** 0.23*** 0.24*** -0.27*** 1 8. ROA -0.02 -0.02 0.02 0.06*** 0.04* 0.04* -0.19*** 1 9. Tobin’s Q -0.02 -0.03 0.02 0.08*** 0.02 0.44*** -0.47*** 0.48*** 1 * p < 0.05, ** p < 0.01, *** p < 0.001
Another check for multicollinearity was conducted through interpreting the variance inflation factor (VIF). This statistic starts at 1 and indicates the percentage, in decimal form, to which the variance is inflated for each coefficient. In the case of this project, the VIF values that are highest coincide with the interaction term including a Digital Strategy and IT Exploitation (VIF = 15.05), Digital Strategy as a single variable (VIF = 8.39) and the interaction term Digital Strategy and IT Exploration (VIF = 5.1). The remaining variables had values that were well under 1.5 and these additional statistics can be found in Appendix A.
Hypothesis Testing
A GLS-random effects model is used due to the dummy-nature of the moderator and the assumption that differences exist across the variables, which may impact the dependent variable. As part of Table 5, three models are displayed which ran on the financial performance
21
indicator. In this section, the models are estimated using stepwise regression in order to test the main effect hypotheses. Subsequently, the moderation term is taken into account for both versions of the third hypothesis. These regressions will be checked for homoscedasticity, endogeneity and autocorrelation using the generalized method of moments made specifically for panel data. In this section, the following will elaborate on the extent to which the hypotheses are supported by the results of the analysis.
To start, H1 assumed a negative relationship between explorative IT capabilities and firm performance. To
test the effect of explorative IT capabilities on firm performance, the independent variable took emerging technologies such as Natural Language Processing and Data Visualization and operationalized these constructs as IT Exploration. Firm performance, on the other hand, was operationalized through Return on Assets. From Model 3, we find evidence for a negative relationship (r = -0.00307, p=0.1) between IT Exploration and Firm Performance such that for each additional unit of IT Exploration contributes to a 0.31% decrease in Firm Performance. Thus, we have support for the claim that IT Exploration has a negative effect on firm performance.
Secondly, H2 considers a positive relationship between exploitative IT capabilities and firm performance. As
mentioned in the former paragraph, firm performance was operationalized as return on assets. On the other hand, exploitative IT capabilities was operationalized as IT Exploitation containing constructs such as Data Management and IT Automation. Following Model 3, evidence for a negative relationship (r=-0.00331, p=0.01) can be found between IT Exploitation and Firm Performance, comparable to the previous findings. In this case, an increase of one unit in IT Exploitation results in a 0.34% decrease in Firm Performance. Following this analysis, we do not have support for the claim that IT Exploitation has an effect on firm performance.
Lastly, H3 and H4 take relatively similar stances in that the third hypothesis suggests that the
negative relationship between explorative IT capabilities and firm performance is positively moderated by a Digital Strategy such that the presence of a Digital Strategy weakens the negative relationship between explorative IT capabilities and firm performance. Model 3 shows that we do not find evidence for the interaction effect in hypothesis three. The fourth hypothesis follows similar logic such that the positive relationship between exploitative IT capabilities and firm performance is positively moderated by a Digital Strategy such the presence of a Digital Strategy strengthens the relationship between exploitative IT capabilities
22
and firm performance. We do find evidence for a significant interaction between a Digital Strategy and IT Exploitation (r=0.0074, p=0.01) such that the presence of a Digital Strategy strengthens the negative effect of IT Exploration on Firm Performance. There is support for hypothesis four. These results can be found in Table 5 on the next page and will be reviewed more elaborately by joining the literature in the discussion section following our analysis.
Table 5: Regression Results
(1) (2) (3)
VARIABLES Model 1 Model 2 Model 3
Firm Size -0.0792*** -0.0830*** -0.122*** (0.0159) (0.0160) (0.0168) RD Intensity -0.150*** -0.159*** -0.194*** (0.0453) (0.0455) (0.0454) IT Exploration -0.00278* -0.00307* (0.00162) (0.00160) IT Exploitation -0.00335*** -0.00331*** (0.00105) (0.00105) Digital Strategy 0.00423 -0.0479*** (0.00436) (0.00940) DS x IT Exploration 0.00120 (0.00109) DS x IT Exploitation 0.00742*** (0.00166) Constant 0.243*** 0.250*** 0.343*** (0.0379) (0.0378) (0.0398) Observations 4,959 4,959 4,959
Number of Ticker Groups 573 573 573
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Robustness
As panel data consisting of multi-dimensional data over the course of a decade is being used, a dynamic panel estimator is appropriate to check for robustness. This is due to the likelihood of panel data suffering from autocorrelation, endogeneity, omitted variable bias, unobserved panel heterogeneity and measurement errors. The most prevailing bias the dataset is likely to suffer from is omitted variable bias due to the nature of jobs postings and the skills (e.g capabilities) that can be developed through other means as well. The generalized method of moments model (GMM) (Arellano & Bond, 1991; Blundell & Bond, 1998), a dynamic panel estimator, was designed to capture the time lag in cause and effect found in panel data. Cited over 7000 times, this system is applicable in Stata through the xtabond2 command (Roodman,
23
2009) and is still proven to be incredibly useful to check for the aforementioned issues (Bennet, 2019; Hernandez-Vivanco et al., 2019; Tchamyou, 2020). To elaborate, lags of the dependent variable are used as explanatory valuables (Ullah, 2018). The system GMM process of removing endogeneity refers to the process in which the variables past value is deducted from its value in the present and orthogonalized (Roodman, 2009), increasing efficiency of the model by the reduction in variables (Wooldridge, 2012). Additionally, there are two transformations known as one-step GMM and two-step GMM. For the purpose of this project, two-step GMM will be used following the advice of Arellano and Bover (1995) to avoid potential data loss. This phenomenon occurs due to the practice of ‘forward orthogonal deviation’ indicating the deduction of the average of prospective observations on a specific variable. For this reason, Ullah et al (2018) argue that a two-step GMM contributes to efficient and consistent estimates for the participating coefficients.
The output for the main two-step GMM is based on 4959 observations scattered over 573 groups based on Tickers with the observations ranging from 2010 to 2019. Two models were set up in order to check for robustness vertically as well as horizontally through the operationalization of firm performance as ROA and Tobin’s Q. The results can be found on the next page, Table 6, and will be elaborated in this section. To start, the GMM model supports the findings for H1 such that IT Exploration negatively affects Firm Performance as
Return on Assets (r=-0.0039, p=0.01) as well as Firm Performance operationalized as Tobin’s Q (r=-0.135, p=0.01). Secondly, our findings for H2 are supported such that IT Exploitation
negatively influences Firm Performance as ROA 0.0037, p=0.01) and Tobin’s Q (r=-0.0057, p=0.01). Overall, Model 1 and Model 2 support the regression models conducted for the first and second hypothesis. Lastly, Model 1 indicates the significance of a Digital Strategy (r=-0.0578, p=0.01).
On another note, regarding the third and fourth hypotheses, Model 1 finds no significant support for the interaction effects. However, Model 2 implies significance for both interaction terms suggesting that a Digital Strategy positively moderates the relationship between IT Exploration and Firm Performance (r=0.0927, p=0.01) such that it weakens the negative effect of IT Exploration on Firm Performance as Tobin’s Q. Model 2 also suggests that a Digital Strategy negatively moderates the negative relationship between IT Exploitation and Firm Performance such that it strengthens the negative effect of IT Exploitation on Firm Performance as Tobin’s Q (r=-0.0655, p=0.05). These findings thus partly support the results
24
found for the fourth hypothesis. There is no support for the third hypothesis. It is important to note that these models come with two caveats regarding autocorrelation and the Hansen test. The former is significant only for Model 1, which influences the coefficients used in this project. The latter considers the restrictions in a model estimated with instrumental variable techniques and is also relatively high for both Models indicating that we have too many instruments in this dataset. Overall, the GMM indicate robustness for hypothesis one and two for both versions of the dependent variables.
Table 6: GMM Results
(1) (2)
VARIABLES ROA Tobin’s Q
Firm Size -0.147*** -1.175*** (0.0226) (0.0117) RD Intensity -0.0997 2.2139*** (0.1174) (0.3597) IT Exploration -0.0039*** -0.135*** (0.0001) (0.0006) IT Exploitation -0.0037*** -0.0057*** (0.0001) (0.0004) Digital Strategy -0.0578*** -0.0288 (0.0189) (0.08) DS X IT Exploration 0.0045 0.0927*** (0.0053) (0.0216) DS X IT Exploitation 0.0059 -0.0655** (0.0072) (0.0283)
Year Dummies Yes Yes
No. of Obs. 4959 4946 F Statistic 3900.94 114.08 Groups/Instruments 573/39 573/39 AR (2) 0.013 0.573 Hansen Test 0.971 0.854 Sargan Test 0.074 0
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
25
Discussion
Theoretical Implications
In the context of this project, the endeavor built on dynamic capability theory and ambidexterity theory to extricate novel insights as to how ambidextrous IT capabilities affect firm performance; and to contribute to the understanding of the mechanisms behind the role of strategy to overcome the tensions that are inherent to the simultaneous pursuit of explorative and exploitative IT capabilities. The tensions were those regarding the extent to which success of repetition deposes demand for exploration while consistent deviation and inexperience overwhelm commitment. This extreme perspective necessitates digital strategy to direct key digital resources to support the strategic ambidextrous imperative. It is for this reason that the role digital strategy is crucial and novel in this project. As such, the following paragraphs will elaborate on how this project advances understanding on Explorative IT Capabilities, Exploitative IT Capabilities and Digital Strategy.
Explorative IT Capabilities reinforce the appropriability of emerging technologies, leading to the possible reduction of costs and improvement in efficiency; both contributing to firm performance. However, the caveat this raised was time dependent such that the suggested positive firm performance through the appropriation of the unabsorbed slack resources may be justified, but only in the long-term. Thus, explorative IT capabilities enable firms to effect integration with emergent market needs yet set cause for organizations to reap positive results further along the horizon. As this project’s dataset was solely set on the past decade, the hypothesis suggested that IT exploration negatively affects firm performance. This claim is supported, suggesting that the caveat to explorative behavior is grounded. Literature suggests that explorative abilities enable firms to create novel knowledge in order to acclimate to environmental change and diminish the risk of extinction of which the results may only be felt in the long run. This finding advances this notion through empirically substantiating the notion that it takes time for firms to benefit from explorative behavior in the realm of appropriating value from emerging technologies.
Exploitative IT Capabilities enable enhanced integration and applicability of IT resources. Strengthening the efficiency of existing services and the optimization of products allow organizations to effectively integrate absorbed slack resources; thereby reducing business uncertainty and encouraging the improvement of firm performance. Through this perspective,
26
firms should continuously update and extend their infrastructures thereby developing operational competence resulting in improved short-term financial performance. This logic led to consider a positive relationship between exploitative IT capabilities and firm performance. This claim was not supported, while evidence of a significant negative relationship was. The finding is counter intuitive to literature yet offers the possibility to underpin the position in ambidexterity theory which encapsulates that the simultaneous pursuit of both activities may not be possible. The counteraction also makes this finding interesting considering that literature suggests that exploitative practices are associated with positive short-term firm performance. The finding leaves room to consider the thought in which financial performance might not be the correct indicator of success in the pursuit of exploitative IT practices. Although literature suggests it should, evidence for a significant negative relationship advances the idea that short term positive outcome may be found along a different indicator.
The imperative of a digital strategy rests on the value that can be derived from technological resources and corresponding business processes. Most importantly, it is these specific processes that should be defined in order to direct key resources. Specifically, digital strategies should enable firms to build on their current technologies and allow for experimentation with emerging technologies. In this project’s case, theory was presented that suggested the strategic importance of Technology Officers in the implementation of digital strategies such that the presence weakens the negative effect of exploration and strengthens the effect of exploitation on firm performance. The findings imply that interaction is found for the latter, advocating that digital strategy may assist in improving firm performance through exploitative IT behavior but does not facilitate the expedition into unchartered territory. The literature allows to cogitate and construct a logical conclusion to this finding, namely, that a firm is more likely to effectively strategize with that which it knows in contrast to that which it does not. These findings contribute to literature such that the imperative of a digital strategy to direct digital resources is predominantly effective when an organization participates in IT Exploitation. On the other hand, it may be that the appropriability of emerging technologies is effective when a digital strategy both deliberately encourages experimentation but leaves enough room for emerging strategies to evolve in order to facilitate IT Exploration. All variables and interaction terms considered, the findings imply the significance of a Digital Strategy such that it should not be discarded but embody both Explorative and Exploitative IT Capabilities in order to effectively overpower the tensions raised through ambidextrous behavior.
27
Limitations & Future Research
The constraints raised in the scope of this project consequentially result in areas in which future research could be directed towards. To start, the dataset used to analyze the sub clusters covers vacancies solely posted on the internet and does not, for example, take replacement hiring into consideration. This entails that skills that are immediately necessary, were not posted online and would thus not be adopted in the dataset. For the purpose of this project, it could be assumed that skills that encourage immediate exploitative behavior are left out of the equation. Similarly, enthusiasm towards emerging technologies and its inherent turbulent environment also sets precedent for relatively abrupt hires in order to secure technologically advanced positions. Similar to replacement hiring, this also brings an interesting point in which upskilling, or M&A activity is not considered. Both could have an impact on the relationship between the development or acquisition of skills for ambidextrous IT capabilities on firm performance. Although endogeneity was controlled for, our findings suggest that it cannot be ruled out. Secondly, the sub cluster constructs used to shape Explorative and Exploitative IT Capabilities could be tested to empirically justify their belonging. Future research could apply a structural modelling approach to discern the effect of the sub clusters to build a deeper understanding of the interaction between IT Exploration and Exploitation. Lastly, the data used to analyze digital strategy was limited in its comprehensiveness due to its dummy nature. Future research could embark on coding annual reports to analyze the extent to which digital strategies are set to explore emerging technologies and exploit a firm’s current technological stack.
Managerial Implications
Building on the findings mentioned in the previous section, this project advocates that managers enact digital strategies to develop explorative and exploitative competencies. Encouraging experimentation with emerging technologies allow firms to develop their explorative IT capabilities which in turn enable firms to effect integration with emergent market needs. This enforces a firm’s forward-looking strategy such that the firm understands that potential positive effects can be sown in time yet allows the firm to perform explorative behavior. Perhaps simultaneously, firms should continuously update and extend their infrastructures in order to appropriate both existing as well as emerging technologies more effectively. In order to overcome the tensions found in the paradoxical conditions of ambidexterity, our findings suggest that, especially in the case of exploiting a firm’s resources, a digital strategy amplifies the positive effect of exploitative IT capabilities on firm performance. On a similar note, when a firm wishes to employ IT ambidexterity it should consider the
28
organization as a whole and not view IT as functional. Thus, a digital strategy is formulated such that it encompasses the “Digital Lens” through which digital becomes ubiquitous by the consolidation of digital information and physical resources. Managers applying this “Digital Lens” in order to concatenate a digital strategy should explicitly view IT Exploration and IT Exploitation as practices crucial to continuously innovating in a world dominated by digital technologies. Specifically, this entails providing guidelines for teams to proactively appropriate emerging technologies, whether it be a simple program for 3D Modelling or a deliberate strategy to incorporate Augmented Reality. On the other hand, guidelines should encourage critique on current processes such that they are consistently evaluated against the firm’s environment. In both respects, these guidelines should still allow for the “un-thought-ofs”, because only then will an organization achieve AmbidexterIT.
29
References
Agarwal, R., Lucas, H., Clemons, E., Sawy, O., & Weber, B. (2013). Impactful Research on Transformational Information Technology: An Opportunity to Inform New Audiences. MIS Quarterly, 37(2), 371-382.
Ahlstrand, B., Lampel, J., & Mintzberg, H. (2001). Strategy safari: A guided tour through the wilds of strategic management. Simon and Schuster.
Allen, F. (1993). Strategic management and financial markets. Strategic Management Journal, 14(S2), 11-22.
Amato, L. H., & Amato, C. H. (2004). Firm size, strategic advantage, and profit rates in US retailing. Journal of Retailing and Consumer Services, 11(3), 181-193.
Amit, R., & Zott, C. (2001). Value creation in e‐business. Strategic management journal, 22(6‐ 7), 493-520.
Andriopoulos, C., & Lewis, M. W. (2009). Exploitation-exploration tensions and organizational ambidexterity: Managing paradoxes of innovation. Organization science, 20(4), 696-717. Aral, S., & Weill, P. (2007). IT assets, organizational capabilities, and firm performance: How resource allocations and organizational differences explain performance variation. Organization science, 18(5), 763-780.
Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The review of economic studies, 58(2), 277-297.
Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of econometrics, 68(1), 29-51.
Atuahene-Gima, K. (2005). Resolving the capability–rigidity paradox in new product innovation. Journal of marketing, 69(4), 61-83.
