1. Introduction
Modern firms struggle to keep up with the rapidly changing technology and business landscape [
1,
2]. This is where the concept of Enterprise Architecture (EA) comes into play. Analyses of Gartner show that EA has now reached the phase ‘Climbing the Slope’ on the Hype Cycle [
3]. What this means is that we better understand what benefits EA can bring to the enterprise. However, EA’s broad market applicability and relevance are not yet clearly paying off. EA’s focus has now changed toward EA-enabled business models and operations, and delivering value under continuously changing conditions is now center stage.
EA is typically conceptualized as the firms’ organizing logic or blueprint [
4]. As such, it describes “what is going on in the firm” right now (often described as the “as-is” situation) in terms of data, process, and information systems (IS) and information technology (IT). EA is also used to describe “what should be going on” (often described as the “to-be” situation) in the business and IS/IT landscape following the firms’ ambitious (digital) strategies and goals. Based on the unfolding gap-analysis, EA provides a roadmap with accompanying agile IS projects and programs to achieve this target from the current state [
4,
5,
6]. Firms are currently embracing EA to leverage their digital investments and facilitate flexible integration of IS/IT assets and resources with business processes to obtain an advantage over competitors [
6,
7,
8]. In practice, EA facilitates firms to translate business strategy into design, daily operations, and master emerging complexity across the enterprise [
6]. EA can be a valuable asset to the firm, as it can unlock the true potential and business value of all firm’s digital initiatives that require enterprise-wide integration of large numbers of heterogeneous and frequently changing systems and information structures [
9]. EA has become crucial for firms that operate in turbulent business environments. Recent work showcases the central role of EA during firms’ digital transformation, where it actively supports decision-makers in making adequate decisions concerning the radically changing business and IT landscape as part of the digital transformation [
10,
11].
However, despite valuable scholarly contributions in the EA domain, there is still a pressing need for empirically validated work that advances our current understanding of EA benefits and value created by EA-driven capabilities [
12]. These capabilities mobilize business, IS/IT assets, and resources in alignment with the firm’s strategic objectives [
4,
13,
14,
15,
16]. The lack of empirically validated work is problematic as the current knowledge-base only delivers provisional conceptions on how EA-based capabilities cultivate organizational change, innovation, and business and IS/IT benefits [
17,
18,
19,
20].
We ground this work within the dynamic capability view (DCV), a leading strategic management framework [
4,
7,
21], and argue that the firm’s innovativeness—the ability to introduce innovation to a firm’s business processes and ability to use the latest technological innovations for new product development [
22,
23]—depends on its EA-driven dynamic capabilities. Similar to Shanks et al. [
4] and Van de Wetering [
16], we consider such EA-driven capabilities as dynamic capabilities [
21,
24]. These dynamic capabilities help firms sense possible business and IT opportunities and transform and deploy these initiatives and opportunities while ensuring that their assets and resources align with the strategic goals and market needs [
4,
16,
21,
24].
The IS and management scholarship has evolved considerably in recent decades. However, there seems to be consensus concerning the key attributes of successful firms in fast-changing economies. Scholars showed that firms’ dynamic capabilities and organizational design are two salient and strategic factors that profoundly affect firms’ innovativeness levels [
25,
26,
27,
28,
29,
30,
31,
32]. These studies support the view that designing organic, firm structures is a crucial strategic choice for firms that complements dynamic capabilities as a key driver of the innovativeness of firms [
29,
30,
33]. Hence, in addition to the EA-driven dynamic capabilities construct, in this study, we, therefore, extend the core argument and claim that the firm’s organic, firm structure influences the firm’s innovativeness.
Organizations that adopt an organic and decentralized structure are typically more innovative than those with a rigid and formalized structure [
34]. Such organizations embrace a culture of informality with decentralized decision making, resulting in their ability to be agile and quick in sensing external business environments, seizing the opportunities and reconfiguring their resources to improve existing products, processes, and services or develop new ones based on the latest technologies [
22,
23,
35]. Hence, a firm’s organizational structure is an important concept to consider alongside EA-driven dynamic capabilities in exploring the influence of a firm’s innovations on organizational benefits.
Hence, this study unfolds the critical intermediate abilities and organizational capabilities (innovativeness) in the value path consistent with previous, dynamic, capability literature, showcasing the direct and indirect effects of dynamic capabilities on other organizational benefits [
36,
37,
38].
Therefore, the current study aims to address the pressing need for empirically validated work in this particular domain and tries to deliver a foundational concept of how EA-based capabilities contribute to the firms’ benefits, and, thereby, enhances our understanding in four ways. First, we unfold the theorized relationships between EA-driven dynamic capabilities and innovativeness using data from 299 CIOs, IT managers, and lead enterprise architects of Dutch firms. The firm’s innovativeness (partially) mediates the relationship between EA-driven dynamic capabilities and organizational benefits. Second, our study shows how firms that have embraced organic organization structures and, thus, a culture of informality and decentralized decision-making will be better equipped to enable EA-driven value activities and, thus, inventiveness. Third, this study unfolds the path through which dynamic capabilities add value to organizational benefits. Finally, it is essential for modern firms to co-evolve their business and IT resources and capabilities to maintain a competitive edge [
39]. Hence, we also investigate the particular conditions and circumstances under which the firm can unlock EA’s value and drive digital and process innovations given their available valuable, rare, inimitable, and non-substitutional (VRIN) resources [
40] and the firm structure [
41,
42,
43,
44]. From the DCV, it can be deduced that firms need to design their organization in such a way as to build dynamic capabilities for innovation [
29,
45,
46]. However, there is no consensus in the literature on how this inter-relationship looks like [
29,
47,
48]. It is evident that there needs to be coherence between dynamic capabilities, resources, and the organizational structure to drive innovativeness. Therefore, this study opts for an appropriate practical methodology that rigorously discovers complementarities between these elements and how they—as patterns—lead to innovativeness [
26,
49].
Based on the above four objectives, this study addresses the following three research questions:
- (1)
To what extent do the firm’s EA-driven dynamic capabilities and organic firm structure influence its level of innovation?
- (2)
To what extent does the firm’s innovation level impact organizational benefits?
- (3)
Which unique configurations of EA-driven dynamic capabilities shape a firm’s innovativeness?
This paper is organized as follows. First, we outline the theoretical background and review the core theories relevant to our work. Then, we synthesize the core literature on EA-based capabilities and, subsequently, develop the hypotheses that underlie the research model. Next, we describe the empirical study, including the data collection, analyses, and the measures used in this study. Finally, we outline the work’s empirical results by first confirming our model’s reliability and validity and then testing the developed hypotheses by drawing on a sample of 299 CIOs, IT managers, and lead architects. This study continues with a fuzzy-set qualitative comparative analysis (fsQCA) [
50,
51] to unfold the particular circumstances in which an organic firm structure is particularly relevant for firms. Finally, we discuss our study findings and conclude the study.
5. Quantitative Data Analysis
The model’s hypothesized relationships were tested using Partial Least Squares (PLS) structural equation modeling (SEM). We used SmartPLS version 3.2.9. to estimate and model parameters [
122]. PLS-SEM assesses both the measurement model, i.e., outer model [
128] and the structural model (i.e., inner model) of the research model [
125]. The PLS algorithm establishes latent constructs from factor scores and PLS, thereby, avoids factor indeterminacy [
129]. Hence, scores can then be used in the following analyses [
130]. PLS-SEM is appropriate for our analyses as our focus predicts as to whether the PLS algorithm assesses the explained variance (
R2) for all dependent constructs [
129].
Figure 2 summarizes the structural model tests and the hypotheses testing with
R2, their associated predictive values, the regression coefficients, and the associated T-values. As can be seen in the figure, all hypotheses can be confirmed as the path coefficient was significant while controlling for the non-significant effects of ”size” (
β = 0.0021,
t = 0.353,
p = 0.72) and ”industry” (
β = −0.0011,
t = 0.204,
p = 0.84). The estimated effect sizes (
f2), i.e., the specific contribution of exogenous constructs to endogenous latent constructs, are for EA-driven dynamic capabilities
f2 = 0.21, organic firm structure
f2 = 0.09, and
f2 = 0.30 for innovativeness.
To assess whether or not innovativeness fully or partially mediates the effects of EA-driven dynamic capabilities and organic firm structure on organizational benefits, we followed mediation guidelines [
125]. Therefore, for EA-driven dynamic capabilities, the direct effect (thus, without innovativeness in the model) was positive and significant (
β = 0.33,
t = 5.855,
p < 0.00001). This outcome fulfills the basic mediation condition [
131]. Additionally, the direct effect of an organic firm structure on organizational benefits was positive and significant (
β = 0.22,
t = 0.353,
p = 0.010). In a subsequent step, we included innovation in the model and assessed the significance of the indirect effects (i.e., mediating paths) integrally (thus including all mediating paths) through a bootstrapping approach using a non-parametric resampling procedure [
125,
132]. At that point, results showed a less strong, but still significant relationship for the direct path (EA-driven dynamic capabilities → organizational benefits) (
β = 0.120,
t = 2.665,
p = 0.008). For an organic, firm structure (organic, firm structure → organizational benefits), this outcome was insignificant (
β = 0.081,
t = 1.350,
p = 0.177). Hence, it can be concluded that innovativeness partially mediates the effect of EA-driven dynamic capabilities as the direct and the indirect effect aim in the same direction (both positive and significant). The outcomes show that an organic firm structure is a key enabler of innovativeness, and that innovativeness fully mediates the effect of an organic, firm structure on organizational benefits.
Further results show that the included control variables showed non-significant effects: ”size” (β = 0.004, t = 0.067, p = 0.95), and ”age” (β = −0.020, t = 0.362, p = 0.72).
Result show that the model explains 2% of the variance for innovativeness (
R2 = 0.22) and 33% of the variance for organizational benefits (
R2 = 0.33). These particular effect sizes can be classified as moderate to large [
125]. Finally, we used a blind-folding procedure in SmartPLS to evaluate the model’s predictive power [
125]. Obtained Stone-Geisser values (
Q2) for the endogenous latent constructs exceed 0, thereby showcasing predictive relevance [
125].
6. FsQCA Configurational Analyses
This study employs fsQCA [
50,
51] to gain insight into the particular circumstances in which the organic firm structure is particularly relevant for firms. It adheres to the fact that an organic firm structure had an unusually small effect size. Moreover, VRIN resources are included as tightly associated with dynamic capabilities and can provide firms with a durable, competitive advantage [
40,
133].
FsQCA is a configurational approach that complements traditional regression-based approaches (including SEM) in the process while showing the particular conditions under which an outcome of interest is obtained in the data. Hence, it does so by examining the specific asymmetric relationship between various (antecedent) constructs and specific outcomes [
50,
134]. A single configuration can be defined as a specific combination of antecedent conditions and factors present in the data so that high levels of an outcome (i.e., innovativeness) are obtained [
51]. FsQCA allows the predictor and outcome variables to be on a fuzzy scale rather than on a dichotomous (binary) scale. Furthermore, it enables the reduction of elements for each pattern. Therefore, configurations only include necessary and sufficient conditions. Therefore, a distinction between core, peripheral, and “do not care” aspects can be made. Within solutions, core elements have a strong causal condition with the selected outcome measure. For peripheral elements, there is a weaker association with the outcome [
135]. As the first step, we defined the outcome and independent measures and calibrated them accordingly into fuzzy sets. These particular sets ranged from 0 to 1, with 0 indicating the absence of a set membership, while 1 denotes full membership. We used fsQCA 3.0 software [
136] to calibrate the items and, likewise, used this procedure to set the membership based on three particular anchors of memberships using a seven-point Likert scale [
137]. Hence, we followed particular guidelines for the process of generating fuzzy set-membership measures [
138,
139] and defined ‘6’ as the full membership anchor (fuzzy score = 0.95), ‘3’ as the anchor value for full non-membership, and ‘4.5’ as a crossover point (fuzzy score = 0.50). The anchor for full non-membership was placed on 3 (fuzzy score = 0.05) instead of 2 due to the distribution of measurement values and the need to adjust scores to respondents’ scores [
140].
After the calibration process, the fsQCA software runs an algorithm to produce a truth table. This table includes all possible combinations of elements, and a row corresponds to a single combination. The number column highlights the frequency of cases of each combination. We have set a minimum of three cases and consider combinations with at least three empirical instances for configurational analysis. The degree of consistency is set to a recommended threshold of 0.75 [
135,
136]. Consistency is a value that ranges from 0 to 1 and reflects the degree to which a set-theoretic combination leads to an outcome [
135], or consistency is analogous to a correlation in statistical analysis [
140]. Coverage concerns the empirical relevance of a consistent subset and helps determine the percentage of the outcome covered by the solutions [
136]. Solution coverage indicates how much is covered by the solution terms and is comparable to the
R2 value [
140]. Raw coverage indicates which alternative path can explain a particular percentage of the outcome. Unique coverage indicates the proportion that uniquely covers a specific outcome [
136,
141]. The obtained consistency and coverage values exceed the minimum thresholds [
136]. A truth table algorithm is then applied to reduce the various combinations into a smaller set of configurations and identify various holistic, interconnected, equifinal solutions that are associated with innovativeness as an outcome [
26,
50,
138].
Table 6 shows the fuzzy set analysis for high levels of innovativeness. The depicted black circles (⚫), and, in this case, core elements denote a condition’s presence, the small crossed-out circles (⊗) indicate the absence of a particular element in the solutions, and peripheral elements and blank spaces indicate a “don’t care” situation where causal conditions may be present or absent [
50,
136,
142].
Outcomes show that achieving innovativeness levels stem from different combinations of capabilities, VRIN resources, and their interplay with the organizational structure. Specifically, the fsQCA results show seven possible solutions (i1-i7). The results showcase that at least one and, in one case, even three EA-driven dynamic capabilities are present as a core element, strengthening the previously outlined results. The first solution applies (i1) to firms that operate under conditions with strong VRIN resources and EA transforming capability, and the absence of EA sensing. Firms capitalize on VRIN resources by reconfiguring business processes and IS/IT rather than focusing on sensing and identifying new business opportunities.
A similar solution is present in i4. Under these conditions, firms should seek to mobilize and seize business opportunities using EA and strong VRIN resources, given that their organizational structure is decentralized and less formal. Solution i2 and i3 apply to firms that have more formalized centralized structures. Under these conditions, firms must develop EA-driven dynamic capabilities to reconfigure business operations as a means to achieve innovativeness adequately. Solution i5 illustrates that innovativeness can be achieved in the presence of organic firm structures and robust EA sensing and mobilizing capabilities. This solution also applies to firms that operate under conditions with an absence of VRIN resources. The firms should seek new innovative business solutions. Solutions i6 and i7 are independent of the organizational structure. For solution i6, VRIN resources, combined with mature mobilizing and reconfiguring capabilities, are crucial in obtaining high innovativeness. Solution i7 shows that innovativeness can be attained with strong EA-driven dynamic capabilities.
7. Discussion
The current study aimed to unfold the theorized relationships between EA-driven dynamic capabilities, innovativeness, and organizational benefits. It also investigated the strategic role of organic firm structures as a driver of innovativeness and a culture of informality and decentralized decisions. Moreover, we wanted to understand the particular conditions and circumstances under which firms can unlock EA’s value given the firm’s available VRIN resources, EA-driven dynamic capabilities, and the firm structure. Following the dynamic capability-based view, we operated a research model and, subsequently, tested the associated hypotheses using cross-sectional data from 299 executives and senior practitioners and found support for the hypotheses. This study also tried to unfold the unique conditions under which firms’ innovativeness levels are obtained through different combinations of dynamic capabilities, VRIN resources, and their interplay with the organizational structure.
This current study has various theoretical and practical implications. First, our findings support our study’s hypotheses and show that EA-driven dynamic capabilities are crucial for organizational benefits through the firm’s innovativeness. This outcome is important as the literature did not fully grasp an adequate understanding of the value-creating process using EA and EA-based capabilities [
4,
14,
16] and the relationship between dynamic capabilities and innovativeness [
87]. Second, another significant result of this study is that we unfolded—using fsQCA—the particular circumstances in which an organic firm structure is particularly relevant for firms and complementary with dynamic capabilities and the firms’ VRIN resources. We show various contingent solutions and alternative paths that drive firms’ innovativeness when particular conditions are present or absent. Hence, achieving high innovativeness levels stems from different combinations of dynamic capabilities, VRIN resources, and their interplay with an organizational structure. This outcome is an essential contribution to the literature. These concepts have predominately been studied using variance-based approaches [
4,
40,
48], thus, neglecting possible specific combinations of antecedent conditions and factors present in the data [
143].
This research suggests two major practical implications. First, the results imply that executives and senior practitioners should actively invest in EA-driven dynamic capabilities as crucial competencies and routines to drive the firm’s innovativeness and strive for higher organizational benefits. The outcomes support the idea of having a more elaborate and coherent perspective when it comes to firm innovativeness and to obtaining higher organizational benefits. Decision-makers should specify three EA-driven capabilities (i.e., sensing, mobilizing, and transforming) as they provide a means to drive the firm’s innovativeness and enhance its evolutionary fitness. Improvement initiatives should not be deployed in isolation, as they will then be unlikely to achieve the desired outcomes since the impact of these complementary practices will be greater than the sum of its parts [
144]. The outcomes facilitate decision-makers with factual business scenarios to achieve innovativeness with their situational capabilities, resources, and organizational structure.
Some limitations of the current study are acknowledged. First, we did not investigate the environmental conditions’ role that could affect the model’s effects. Second, this study did not test potential differences between sample (sub)groups (and their interactions). Additionally, we did a measurement at a single point in time. Therefore, it is difficult to truly establish causality as a firm’s innovativeness and organizational benefits may vary over time. A longitudinal approach could enrich our perspective by providing valuable insights into the study’s evolutionary nature’s construct over time as punctuated equilibrium models [
145]. Finally, we used self-reported measures, and triangulation with archival data could further strengthen the outcomes. However, perceptual data is typically associated with objective measures [
146].