- Service-Oriented Architecture (SOA): An Empirical Evaluation of Characteristics, Adoption Determinants, Governance Mechanisms, and Business Impact in the German Service Industry (2012)
- Since the first introduction of the term Service-Oriented Architecture (SOA) by Gartner analysts in 1996, the idea of service orientation of IS architectures including SOA and its supporting technologies has become a widely adopted concept both in industry and research. SOA leverages the principles of object-oriented and component-oriented programming not only for the development of single systems but for entire system landscapes by adding a new service layer consisting of loosely coupled modular services. In recent years, a number of definitions for SOA appeared that varied with respect to their scope. Researchers applied the concept of service orientation not only for designing business application systems (i.e., a narrow scope of SOA), but also for designing the business architecture and so introduced the concept of a “Service-Oriented Enterprise (SOE)”. Early SOA research focused on technical aspects regarding the characteristics of SOA. However, while providing a solid technological base for implementing SOA, the business perspective on SOA has been largely neglected and gained increasing attention in recent years. Even though recent research has already increased the understanding of SOA from a business perspective, researchers struggle to provide theoretical foundations as well as to apply quantitative empirical research to investigate the relationships on a larger empirical basis. The cumulative dissertation thesis at hand addresses these areas by drawing on a multi-theoretical foundation and applying a multi-method approach in order to address the following research directions: Adoption: Why and under what circumstances do organizations introduce SOA? Business impact: What are the benefits resulting from SOA? Governance: How should SOA introduction and management be governed to achieve the benefits? These research directions are targeted by the eight research articles included in this dissertation thesis following a quantitative approach. First, the foundation for the thesis is laid by reviewing the literature and identifying research questions. Next, nine case studies were conducted to understand SOA adoption, its implications for the organization as well as the used governance mechanisms in depth, and to refine the research models based on further insights from practice. Third, necessary prerequisites for applying quantitative methods are developed in terms of conceptualizing SOA adoption and developing measurement instruments for SOA adoption and SOA maturity based on their characteristics. Based on this preliminary work the developed research models that draw on different theoretical foundations – e.g., modular systems theory, dynamic capabilities theory (DCT), or the technology-organization-environment (TOE) framework – were tested using quantitative methods. Besides different theoretical foundations, the individual papers leverage different statistical analyses, such as structural equation modeling (SEM), factor analysis, or cluster analysis. For our quantitative analyses, we used the data of 247 participating organizations collected through a survey carried out in the German service industry. The results of the analyses are documented in eight research articles that are partly published and partly under review at international journals. Five different approaches to SOA adoption are identified: no SOA adoption, service-oriented business architecture, medium extent among all four SOA dimensions, medium extent with high service orientation of IS architecture, and high extent among all four SOA dimensions. SOA adoption is based predominantly on organizational prerequisites such as IT employees’ IT expertise, top management support, and the compatibility of SOA with an organization’s IT setting. Further determinants are relative advantage, costs, and organization size. Notably, the complexity of implementing SOA is seen as a cost driver but not as an inhibitor. The results show that an organization’s SOA adoption is also influenced by observable SOA adoption from other firms (“same group”, i.e., management fad) and from media, consultants and other sources (“outside group”, i.e., management fashion). This thesis offers a theoretical and empirical explanation for SOA’s business impact. The results substantiate the oft-proclaimed business value discussion of SOA by showing quantitatively significant relationships between the degree of SOA and business-relevant value dimensions in terms of reducing costs and increasing business agility, data quality, process monitoring, internal straight-through processing (STP), and business-to-business (B2B) integration of processes. Also, we offer a theoretical foundation by drawing on the concept of dynamic capabilities to argue theoretically and to show empirically that IT flexibility is a meaningful explanatory factor for SOA’s business value. Further, the results show a diminishing marginal utility of increasing SOA maturity. In addition, this is the first study that conceptualizes and empirically evaluates a measurement instrument for business process management (BPM) and offers first insights into how BPM and SOA jointly affect business process quality. Moreover, SOA is identified to be a good foundation for business process outsourcing (BPO) as it increases related benefits and flexibility while reducing associated risks. This thesis offers an evidence-based contribution to the discussion of the role of SOA governance mechanisms and a validated measurement model that is useful when bringing together managerial and technical perspectives regarding service orientation. Notably, establishing new SOA decision-making bodies, such as the often called for SOA centers of excellence (CoE), hampers modularity, integration and reuse in the beginning. Thus, one important managerial result is that organizations do not necessarily need to implement new decision-making bodies but should prefer adapting existing ones. Among others, standards, employee qualification, and IT/business communication are important for these purposes. Managers pursuing better IT scalability are also well advised to follow this recommendation, as these three mechanisms (but only these of the investigated) also drive scalability. For improving modularity, clear service development processes and the collaborative work of business units are applicable governance mechanisms. Further, the findings suggest that managers should not hope for better alignment automatically following the SOA implementation, but instead explicitly implement management actions that facilitate collaboration between business and IT as this further improves the outcomes of implementing SOA. Answering these research questions contributes to theory in multiple ways by offering theoretical and empirical explanations for the determinants of SOA adoption, governance mechanisms, and SOA’s business impact, as well as providing a newly developed multidimensional SOA research concept and a measurement instrument for SOA maturity. Moreover, answering these research questions also contributes to a topic of high practical relevance. According to Forrester Research, 84% of the global 2,000 enterprises (i.e., organizations with 20,000 or more employees) use SOA. Despite the high rate of SOA adoption among very large organizations the same study shows that only 12 percent of the global 2,000 enterprises achieve the intended SOA benefits. Thus, insights into questions surrounding important adoption determinants for implementing SOA, SOA’s realistic potential business impact, and effective SOA governance mechanisms are of particular importance for managers, especially given that typically there is a high cost associated with introducing SOA.