Context-capability coalignment in DevOps teams: An action design research approach to improving software delivery performance

Many organizations are transforming their IT service functions in response to changing market demands. Agile software delivery approaches such as DevOps have seen widespread adoption, promising rapid delivery of new functionality and swift resolution of incidents. However, successful adoption of the DevOps paradigm requires not only new tools and structures but also the development of organizational and cultural capabilities tailored to the specific environment of the team. Despite its growing popularity, research and practice still offer limited guidance on how to systematically build and align these DevOps capabilities with contextual drivers.

Based on the principles of dynamic capabilities and contingency theory, the goal of this dissertation is to design and evaluate a context-capability coalignment method for DevOps teams. We adopt an action design research (ADR) approach involving researchers and practitioners. The study comprises two cycles of building, intervening, and evaluating the method. Each cycle was conducted as participatory action research in collaboration with a DevOps team of differing maturity and context within Dutch public organizations. Situational method engineering principles were applied to integrate the method fragments.

The ADR results in two possible configurations of a nine-step situational method and an underlying method base composed of models, tools, and templates that enable the creation of additional configurations tailored to the needs of the target organization. The results suggest a positive contribution of the situational method towards the coalignment of context and capabilities.

We discuss three layers of generalization. First, the problem instance may be generalized to DevOps teams operating in other contexts and maturity levels than those investigated. Second, many of the models and tools presented in this dissertation may be generalized to technology-innovation contexts beyond the immediate problem space. Finally, the dissertation proposes six design principles that convey directions and insights regarding the situational nature of context-capability coalignment and performance measurement. This research contributes to the extant knowledge base by extending perspectives on team performance, contingency theory, and dynamic capabilities into the DevOps domain, and by providing organizations with a method to improve the performance of software delivery teams.