Application Portfolio Management from an Enterprise Architecture Perspective

Abstract

In the past decades organizations have collected up to thousands of different applications supporting their business. This immense growth of the application landscape has been stimulated by advances in technology, mergers, acquisitions and organic growth of organizations. Organizations aim to reduce the complexity of the application portfolio and the related costs, but face a challenging time as their budgets for innovation is typically small. Application Portfolio Management (APM) is considered as the discipline to assess individual and application portfolios to justify their ‘value’, in order to make better decisions regarding the future state. APM is closely related to Enterprise Architecture (EA). EA can be used to unveil the relations between business processes, applications and underlying infrastructure. Moreover, EA can be used to effectively detect misalignments and redundancies in the application landscape. This research aimed to create a new APM method and framework from an EA perspective. Method Engineering (ME) was used as technique to model existing APM approaches from literature, practice and Deloitte. In total, eight existing APM approaches were modeled using ME. A workshop with three experts was organized and functioned as construction instrument and resulted in a new APM method. The proposed APM method is clearly influenced by EA, both in activities as in the roles defined. More than twenty experts were interviewed and provide insight how APM is being applied in practice. Moreover, the interview results led to a set of lessons learned of applying APM. The proposed APM method includes the roles and responsibilities defined for each activity using a RACI model. The deliverables such as the lessons learned and the RACI model complement the proposed APM method and can be considered as the framework. Finally, the proposed APM method was evaluated by three experts and the majority of the method was considered credible.