A qualitative study of success factors for portfolios of agile projects in large organisations

Purpose Agile project management (APM) has significantly changed the landscape of software development, and IT projects in recent years. To embrace agility, large organisations scale up their agile methods and practices and integrate them into the top-down traditional hierarchical processes from where the project portfolio is managed. While agile methods proved to be extremely efficient in individual and small projects, they can impose significant challenges at the portfolio level of large organisations. Agile projects largely increase the level of team interactions and necessitate a customer focus that makes the management of portfolios of agile projects highly challenging and complex. Despite the importance given to the scalability of agile methods by practitioners, little research has examined portfolio success when portfolio comprised multiple agile projects. Also, while some critical success factors (CSFs) are proposed for individual agile projects in the existing literature, but the effects multiple agile projects have on portfolio success is mostly unexplored. Therefore, this research aims to explore portfolio success and its influencing factors in large organisations, where multiple agile software development and IT projects are governed by project portfolio management (PPM). The researcher also explored agile scaled practices that are employed by case organisations at their portfolio level. Design/methodology/approach The research approach taken is multiple qualitative case study via an interpretivist research philosophy. Using project contingency theory (PCT) as a lens, the study is based on a thematic analysis of 14 semi-structured interviews from the agile project and portfolio practitioners within eight case organisations. All case organisations were practising PPM and using APM for their software development and IT projects. The interviews were audio-recorded and transcribed. The transcripts were inductively coded using NVivo 12. Then, the findings were analysed and compared using visual mapping strategy to find common factors, themes and patterns. Findings The data analysis identified a set of 26 CSFs for the agile project success, and portfolio success, out of which 13 were contextual, and 13 were at project-level. Each one of the identified CSFs connected to both agile project success and portfolio success, either directly or indirectly. Among all factors, nine are new (six contextual and three project-level) while the rest find support in previous literature and are validated for use at scale. The new contextual CSFs for agile project success are appropriate funding model, training customer and flat organisational structure. The new contextual factors for portfolio success are APM-PPM synchronisation, team/customer involvement and IT and infrastructure support. The new project-level CSFs that connected to portfolio success are team autonomy, agile delivery metrics and sufficient amount of work. The research also found nine exclusive agile scaled (portfolio) practices at the portfolio level of large organisations. These practices are portfolio wall (or board), portfolio stand-up (or SoS), planning sessions, portfolio backlog, project discovery, sprint review (showcase), roadmapping, Spotify practices and portfolio retrospective. The identified factors and practices were then mapped into an agile project-portfolio success framework. Research limitations The findings in this thesis are based on a multiple case study of eight large organisations from a diverse range of industries which allowed an effective cross-case analysis. The research participants were all experienced and knowledgeable in the field, but they were working in large organisations, and their standpoints might not represent the organisations. Also, agile technical CSFs were deliberately excluded from the study. Therefore, the emerged framework can be further expanded by identifying more factors. Further research should investigate the relationship between the identified CSFs and determine the extent to which each factor can affect the agile project and portfolio success. The emerged framework needs to be tested employing a quantitative method with a larger sample set to allow the proposed model to be tested empirically. Practical implications The study findings have two clear implications for managers and professionals. Firstly, the emerged framework and its embedded CSFs can serve as guidelines for large organisations planning to adapt and embrace agile methods and practices in their projects and portfolios. Secondly, the identified CSFs support organisations in making agile scale-up decisions. Checklists, metrics and benchmarks can be set based on the proposed CSFs enabling organisations to assess the effectiveness of performed actions and practices during and after the adoption of agile methods at scale. Originality/value The theoretical research contribution has three distinctive aspects. Firstly, the set of new and confirmed CSFs, agile scaled (portfolio) practices and the emerged framework contributes to the theoretical conceptualisation, applicability and clarity of agile methods use; the areas that are found controversial and problematic before. Secondly, the research setup integrates two specific domains, i.e. agile projects and portfolio(s), instead of focusing on one area only. This integration allowed a broader systemic context and provided a more comprehensive view of how portfolios of multiple agile projects are managed in large organisations. Thirdly, the study shed light on key contingency factors (CSFs) that influence the strength of agile project success and portfolio success relationship. This study appears to be among the first that applied PCT as a lens to address agile methods used in portfolios of large organisations. Various organisational elements should optimally fit the multiple contingencies identified in the research to ensure agile project success and further portfolio success. Although the emerged framework does not prescribe solutions concerning the fit of organisational variables to identified contingencies, it can be viewed as a good starting point for more exploratory studies.