28th International Congress on project Management "Project Management in the Digital Decade"

Data is increasingly pervasive in organisational contexts, shaping investments from inception to benefits realisation. Data Science Initiatives (DSIs) have gained traction to unlock value from data, yet their success rate remains a concern, with estimates suggesting a significant failure rate. This paper employs a multi-method approach, utilising six DSIs as case studies within Transport for NSW, supplemented by semi-structured interviews with practitioners across various Australian organisations. The study reveals that DSIs possess unique characteristics that render traditional ICT-enabled program management practices ineffective. It offers practical guidance for practitioners to enhance DSI delivery efficiency. Furthermore, the paper delineates three key stages – Exploitation, Transition, and Exploration – characterising the delivery of DSIs. Theoretical contributions extend to the exploration of uncertainty within data science projects, elucidating their impact on DSI success or failure. Future research avenues may entail validating these findings across diverse public and private sector organisations internationally.