Minimum viable governance for data science initiatives

Too much governance can stifle innovation in organizations. Too little governance can waste precious organizational resources. Business agility demands empowerment of people to take decisions on initiatives designed to deliver innovative products and services. Traditional monthly and quarterly governance forums such as steering committees and program boards for decision-making potentially impede the flow of work when the delivery of a program or project is done using agile methods in two-weekly sprints. Data science initiatives (DSIs), which are exploratory and innovative in nature, follow agile delivery methods and often require decisions in a different and more frequent cadence. This chapter is an exploratory study of implementing governance for DSIs based on a single case study at Transport for New South Wales. It investigates agile governance at project, program, and portfolio level for DSIs and suggests eight guiding principles focusing on product and portfolio governance. It is targeted at practitioners to guide them in setting minimum viable governance to ensure value is realized from their DSIs and for academics to advance research in governance of DSIs.