Unpacking human systems in data science innovations: Key innovator perspectives

Abstract

Despite optimistic forecasts, industry innovations in data science have extraordinarily high rates of failure. It is essential to minimise the failure of data science projects, for both businesses and data professionals. Human systems are critical to the success of data science innovations. However, the human aspects of innovation management are often neglected or omitted in most guidelines and frameworks for data science. This provides limited guidance about the necessary human conditions for successful data science innovations. In this article we address this concern by developing a systematic framework for understanding human systems that support data science innovations. We first reviewed the elements of human system at different levels of analysis and how they contribute to innovation. Substantial research and theory indicate successful innovation requires supports at different levels of organisations, which combine to create the organisation’s innovation capability. The review provided an initial framework for integrating human systems with data science innovations. Then, we drew on a series of interviews with key innovators engaged in developing current data science innovations. The interviews generated a more complete picture of human systems in practice. The findings support a range of practices to energise and facilitate innovation as an integral part of strategic planning and business processes. This study contributes to the advancement of innovation management theories and calls attention to guiding and engaging individuals through providing support and removing barriers to data science at different organisational levels. More practically, innovation managers could use this as a guide to optimise work systems and inform pathways to improve organisation data science efforts.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 9 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.