From data to impact: A multivocal review of success factors and criteria for data science projects
Abstract
Background. Organizations increasingly invest in data science projects for their transformative impact potential, but high failure rates highlight the necessity of effective success management strategies to realize data’s strategic value. This study aims to support such strategies by identifying and synthesizing two key dimensions: critical success factors (CSFs) and success criteria.
Methodology. A multivocal literature review (MLR) of 3,134 academic and grey sources was conducted, from which 57 high-quality publications were synthesized. To improve practical relevance, findings were refined through an expert panel of data science professionals. The CSFs were organized using the P-DOTS (Project, Data, Organization, Technology, Strategy) framework for comprehensive coverage. Success criteria were grouped into six outcome-focused categories and paired with example metrics to support evaluation.
Results. Key CSFs, such as goal setting, team capability, and stakeholder involvement, emerged as universally significant across data science and related domains. Generic factors often hold equal or greater importance than data-specific ones like data quality, access, and governance. Our analysis of success criteria confirms a shift in evaluation focus, from traditional project performance toward criteria such as stakeholder satisfaction, product quality, and value delivered. Overall, this study provides an extensive overview of the current research landscape and expert-validated success dimensions.
Conclusions. This is the first MLR to investigate success dimensions in data science projects, bridging academic and industry perspectives. The identified dimensions provide a practical basis for defining, evaluating, and improving project outcomes and establish the building blocks for success management approaches tailored to data science.