Connecting Organizational Goals to Smart Manufacturing: A Framework for Minimum Effective Data

Abstract

The rapid adoption of Industry 4.0 technologies has led to a substantial growth of sensors and connectivity in manufacturing systems, resulting in the generation of high-dimensional, memory-heavy datasets. Despite this abundance of data, many manufacturers struggle with data overload, poor utilization, and fragmented data infrastructures, which hinder the deployment of advanced analytics and trustworthy AI. As the sector transitions toward Industry 5.0, with its emphasis on human-centric, resilient, and sustainable manufacturing, the challenge is no longer how to collect more data, but how to identify and exploit the minimum set of data that meaningfully supports decision-making. This paper addresses the central research question: Is there a generalized framework to systematically identify and extract the minimum “smart data” required for specific manufacturing performance indicators? Through a structured review of recent academic and industrial literature, the paper evaluates emerging concepts including smart data, targeted data, minimum effective datasets, AI-driven feature selection, edge-based data filtering, and human-in-the-loop analytics. Building on these insights, the paper proposes a generalized, systematic framework grounded in the Hoshin planning principle. The framework links strategic manufacturing objectives to operational metrics and data requirements, ensuring consistent, goal-aligned data minimization across all organizational levels. The key takeaway is that a principled shift from indiscriminate data accumulation to minimum effective data enables improved model performance, greater trust and interpretability, reduced system complexity, and enhanced human-centric decision-making. The proposed framework aims to close a critical gap between fragmented research and practical industrial application, offering a scalable foundation for next-generation smart manufacturing.