Into the Unknown: Synthetic Confidence Intervals for Sustainability Trade-Off Analysis in Early Product Development

Sustainability assessment in early-stage product development is challenged by high uncertainty, limited structured information, and a lack of decision-support tools that help anticipate systemic trade-offs. Established methods such as Life Cycle Assessment typically apply retrospectively and are often unavailable or unsuitable at this early stage, when critical design decisions must already be made. This paper presents a method to support sustainability-aware decision-making by generating and visualizing synthetic confidence intervals. In response to the inherent complexity and interdependence of factors in early development, the approach systematically varies known design parameters to create plausible contextual scenarios. These scenarios reflect a spectrum of plausible design contexts that extend beyond the narrow scope of typical empirical datasets. The resulting variations in sustainability impact are used to build index-scaled confidence intervals, which are visualized in bar charts indicating expected positive to negative effects across multiple sustainability indicators. The visual format enables rapid interpretation and supports intuitive recognition of potential sustainability trade-offs. It is designed to complement limited explicit knowledge by aiding developers in reasoning under uncertainty and informing early-stage design decisions. Users can specify a desired confidence level, such as 80, 90, or 95 %, to control the robustness of the interval and filter out statistical outliers. The method is demonstrated through a case study that evaluates the plausibility and interpretability of the results. Rather than prescribing precise outcomes, the approach provides orientation within complex design spaces. It functions as a form of methodological guardrail that enables more informed and context-sensitive sustainability reasoning when information is limited but decision freedom is high. Its structured yet flexible logic holds promise for broader application in complex systems design, where conventional data-driven methods fall short in early development stages.