Revisiting Cost Forecasting Methods in Construction Projects

Abstract

Accurate cost forecasting is fundamental to construction project management, yet the basis for selecting appropriate forecasting methods across different project conditions remains insufficiently articulated in the literature. This paper presents a structured conceptual review classifying construction cost forecasting methods into five groups: deterministic and rule-based, probabilistic and uncertainty-based, process- and physics-based, data-driven and machine-learning, and hybrid and integrative. A comparative SWOT analysis examines the analytical strengths, weaknesses, opportunities, and threats associated with each method group. A contingency-based method-selection framework is then developed, supported by six testable propositions linking project context, method choice, and expected forecast performance. The central argument is that method selection should be governed by methodological fit with the project context, including in off-site, modular, and industrialised construction, where cost behaviour reflects logistics, and supply-chain dynamics alongside site-based variability.