Investigation of the Correlation Between Initial Parameters and Success Factors in Completed Investment and Industrial Plant Construction Projects

This thesis develops the conceptual foundation for PEER (Project Evaluation, Expectation, and Recommendation), an early-stage project success prediction tool for industrial plant construction. The goal is to identify structural risks, behavioral weaknesses, and early warning signals before execution starts, using the limited and often qualitative information available during early project phases. Unlike traditional project management tools that depend on finalized budgets, schedules, or detailed work breakdowns, PEER is designed to work under uncertainty, providing structured recommendations based on early inputs. To prepare the concept behind PEER, traditional forecasting methods such as PERT, CPM, EVA, and Monte Carlo Simulation were reviewed. These methods, while useful during later execution stages, were found unsuitable for early-phase applications. Their reliance on finalized project structures and complete cost frameworks prevents them from supporting decisions when uncertainty is still high. Because of these limitations, PEER was developed to close the gap. It is not designed as another tracking or control tool, but as a system that interprets fragmented project information, flags structural weaknesses, and provides early, practical recommendations before risks escalate. More than seventy project parameters were defined and grouped into areas such as contract setup, innovation intensity, client behavior, stakeholder complexity, cash flow structure, and external governance conditions. A dataset of 47 completed industrial projects was manually analyzed, focusing on profitability evolution, value shifts, schedule changes, cash flow effects, and client-side project management competence. The analysis showed clear patterns between structural setups and project outcomes and confirmed that early project structures have a major influence on financial and schedule performance. The concept for PEER is based on pattern recognition, risk grouping, and proactive recommendations, rather than deterministic forecasting. It focuses on signals that are usually missed or ignored during early traditional planning. The thesis concludes that early-stage project success prediction is possible when structured reasoning, behavioral indicators, and historical patterns are combined. PEER offers a first step toward more proactive project intelligence, focusing on the phase where decisions have major impact on outcome and costs. Future work will include expanding the dataset, building a digital prototype, testing the system on live projects, and developing a financial extension module to integrate preliminary economic evaluations alongside structural assessments.