The goal of the thesis is to contribute to reducing the estimation error in software development projects by coming to a better understanding of the shortcomings of, and how to improve, expert judgment-based processes of effort estimation. The following topics are addressed: • When, and how much, do judgmental biases and inconsistencies in effort estimates impact the estimation error? • What is the software clients' role in reducing effort estimation error in software development projects? • What is the state of practice of software effort estimation error measurement and analysis?