Software Project Managers are often faced with the dilemma of estimating the cost and duration of new projects and enhancement requests. The software industry is relatively new, and surveys conclude that software size and effort estimation is a challenge and needs research solutions. In the last three decades, research has come up with a substantial amount of analysis and study resulting in a variety of software estimation models. Recently, with the onset of Machine Learning and AI related technologies, numerous models have been evidenced that provides better technical accuracy to the problem of software estimation. At the same time, research still evidences that Estimation based on Expert Judgment remains the most widely used methodology. While estimating smaller units of work as in an enhancement or maintenance project is less error-prone, previous research does not report enough accuracy. The objective of this research is to bring a structure to the EJ methodology and bring in consistency and accuracy to an otherwise human-centered and intuition led approach. We employed a lean and straightforward work break down (WBS) to each of the enhancement requests that were put up to a support team for estimation purposes. We evidence that our approach of employing this work-breakdown can bring in two-fold improvement to the estimation process while using expert judgment. This approach also ensures a stable estimation process as evidenced through experiments using Statistical Process Control techniques. We applied our methodology to one of the projects in an IT organization and collected data from two years of operations. Comparing our results with other previous studies, we were able to reduce the error to 25% of the value of this metric, while more than doubling the accuracy of predictionat PRED(25). © BEIESP.