Early effort estimation in web application development

Project planning in software industry represents one of the most complex tasks, especially when there is a need to estimate the time, cost and effort needed for development of software projects. In the field of development effort estimation for classical software projects a number of methods have been developed, tested and successfully implemented. Web projects are, by their nature, different than classical software projects, and there is a lack of methods and models that provides a high degree of confidence in development effort estimation. This paper analyzes the possibility of using a combination of functional size and conceptual models for the purpose of web application development effort estimation. Measurement of functional size can be effectively applied to the conceptual models of the data-driven web applications because of the existence of extensive count of data movements. For the purpose of this study 19 web applications with their conceptual models were employed. An effort model was built using simple linear regression analysis. Upon construction, evaluation and validation of the effort model prediction accuracy elements, R2, MMRE, and Pred(l), showed promising results for web projects used in the model construction and validation process.