Dynamics of Pledge Behavior of Crowdfunded Projects

摘要

This paper aims to provide ways to enhance overall performance of crowdfunding platforms by improving success prospects of projects post-launch. Pledge behavior at the initial stages of project launch is a key indicator of project success. So, this work identifies projects to be promoted on basis of their pledge behavior at such a crucial phase. The time series of pledge amount is analyzed to understand dynamics of funding pattern and to predict a project's chances of successful funding. Statistical analysis was performed on two different datasets of projects launched over crowdfunding platform Kickstarter. The results obtained provide better understanding of the funding pattern of successful and unsuccessful projects. On the basis of behavior pattern, projects are classified as overfunded, funded, potential and low potential. To classify a project, Euclidean distance of the target project with median of the funding pattern of different categories is used to find closest category to which a project belongs. This process is effective and less expensive in terms of computation.