Towards Understanding the Success of Crowdsourced Ideas: An Analysis of the My Starbucks Idea Platform

In the last decade, idea platforms have been developed to channel customer wishes and to crowdsource new ideas for services and products. Due to their popularity and the high engagement of the customers on those platforms, the question why are certain ideas successful and the prediction of the idea successes is a current research topic. In this thesis, we exploit data crawled from the My Starbucks Idea Platform. We address the question how it is possible to model and cluster ideas to topics based on word embedding, a state of the art text mining technology. In a second step, we address the question if those topics change over time and if the chance of the success of ideas is significantly higher in different topics.