Examining the Success of Crowd-funded Movies Using Topic Mining and Matching: An Empirical Study of Kickstarter

Previous studies have examined various factors such as campaign characteristics, private and social networks that might influence the outcomes of the fund-raising campaigns. However, as a key part of any crowdfunding proposal, the written descriptions of the projects have rarely been studied for analyzing crowdfunding successes. Many entrepreneurs are struggling with the question of how to market new ventures ideas to backers in a more effective and successful way on crowdfunding platforms.

In this proposed research, we examine how different features of textual information are related to the success in raising funds in the context of movie projects. Specifically, we develop a framework to delineate the key features from project descriptions that can affect the fund-raising outcomes and determine what makes a good fit between the pitch presentation and the genre of movies. We use natural language processing techniques (LIWC) to extract linguistic and topical features from a sample of 1559 crowd-funded movie projects collected from Kickstarter. We then test the effect of these key features on the campaign performance in terms of topical features and language style.

The results show that the appropriate description of fundraiser’s cognitive process and appropriate use of first-person plural and negative emotion can increase the likelihood of success. However, the frequent use of words related to perceptual process like “see”, “hear” might have an adverse effect on the project success. The managerial implication of our research is that new ventures can apply the proposed methodology to identify the most influential topical features embedded in project descriptions for different genres of movies and hence to better promote their startup projects and improve the chance of fund-raising success.