Perceived Value of Technology Product Features by Crowdfunding Backers: The Case of 3D Printing Technology on Kickstarter Platform

Abstract

Crowdfunding is an activity that gathers funds by drawing on a relatively small contribution from a relatively large number of individuals using the internet. One of the crowdfunding purposes is to fund entrepreneurial ventures. Modern crowdfunding activities--that utilize the internet--go back to 1997 and gained popularity in the music and video community. However, the most common platforms for entrepreneurial activities, including Kickstarter and IndieGoGo, have been established as recently as 2008. Thus the understanding of crowdfunding's dynamic is in its infancy. Crowdfunding has been studied from various perspectives, primarily focusing on the factors that increase the platform's participation and determinants that make a campaign successful. Most literature considered general determinants for analyzing the outcome of a campaign. These approaches mediated the differences in project type by adjusting the impact of determinants from one project category to another. There are indications that mediation is not sufficient to explain the differences, especially in the technology category with a unique behavior--for instance, the technology category has the lowest rate of success yet the third-highest amount of raised money for successful projects on Kickstarter platform. It is believed that the presence of videos and pictures on a project's campaign page has a positive influence on the campaign's success. However, a mandate of providing videos and pictures for technology products is not helpful to improve the success rate. On the other hand, a higher complexity compared to other types of products such as art, music, film, or game explains the lower success rate. According to the diffusion of innovation theory, complexity impedes product adoption. The relative advantages of complex innovations are a vital attribute to overcome complexity impediment, especially when the decision for adoption is taken under a high amount of uncertainty. In this dissertation, I studied the perceived value of technology-product features by crowdfunding backers to provide insights into what appeals to technology backers to support a complex and risky project. This approach combines aspect and opinion extraction--Double Propagation--to efficiently extract a comprehensive set of product features and regularized logistic regression to deal with the sparsity of product features and analyze the impact of technology features on the campaigns' success. Furthermore, I overcame the trade-off issue between statistical validation and detecting the impact of non-dominant features by utilizing a bootstrapping technique and marking identified advantages as "statistically verified" or "verified by subject matter expert." This work mainly makes contributions to crowdfunding theory, including establishing