This project will explore some technical problems in the area of "crowdfunding:" inducing members of the public to contribute small amounts of money at a time to fund projects. Crowdfunding web sites are an instance of a commonly studied problem called a "matching marketplace:" a central place where two types of people come to be matched according to some criteria. Other examples of this are job websites, where employers and employees come to be matched (find jobs), and online dating sites. Crowdfunding exhibits a "complementarity:" a project cannot start until enough people donate money. One person's donation is not enough to support a project, but if a person waits for others to donate first, valuable projects may be left unfunded. Many real-world matching marketplaces involve complementarities, including couples finding jobs together and a project manager trying to put together a team that works well together.
This project brings together ideas from both computer science and economics. Both disciplines study matching problems, but neither has found a completely satisfactory solution in the presence of complementarities. In economics, the most common solution is a "deferred acceptance" mechanism, which breaks down when dealing with complementarities. Computer scientists frequently use "recommender systems" that make recommendations to people to help improve matches. But it isn't even clear what should be recommended when there are complementarities. This research will develop an understanding of the role of complementarities in both of these types of solutions, and apply insights across fields to design better methods of matching.
The project will test these ideas in the real world by developing a crowdfunding system. This system will provide a test bed for new algorithms and mechanisms. It will also support a vibrant community of students by helping them to fund small creative projects that develop important skills. Crowdfunding is increasingly being applied to fund high-quality professional journalism (on sites such as Spot.Us), and improved crowdfunding systems have the potential to be a new funding source for this struggling industry. This research project will also train both undergraduate and graduate students in the conduct of interdisciplinary research.
