Machine learning reveals factors for successful crowdfunding

Modern crowdfunding has grown from relatively modest beginnings in the late 1990s to a multi-billion-dollar financing market for all kinds of early-stage innovations. The platform Kickstarter alone went from $276 million pledged in 2012 to $7.8 billion in 2024. There are even professional project designers to help craft that winning proposal.

With stakes like those, getting the pitch right is everything.

Enter machine learning to assist. Researchers from the University of Toronto’s Rotman School of Management put four different types of this artificial intelligence application to the test, including Deep Learning. Machine learning proved not only superior to conventional statistical methods for predicting whether a crowdfunding campaign would reach its goal, it also identified which elements helped most, and how.

The findings are published in the Journal of Business Venturing Design.

“Running a crowdfunding campaign is costly and could fail,” says Ramy Elitzur, a professor of accounting at the Rotman School. “Our analysis shows project creators ways to improve their chances of success, or, alternatively, whether they should pursue a different project funding strategy.”

Kickstarter is an all-or-nothing platform, meaning that project creators don’t receive any money unless they meet their fundraising goal. Analyzing more than 100,000 Kickstarter projects, Prof. Elitzur, Prof. David Soberman, who is the Canadian National Chair of Strategic Marketing, and other researchers found that the size of the campaign’s monetary goal accounted for more than half of a project’s success. The creator’s social capital, the number of reward options offered, and the campaign’s duration were also top factors.

Machine learning also got into the nitty-gritty of how much and how long. A project’s chances of success remained pretty much the same up to a fundraising goal of $100,000, then began to drop beyond that, with a sharper drop-off over $133,300. Conventional standard regression models, however, predict that the chance of success consistently drops as the monetary goal increases. That’s because these models lean towards identifying “linear” relationships where influencing factors and outcomes move in only one direction.

Crowdfunding, like many things, is more complex, with multiple variables acting on each other as well as the outcome.

“One of the things machine learning does is model all possible interactions among variables,” says Prof. Elitzur. “It gives us the direct effect on the outcome of each variable and the total effect of the interaction with other variables.”

While standard regression showed success increased with greater social capital—measured through the number of comments a project earned—machine learning revealed that success actually increased up to about 750 comments, then leveled off.

Its results also suggested the sweet spot for campaign duration was 10 to 15 days and that the number of project reward options has a moderate positive effect on success up to about 15 reward options, then slightly negative effects between 15 and 20 options, followed by a positive effect in waves between 20 and 50 reward options, finally plateauing after 50 options.

When machine learning’s text analysis capabilities were deployed—something numerically based standard methods can’t do—they could reach beyond Kickstarter’s 15 main project classifications to identify “gadgets” as the least successful project type.

It turns out that creators looking for that winning proposal should steer clear of flyable Second World War aircraft. In that case, “the deck is stacked against you and you would have a lower likelihood of success than any other domain,” says Prof. Elitzur, who is currently applying the same methods to predicting high-tech start-up success.

In addition, the text analysis capabilities of the models illustrate that as with real estate, the location of the project counts.

The research was also co-written with Noam Katz of Ben Gurion University in Israel and Perri Mutath of The Israel Innovation Authority.