The social media effect on the success of Leetchi crowdfunding projects


Karina Sokolova, Charles Perez


Keywords : Crowdfunding, Factors of success, Social network analysis, Twitter campaign, Decision support 

Submitted :  MSNDS 2017: The 8th International Workshop on Mining and Analyzing Social Networks for Decision Support



Crowdfunding platforms have known an increasing interest during the last decade. The global crowdfunding industry estimates the fundraising volume to be up $60 Billion in 2016, surpassing traditional forms of finance. Crowdfunding by counterparts and lending have been extensively studied but, until now, only a very small amount of research were focused on donation types of platforms. In this paper, we propose a first analysis of the Leetchi donation crowdfunding platform from a social network activity perspective and investigate the social media factors of project success. This study aims to provide some first key elements of project success that could be used for guiding project owners into more successful decisions and strategies on social media. We collected and analyzed the Twitter activity related to any Leetchi projects during a period from November 1st to November 30th 2016. We finally obtained a total set of 19,000 tweets mentioning 430 Leetchi projects. We define and investigate seven social media features: the number of tweets, number of profiles, number of influencers, audience of the top influencer, the average number of followers and friends and statuses of people mentioning the projects. In addition, social network analysis features are generated by modeling Leetchi social media campaigns as a diffusion graph. We therefore extend our analysis with eight additional features: the order, the size, the number of connected components, average in/out-degrees, weighted average in/out-degrees, the density of the diffusion graph. We, finally, study and highlight the predictive power of those 15 features on the various definitions of success of Leetchi projects using machine learning algorithms. We found that while traditional social media metrics are important for predicting the reach of small founding (from 1€ to 100€), social network features are better indicators for predicting projects with a higher level of founding (from 100€ to 1,000€).