Thesis: By its prototypical nature, each musical good, and by extension each cultural good, is new. The aim of this thesis is two study the consumption and the funding of musical novelty, using two experimental approaches : the use of in-lab experiments to study demand (part I) and the use of experimental measures to understand field behaviors (part II). The first part explores the determinants and characteristics of demand for novelty. In the first chapter, we study the demand concentration when consumers can choose between established artists and new entrants. The second chapter presents estimations of an almost ideal demand system for four musical genres. The second part of this thesis focuses on contributors’ behaviors of a reward-based crowdfunding platform. In a third chapter, we propose a model of decision to contribute to a musical project, based on the observation that contributors are exposed to two types of risk : a risk of coordination failure and a risk of non delivery. With this in mind, illusion of control allows to understand the timing of decision. A closer look at the role of risk preferences shows that risk aversion is negatively correlated with contributions when coordination is ensured. On the contrary, the correlation becomes positive at the beginning of a campaign. In the last chapter, we investigate the mixed nature of crowdfunding. Results suggest that the decision to contribute falls within a donation logic while the decision on how much to contribute falls within a consumption logic.