We develop a model where reward-based crowdfunding enables Örms to obtain a reliableproof of concept early in their production cycle: they learn about total demand from a limitedsample of target consumers pre-ordering a new product. Learning from the crowdfunding samplecreates a valuable real option as Örms invest only if updated expectations about total demandis su¢ ciently high. This is particularly valuable for Örms facing a high degree of uncertaintyabout consumer preferences, such as developers of innovative consumer products. Learning alsoenables Örms to overcome moral hazard. The higher the funds raised, the lower the Örmsíincentives to divert them, provided third-party platforms limit the sample size by restrictingcampaign length. While the probability of campaign success decreases with sample size, theexpected funds raised are maximized at an intermediate sample size. Our results are consistentwith stylized facts and lead to new empirical implications.