We investigate the extent to which a company’s drug successful development benefits from previous - successful or failed - related development efforts in-house (experiential learning) or from other companies (vicarious learning). In a comprehensive framework, we also distinguish learning within and across therapeutic domains, as well as learning in the form of belief formation (at the start of a project) from learning as belief updating (during a project). Based on extensive information on 6,518 drug development projects, we find that projects that build on previous successes are more likely to develop marketable drugs, with the effects most pronounced for experiential learning, within-domain learning, and belief updating. Learning from previous failures also occurs, but only in the context of updating beliefs within the same domain. Belief formation learning from prior failure reduces the probability of drug development success if it concerns vicarious learning across domains.
