We develop a modeling approach to explain demand for an online platform of user-generated content, and measure the impact of marketing activities on decisions to visit the platform and on decisions to create and buy content. The model explains individual-level decisions as a function of consumer characteristics, marketing activities, and behavior of other online users, allowing for the possibility of interdependence of decisions and inuence of other agents. Empirically, we apply our model to

more »

... Hewlett-Packard's print-on-demand service of user-created magazines, named MagCloud. We use two distinct data sets to show the applicability of our approach: an aggregate-level data from Google Analytics, which is widely available to managers of websites, and an individual-level data set from HP. Our results show that price and content creator marketing actions have strong eects on the number of purchases, while promotion by the rm inuences visits and creation of content. We provide recommendations to the level of marketing investments and quantify the benets of free promotional activities from content creators. According to our ndings, 8% of MagCloud's prots are directly related to these actions. This type of free marketing is likely to have a substantial presence in most online services of usergenerated content.