This paper develops a method to estimate and simulate the adoption of a network good. I estimate demand for mobile phones as a function of individuals’ social networks, coverage, and prices, using transaction data from nearly the entire network of Rwandan mobile phone subscribers over 4.5 years. Because subscribers pay on the margin for calls, the calls placed reveal the value of communicating with each contact. This feature allows me to overcome traditional difficulties in measuring network effects, by estimating the utility of adopting a phone based on its eventual usage. I use this structural model to simulate the effects of two governmental policies. An adoption subsidy had a high social rate of return, and spillovers accounted for a substantial fraction of its impact. A requirement to serve rural areas lowered the operator’s profits but increased net social welfare. Benefits from this policy were widely dispersed, with the majority accruing to individuals in areas where coverage was not affected.