This paper examines if voters act upon information about the quality of public service delivery. We explore a natural experiment in Brazil, which provided an objective measure of quality for some public schools, but not others. This creates variation in the availability of information to voters about the quality of schools inside a municipality. To use this variation, we look at polling stations located in municipal schools and compare electoral outcomes in mayoral elections in informed and non-informed groups of voters, before and after the information release. We find that, when the information received by voters is good news, the support for the incumbent increases. For the worst performers in our sample, providing information about school quality implies a decrease in the vote-share of the incumbent. We find that these effects are stronger in municipalities without local radios and/or newspapers. We do not find relevant heterogeneities when decomposing our effects according to schooling.