Altruists’ impacts often fall short of their intentions. To examine this I model an altruist who derives warm glow from his perceptions of outcomes, as opposed to the outcomes themselves. For example, a donor who sponsors a child in a developing country feels good about his perceptions of the child’s wellbeing, accurate or not. Such an altruist rationally avoids most (but not all) information that could guide his actions, and all feedback on their effects. On the supply side, intermediaries such as charities can increase revenue by limiting the availability of specific kinds of information, depending on donor motives. These incentives are not unwound by competition. Optimal policy trades off the quantity and ``quality’‘ of giving.